EVALUATION OF THE MEDICARE PREFERRED PROVIDER ORGANIZATION
(PPO) DEMONSTRATION

Final Report

Authors: Gregory C. Pope, M.S.
Leslie M. Greenwald, Ph.D.
John Kautter, Ph.D.
Nathan West, M.A.
Shula Bernard, Ph.D.
Wayne Anderson, Ph.D.
Lee R. Mobley, Ph.D.
Judith Lynch, B.A.

Project Director: Gregory C. Pope, M.S.

Associate Project Director: Leslie Greenwald, Ph.D.

Federal Project Officer: Penny Mohr

RTI International

CMS Contract No. 500-00-0024 T.O. #5

March 2008

This project was funded by the Centers for Medicare & Medicaid Services under contract no. 500-00-0024 T.O. #5. The statements contained in this report are solely those of the authors and do not necessarily reflect the views or policies of the Centers for Medicare & Medicaid Services. RTI assumes responsibility for the accuracy and completeness of the information contained in this report.

ACKNOWLEDGMENTS

The authors thank the following current or former RTI staff for their contributions to this report: Eric Olmsted, Brian Dulisse, Melvin J. Ingber, Adam Hinman, Scott Scheffler, Mark Bruhn, Linda Andrews, Philip Salib, Susan Haber, Jiantong Wang, and Gordon Brown. Norma DiVito, Terry Hall, Michelle Bogus and Nanci Pepoli produced the final report. Aleksandra Petrovic, Nora Rudenko, Robert Baker, Helen Margulis, and Jenya Kaganova provided computer programming support. Above all, the authors thank our former CMS Project Officer, Victor McVicker, for his steady and insightful guidance, assistance, review, and input to this report. The authors also thank other CMS staff who reviewed and corrected this report. Finally, the authors thank the health plan representatives who participated in our site visit interviews and gave us their perspective on the Demonstration. Any remaining errors are our responsibility.

Symbols
- - -	Data not available
•••	Category not applicable
-	Quantity zero
0.0	Quantity more than 0 but less than 0.05
*	Figure meets standards of reliability or precision

EXECUTIVE SUMMARY

ES.1 Background

The purpose of this project was to evaluate the Centers for Medicare & Medicaid Services' (CMS's) Medicare Preferred Provider Organization (PPO) Demonstration, which began offering services to Medicare beneficiaries on January 1, 2003 and ended on December 31, 2005. By initiating this Demonstration project, CMS had the following policy goals (CMS, 2006):

Fulfill the ideals of the Medicare health plan program by expanding the number and types of managed care products available to Medicare beneficiaries as alternatives to traditional Medicare fee-for-service (FFS);
Provide greater access in Medicare to PPOs, which are popular in the employer-sponsored private commercial insurance market;
Test the impact of enhanced payment and risk-sharing arrangements between CMS and the plans on the range of options and benefits available to beneficiaries.

At the time the Demonstration was announced, only 2 PPO plans (iterations of an earlier demonstration) were offered to Medicare beneficiaries, despite widespread availability in the private sector. Thus, before 2003, Medicare and its beneficiaries had very limited experience with the PPO model.

CMS solicited Demonstration contracts in an April 15, 2002 Federal Register notice, with proposals due by May 30, 2002. CMS awarded contracts to 17 organizations comprising 35 individual Demonstration plans in 23 states. Fourteen of the new PPO options were located in market areas from which Medicare health plans had recently exited. All but one of these Medicare PPOs were products of organizations already offering Medicare health plans.

The key features of the PPO Demonstration were

A requirement for Demonstration plans to provide coverage for services obtained by their enrollees from providers not participating in the plan's contracted provider network;
Higher capitation payment rates to Demonstration plans in some counties (payment of the greater of the standard rate or 99 percent of per capita FFS expenditures);¹
Optional symmetrical risk sharing between CMS and the plan around a targeted medical loss ratio;
Greater actuarial flexibility in benefit and cost sharing design (the requirement that cost sharing not exceed Medicare FFS in actuarial value was relaxed);
$100,000 in start-up funding for Demonstration plans;
Streamlined application and reporting requirements for Demonstration plans. An Adjusted Community Rate proposal was not required, and quality measures were selected that were appropriate for PPOs (given the possibility of out-of-network provision of services).

ES.2 Key Evaluation Findings

ES.2.1 Plan Participation and Availability

Participation

Seventeen parent companies participated in the Demonstration, including a mix of large national and regional insurers and smaller, local independent and provider-owned plan sponsors.
Only one participating parent company was new to Medicare.
The Demonstration began on January 1, 2003 with 31 contracts offering 60 plans. Of 60 plans offered, 53 were open access and 7 were employer-specific plans. In 2003, Demonstration offerings dominated available Medicare PPO plans.
In 2004, the number of PPO Demonstration contracts rose by 4 to 35, and the number of Demonstration plans increased from 60 to 84. The majority (16 of 24) of the new plans were employer-specific.
In 2005, PPO Demonstration contracts declined by one to 34 (one contract withdrew), but plans offered continued to rise, from 84 to 125. Non-Demonstration PPO offerings increased dramatically in 2005 because of higher Medicare Advantage (MA) payment rates, the anticipation of the moratorium on offering "Local" PPOs in 2006 and 2007, and the anticipated introduction of the Medicare Part D prescription drug benefit in 2006.
The third and last year of the Demonstration was 2005. Of the 34 Demonstration contracts in effect in 2005:

Twenty-seven were offered as Local PPOs, a regular part of the MA program, in 2006.

Two were consolidated into previously existing MA contracts in 2006;

Five were not offered in 2006;

An additional 5 of the original Demonstration contracts still offered in 2006 were no longer offered in 2007.

Availability

Demonstration contracts were located in 21 states in all 4 Census regions and in 9 of the 10 CMS regions, but were concentrated in the Mid-Atlantic, Midwest, and Southeast states (29 of 35 contracts). Notably, no Demonstration contracts operated in California, the largest Medicare managed care market. The California market was dominated by successful Medicare Health Maintenance Organizations (HMOs), making entry of PPOs difficult.
Overall, open-access Demonstration plans were available in only 6 percent of counties in 2003, rising to 10 percent by 2005. But Demonstration plans were offered in 22 percent of large urban counties in 2003, rising to 34 percent by 2005. Demonstration plans were less available in small urban and rural areas.
By region, open-access Demonstration plans were most available in the Northeast. A Demonstration plan was available in one-third of Northeast counties in 2003 and 2004, rising to over 40 percent in 2005. Availability was markedly less in other regions, although it grew rapidly in the Midwest over the course of the Demonstration.
PPO Demonstration plans were primarily an urban phenomenon. In 2003, over 80 percent of counties (151 of 185) in which open-access Demonstration plans were offered were urban counties, and the only rural counties in which plans were offered were adjacent to urban areas. The Demonstration provided no evidence that local PPO plans were more likely than other MA plan types requiring a contracted provider network (primarily HMOs) to expand Medicare managed care options in rural areas. Demonstration PPOs were more likely to serve larger urban counties than existing Medicare coordinated care plans (CCPs, mostly health maintenance organizations—HMOs).
Descriptive and multivariate analysis showed that the most powerful predictor of PPO plan entry was greater existing Medicare or commercial managed care presence in an area. Demonstration plans located mostly where other Medicare coordinated care options were available and relatively successful. Two reasons for this were the aggressive start date of the Demonstration (less than a year after the Demonstration was announced and applications were due) and the difficulty of building provider networks in new service areas, as opposed to executing an addendum to existing provider HMO contracts. But Demonstration plans did increase beneficiary choice of Medicare coordinated care options, and in a modest number of counties (21 in 2004), Demonstration PPOs were the only coordinated care option available to Medicare beneficiaries. We found no evidence that the higher capitation payment rates (100 percent of fee-for-service expenditures) offered under the Demonstration in some counties in 2003 induced greater plan entry in those counties.

ES.2.2 Plan Premiums, Benefits, and Cost Sharing

Premiums

The mean Demonstration plan premium was $90.38 in 2003, then declined to $75.63 in 2004 and $60.69 in 2005. These premium reductions mirror an overall trend for MA plans following the Medicare Modernization Act (MMA)-mandated increase in MA capitation payment rates that took effect in early 2004. Most commonly, Demonstration plans' premium was between $50 and $75 in 2003 and 2004, and between $25 and $50 in 2005. In each year, a few Demonstration plans charged no premium; the maximum premium charged by any Demonstration plan varied from $170 in 2005 to $227 in 2004.
PPO premiums were generally higher than competing HMO options, but lower than the most popular Medigap plan. PPOs are therefore a mid-range product, offering higher premiums but more provider access than HMOs, but lower premiums and less provider access than Medigap.

Benefits

All Demonstration plans provided out-of-network benefits, the key distinguishing feature of PPOs from HMOs.
Most Demonstration contracts (91 percent) offered a plan with an outpatient prescription drug benefit. Demonstration plans were more likely than competing CCPs to offer a drug benefit (91 versus 79 percent). However, when offered, the PPO drug benefit was less generous on average than that of competing CCPs.
In addition to out-of-network benefits, Demonstration plans enhanced benefits and covered additional services as compared to the standard Medicare FFS benefit package. But a lower proportion of Demonstration plans than competing CCPs provided supplemental vision, hearing, and dental benefits.
Demonstration PPOs provided less restrictive access to network physician specialists than CCPs. Seventy-two percent of competing CCPs required referrals for a specialist visit compared to only 10 percent of Demonstration PPOs.

Cost Sharing

Thirty-nine percent of Demonstration plans had in-network out-of-pocket maximums while only 23 percent had out-of-network out-of-pocket maximums. Among PPOs that had a maximum, the in-network out-of-pocket maximum was typically about $1,800. The out-of-network out-of-pocket maximum was typically about $3,250, when it existed.
PPOs (and CCPs) occupied an intermediate position between FFS and Medigap in terms of out-of-pocket costs and risk protection. PPOs had higher predicted enrollee expenses than FFS for enrollees in good health but lower expenses for enrollees in poor health. Compared with Medigap, the situation was reversed: PPOs were less expensive for enrollees in good health, but more expensive for enrollees in poor health. PPOs exposed enrollees to more health-related financial risk than Medigap plan F, but less than FFS alone.

ES.2.3 Plan Enrollment, Market Share, and Prior Enrollment Status

Enrollment

Initial (2003) enrollment in the Demonstration was dominated by one contract, Horizon Healthcare of New Jersey. Nearly 50,000 of the approximately 70,000 Demonstration enrollees in 2003 were in Horizon. Almost all of the initial Horizon enrollees transferred from a 2002 Horizon HMO product that was discontinued effective contract year 2003.
Although much smaller than Horizon, several other Demonstration plans attracted significant 2003 enrollment. The second-largest enrollment plan, also in New Jersey and also with antecedents in a discontinued product, enrolled nearly 6,000 beneficiaries in 2003. Five other contracts enrolled more than 1,000 beneficiaries in 2003. The other 24 Demonstration contracts available in July 2003 had relatively modest initial enrollments of less than 1,000.
Horizon enrollment was relatively flat throughout the Demonstration (2004 and 2005), but non-Horizon enrollment rose steadily, if not dramatically. In 2004, total Demonstration enrollment exceeded 100,000 beneficiaries, more than half of whom were enrolled in non-Horizon contracts. In 2005, the Demonstration's last year, enrollment was nearly 125,000, with more than 70,000 of the total in non-Horizon plans.
By 2005:

Two contracts enrolled more than 10,000 beneficiaries (Horizon in New Jersey, Health Net in Oregon/Washington);

Four contracts enrolled between 5,000 and 10,000 beneficiaries (OSF in Illinois, Aetna in New Jersey, Group Health Incorporated in New York, Coventry in Ohio/West Virginia);

Six contracts enrolled between 2,500 and 5,000 (United in Illinois/Missouri, Aetna in Maryland, HealthNow in New York, United in North Carolina, Aetna in Pennsylvania, United in Florida);

Three contracts enrolled between 1,000 and 2,500 (Humana in Florida, Advantage in Indiana, United in Florida);

The remaining 19 contracts offered in 2005 enrolled fewer than 1,000 beneficiaries.

Contracts that experienced rapid enrollment growth over the Demonstration period included Health Net in Oregon/Washington, OSF in Illinois, Group Health Incorporated in New York, and Coventry in Ohio/West Virginia. There is no obvious similarity in these "successful" contracts, as they ranged from the West Coast to the East Coast, from urban to rural service areas, from large national/regional MA plan sponsors to small local plan sponsors, and from plans facing little MA competition in their service area to plans facing substantial competition.

Market Share

As of March 2004, Demonstration PPOs' enrollment share of all Medicare beneficiaries in their combined service areas was 1.0 percent, and their share of total Medicare health plan enrollment in their service areas was 4.6 percent. Only 4 Demonstration contracts had more than a 1 percent enrollment market share in their service areas: Horizon in New Jersey (4.1 percent), Coventry in Ohio/West Virginia (3.5 percent), OSF in Illinois (2.8 percent), and Aetna in Maryland (1.1 percent).

Prior Enrollment Status

Excluding enrollees continuing in the Horizon Demonstration contract from a prior discontinued HMO product, PPOs drew about the same proportion of their enrollees from FFS as did competing CCPs (42 percent versus 40 percent). Also, compared with CCPs, PPOs drew a somewhat lower proportion (15 percent versus 23 percent) of their enrollees from recent Medicare program enrollees (beneficiaries new to the Medicare program during the Demonstration period).
Among the 43 percent of non-Horizon PPO enrollees previously in Medicare health plans, about two thirds (64 percent) were previously enrolled in unaffiliated plans and about one third (36 percent) were previously enrolled in affiliated plans.² Thus, the Demonstration PPOs were not simply siphoning affiliated HMO enrollment—only about 15 percent of total PPO enrollment came from this source.

ES.2.4 Findings from Case Study Plan Interviews

Reasons for Participation in the Demonstration

Two payment enhancements (higher capitation rates and risk sharing) were the only reasons for Demonstration participation rated as very important by any of the organizations.

Eleven of 16 organizations were attracted to the Demonstration by the higher county payment rates (based on 99 percent of FFS rather than the existing rate book), with 5 rating it a very important reason for their participation in the Demonstration.

The ability to share risk with CMS was mentioned by 9 organizations as an appealing feature of the Demonstration, with 4 rating it as very important for their participation decision.

The very aggressive time frames for the Demonstration implementation—less than 1 year between Demonstration application due date and first enrollments—may have all but eliminated the chance that organizations new to Medicare would participate in the Demonstration.
The features of the Demonstration—primarily the favorable payment rates—did keep some organizations in the Demonstration when they might otherwise have left Medicare altogether.
In addition, some organizations either reentered Medicare markets from which they had previously withdrawn or expanded into areas where they did not offer a plan.

Product Pricing and Design

Many organizations wanted their Demonstration plans to be competitive with Medigap alternatives and priced them accordingly (i.e., under competing Medigap products). But PPOs were price above more restrictive and tightly cost-managed HMO products.
Cost sharing was set to be able to attain the desired premium (e.g., to underprice Medigap). Organizations typically set in- versus out-of-network cost sharing to give enrollees strong incentives to use in-network providers.
Many organizations believed that drug benefits were felt to be necessary to attract enrollees to the PPOs from Medicare supplements and to be competitive with other Medicare health plan offerings.
Plans did not have a uniform perspective on expected risk selection in PPOs. Some thought that sicker beneficiaries who wanted to use out-of-network providers would be attracted to PPOs, creating adverse selection. Others thought that beneficiaries who were healthy and wanted to travel would be especially attracted to the PPO out-of-network benefit. In general, organizations were more concerned about adverse selection from plans' drug benefits than from the PPO out-of-network benefit or plan type per se. They kept the drug benefits similar to other Medicare health plan drug benefits to avoid adverse selection.
Offering employer group-only plans was attractive to some organizations because group rather than individual enrollment resulted in lower per member acquisition costs and a lesser possibility of adverse selection.

Factors Influencing Beneficiary Enrollment

For most organizations, with the exception of Horizon and Aetna (whose enrollments were high during the initial year of the Demonstration), Demonstration enrollment was slower than projected. Most organizations with disappointing enrollment concluded that Medicare beneficiaries simply needed time to "get used to a new product." Most organizations marketing the PPOs to Medicare beneficiaries found that getting beneficiaries to even consider changing insurance options was a major hurdle.
Price appeared to be the factor that most influenced Medicare beneficiaries' willingness to consider PPOs. The organizations we spoke with observed that many beneficiaries are willing to pay sometimes high Medigap premiums to maintain freedom of provider choice and access to all services without referral. That said, many organizations also noted that beneficiaries are sometimes willing to consider other options if they can save money, particularly among beneficiaries who may be feeling increased financial pressures. Beneficiaries did not seem willing, at least so far, to pay higher premiums for a PPO/POS than for available HMOs or to switch from Medigap without substantial savings. The Demonstration organizations that were most successful tended to offer a well-priced PPO option.
Some organizations felt that their initial Demonstration PPO offerings were not sufficiently differentiated from their HMO product(s), which might have hurt first-year PPO enrollment. For example, the PPO and HMO provider networks and benefits within the same organization were often the same, except for the PPO's out-of-network benefit. These organizations planned to better differentiate their PPO from their HMO products in the future. For example, the future PPO product might have a wider provider network or richer benefits than the HMO.
The Demonstration organizations reported that the PPO coverage of out-of-network providers did not attract large numbers of beneficiaries; the out-of-network benefit per se was not valued enough by most beneficiaries to draw in large enrollments or command a large price premium over HMOs offered in the marketplace. It is also true that organizations required substantial out-of-network cost sharing, generally 20 percent or 30 percent coinsurance, often with a deductible and a high (or no) out-of-pocket maximum. Although this served to give enrollees strong incentives to use in-network providers, it meant that utilizing the PPO out-of-network benefit was very expensive for them, limiting its attractiveness. In our site visits, several organizations told us that beneficiary concerns over possibly high out-of-pocket costs, or uncertainty about what out-of-pocket costs might be–particularly if they decided to use out-of-network services–was a major reason beneficiaries cited for choosing not to enroll in the PPO (after initially considering the product). Although many organizations used the availability of out-of-network coverage as a primary selling point of PPOs, beneficiaries seemed to focus on total monthly premiums.

Impediments for Medicare PPOs

A major challenge for Medicare PPOs was that their low enrollments limit their bargaining power with providers.
Another way in which PPOs could offer a lower price than Medigap is through managing beneficiary utilization to eliminate unnecessary or low-value care. But the organizations we interviewed did not expect significant savings relative to Medicare FFS through care management. Virtually all of the PPOs engaged in some combination of disease management, utilization review, physician profiling, or prior authorization. But the lack of primary care physician gatekeeping, the out-of-network benefit, little risk sharing with providers, and the sometimes broader PPO provider networks limited the ability to manage care in a PPO compared with an HMO. The PPO is more an "open access" model that focuses on beneficiary choice of provider-- not the tight utilization controls, provider risk sharing, and narrow networks of the most efficient providers that may exist in HMOs.
The Demonstration organizations cited several factors limiting the viability of Medicare health plans, especially those requiring contracted provider networks such as PPOs and HMOs, in rural areas. CMS currently maintains the same provider network requirements for rural areas as for urban areas. However, meeting these requirements in many rural areas can be nearly impossible, the organizations told us. The supply of physicians and hospitals, as well as other providers, in rural areas can be very limited. Because of this relative scarcity, rural providers have market power and sometimes require more than 100 percent of FFS Medicare payments in order to participate in managed care networks. Some organizations noted that rural providers were "not organized," and that contracting with individual providers in rural areas was expensive and labor intensive. Also, it is difficult to achieve the necessary scale for plan viability of 5,000 to 10,000 enrollees in rural areas. As a result, many organizations viewed rural areas as less attractive potential markets.

Marketing of Medicare PPOs

For those organizations already offering a Medicare HMO product, the target population was typically younger beneficiaries who have middle or higher incomes. PPOs were expected to appeal to a more affluent clientele than HMO enrollees, and to beneficiaries who were willing to accept some financial risk in the form of higher and less predictable out-of-pocket costs than in HMOs or in Medicare supplements.
Many organizations targeted beneficiaries who travel a lot or live as "snowbirds" for half the year and therefore value an out-of-network benefit.
Other organizations directly targeted Medigap purchasers to the PPO, particularly in markets where Medigap premiums were high and were potentially pricing beneficiaries out of that product.
A final target group many of the organizations marketed to were age-ins, or those who aged into the Medicare population at age 65. This group is seen as having the largest long-term enrollment potential, largely because these younger groups were more familiar with the PPO concept and with managed care in general.
Nine of the 16 organizations used the acronyms "PPO" or "POS" in their plan's marketing name. Most organizations told us that most Medicare beneficiaries did not have a good understanding in the abstract of the differences between the acronyms "HMOs" and "PPOs" or "POS," and that the organizations did not emphasize the acronyms in their marketing. In a few markets (e.g., New York City), some organizations believed that "PPO" was understood well enough to be emphasized in their marketing.
Key marketing themes were "choice," "value," "freedom," and "convenience." The freedom to choose providers and the lack of need for referrals were particularly emphasized.

Provider networks and reimbursement

Because of the quick timing involved in implementing the PPO Demonstration, the organizations relied on their existing provider networks in establishing their Demonstration service areas.
Even in the longer run, several organizations saw HMOs and PPOs as complementary offerings that would be viable in the same types of markets and planned to offer PPOs in areas where they also offered HMOs. Organizations could amortize the fixed costs of network development and contracting across multiple product types. Offering Medicare health plans—including PPOs—was most attractive in areas where competitive provider networks could be contracted at sufficient discounts relative to the Medicare county rate to make the product financially viable. These were generally urban areas where competition among providers could be used to obtain discounted payment rates.
Almost universally among the Demonstration organizations, in-network providers were reimbursed at discounted Medicare FFS rates. Very few organizations reported to us any risk-bearing payment arrangements among providers; in these few cases, only primary care physician groups were paid on an at-risk basis.

Overall perceptions and comments about PPOs

In general, the Demonstration organizations had a favorable view of the PPO product for Medicare, despite slower than expected enrollments. These organizations remained committed to the PPO Demonstration product and were willing to give it some time. Many told us that launching a new product for Medicare was often a slow process because beneficiaries were frequently slow to respond to anything unfamiliar.
To make the PPO product viable for organizations, we heard most often that payment rates were a critical factor in the success of the PPO (as with any of the Medicare managed care products).
Finally, many Demonstration organizations told us that education was the key to success for the Medicare PPO product

ES.2.5 Findings from Beneficiary Surveys

Beneficiary Choice of Plan

Beneficiary awareness and knowledge of the terminology and features of PPOs in the abstract were generally low, even among those enrolled in PPO plans. But this does not necessarily mean that beneficiaries lack sufficient understanding of specific features of plans they are evaluating to make informed choices among specific plans.
Among beneficiaries who had recently switched plans, the most important reason for choosing their current option among current PPO and HMO was a desire to control or reduce their out-of-pocket costs. FFS enrollees most often cited their desire to choose a doctor or hospital.
Satisfaction with their current insurance and lack of understanding about PPO options were the two leading reasons cited by beneficiaries for not joining a PPO, with little difference in the frequency cited between current HMO and FFS enrollees. These reasons were followed by prescription drug coverage or other benefits not being good enough, and PPO plans costing too much. Cost was more frequently cited as a reason for HMO enrollees surveyed, and lack of desired benefits was cited most by FFS beneficiaries surveyed.

Beneficiary Plan Experience

PPO enrollees on average were less satisfied with their current health insurance coverage than were HMO or FFS enrollees.
More PPO enrollees reported higher out-of-pocket costs, and a smaller proportion reported no out-of-pocket costs, than HMO or FFS enrollees.
Anxiety about out-of-pocket costs was as high or higher for low- to middle-income PPO enrollees ($10,000 to $30,000 income) as for low-income PPO enrollees. Even among upper-middle-income PPO enrollees ($30,000 to $50,000 income), nearly half (47 percent) were very worried about out-of-pocket expenses. Only among high-income beneficiaries ($50,000+) did concern abate, and even in that category, more than one-quarter of PPO enrollees were very worried about costs. Although PPO enrollees frequently cited a desire to control costs as a reason for enrolling, many appear to have experienced significant out-of-pocket costs. This may account, in part, for the somewhat lower overall plan ratings cited by PPO enrollees.
Beneficiary-reported access to care in PPOs is as good as, if not slightly better than, in HMOs and FFS.
The overall disenrollment rate for PPO Demonstration plans in 2004 was more than 3 percentage points higher than the overall rate for HMO plans in the same market areas. The disenrollment rates for PPO Demonstration plans ranged from 3.5 percent to 46.3 percent with an average overall rate of 10.5 percent. The disenrollment rates for the 80 HMO plans in the same market areas ranged from 1.6 percent to 31.0 percent with an average overall rate of 7.4 percent.
PPO Demonstration disenrollees were significantly more likely to cite information problems than were HMO disenrollees. The most frequently cited reason by both groups was "After joining the plan, it wasn't what you expected." HMO disenrollees were more likely to cite Doctor Access reasons and problems getting care than were PPO Demonstration enrollees. However, the latter were significantly more likely than HMO disenrollees to cite problems with the affordability, cost, and premiums of the plan.

ES.2.6 Impact of the Demonstration on Medicare Program Payments

We estimated that the Medicare program paid more for the 89,334 beneficiaries enrolled in PPO Demonstration plans in 2003 than it would have paid in the absence of the Demonstration. The total estimated cost impact was approximately $41 million. This amounted to $457 per PPO enrollee, and 9.3 percent of estimated expenditures without the Demonstration. Four factors accounted for the higher expenditures under the Demonstration:

The 99 percent of FFS per capita expenditures payment rate paid to Demonstration plans where this payment exceeded the Medicare Advantage rates in the applicable counties in 2003. This factor accounted for about $21.6 million of the cost impact.

Demonstration plans were offered risk sharing with Medicare, which was not available in the regular Medicare program. Net Medicare risk sharing resulted in about $6.8 million in additional payments to plans.³

Demonstration plans enrolled a favorable health status selection of beneficiaries previously enrolled in the original FFS program. Capitation payments under the Demonstration were greater than estimated FFS expenditures. This factor, together with the next, accounted for about $12.1 million of the cost impact.⁴

The usual Medicare capitation payment rate was higher than average FFS per capita expenditures in some counties. This factor—which operated for all Medicare capitated plans, not just Demonstration plans—increased Medicare expenditures whenever an FFS beneficiary enrolled in a capitated plan, even with a neutral health status risk selection.

Although expenditures were higher under the PPO Demonstration, the Demonstration may have had offsetting benefits—such as expanding the range of plan choices available to beneficiaries and retaining some plans in the Medicare program—that justified higher expenditures in the eyes of policy makers. The first two factors above that led to higher expenditures under the PPO Demonstration were unique to the Demonstration; they were not a feature of Local PPOs under the regular Medicare Advantage program in 2006 and beyond. Comprehensive diagnosis-based risk adjustment for Medicare Advantage fully implemented in 2006 should lessen or eliminate the impact of the third factor. Thus, higher Medicare expenditures under the Demonstration do not imply that Local PPOs were raising Medicare payments in 2006 (and after) more than other MA plan types.

ES.2.7 Demonstration Risk-sharing Payments to Plans

For the 2003 contract year, risk-sharing transactions included payments from CMS to 14 PPO contracts of $7.8 million and payments from 5 PPO contracts to CMS of $1.0 million. On net, CMS paid $6.8 million to PPO contracts. This represented $295 per enrollee in a Demonstration contract with a risk-sharing arrangement with CMS in 2003 and for which risk-sharing amounts were available for our analysis, and 8.8 percent of total 2003 CMS payments to these Demonstration contracts. The $6.8 million in net risk-sharing payments accounted for 16.6 percent of the $40.9 million extra costs of the Demonstration.
Three Demonstration contracts received $5.5 million of the total $6.8 million in net risk-sharing payments, or 82 percent. Risk-sharing payments to some contracts were quite significant on a per capita and a percentage basis. The largest per enrollee risk-sharing payment to a Demonstration contract was $1,155 and accounted for 27.7 percent of total CMS payments to that contract. The largest total plan risk-sharing payments to CMS were $426,674, and the largest per enrollee payments to CMS were $372.

ES.2.8 Health Status and Other Characteristics of Demonstration Enrollees

We estimated that the predicted medical expenditures (costliness or health status) of PPO Demonstration enrollees in 2003 and in 2005 were about the same as those of HMO enrollees in the PPO plans' service areas. Both PPO and HMO enrollees were predicted to be substantially less costly than service area FFS beneficiaries, on average. Hence, PPOs experienced about the same degree of favorable selection relative to FFS as HMOs did.
In our beneficiary survey, we found no meaningful differences in health between PPO and HMO enrollees. However, FFS beneficiaries were sicker than PPO and HMO enrollees. Only 5 percent of PPO and HMO enrollees reported "poor" general health compared with 9 percent of FFS beneficiaries. PPO enrollees were considerably less likely (9 percent) than FFS beneficiaries (15 percent) to have severe activities of daily living (ADL) limitations (3 or more). We found no meaningful differences between PPO enrollees and HMO enrollees in ADL functioning.
PPO enrollees had a 1-year (2005) mortality rate of 3.3 percent, compared with 4.6 percent for FFS and 3.9 percent for HMO.
Compared with FFS enrollees, PPO enrollees were more likely to be non-Medicaid, non-disabled, and younger among the elderly. PPO enrollees were more similar to HMO enrollees than to FFS beneficiaries. But they were more likely to be non-Medicaid, disabled, and younger elderly than HMO enrollees.
PPO Medicare/Medicaid dual-eligible enrollment (2.5 percent) was distinctly lower than HMO (7.8 percent) and much lower than the FFS Medicaid proportion (16.0 percent).
PPO enrollees were more likely to be white and less likely to be black or Hispanic. PPO and HMO enrollees had fewer years of education compared with FFS beneficiaries.
PPO enrollees reported higher income than did HMO enrollees. One-third (34 percent) of PPO enrollees had income greater than $30,000, compared with only 24 percent of HMO enrollees. PPOs enrolled a smaller proportion of both very-low-income and very-high-income beneficiaries than did FFS.

ES.3 Conclusions

The PPO Demonstration succeeded in one of its major goals, which was to increase offerings of PPO plans to Medicare beneficiaries. A large number of both prominent national or regional parent companies and local sponsors offered a substantial number of plans under the Demonstration in a wide variety of geographic areas. This was an important accomplishment given the very tight time frame of the Demonstration from announcement through application to implementation. PPO plans were offered mostly where other Medicare managed care plans were already offered, but this was probably inevitable given the tight time frame of the Demonstration and the difficulties of developing provider networks in areas not already served by managed care plans.

The higher payment rates and other features of the Demonstration encouraged some plans to remain in the Medicare market or to reenter market areas they had exited previously. This was an important accomplishment given the large number of plan withdrawals from the Medicare program around 2003. The risk sharing available under the Demonstration made some plans willing to offer the PPO product, with its potentially risky (to plans) out-of-network utilization feature.

Although many plans were offered, enrollment and market share of the Demonstration plans was modest, especially considering that some of them captured large "rollover" enrollments from prior discontinued products of the same plan sponsor. The PPO premiums were higher than HMO premiums and so were at a competitive disadvantage for less affluent or price-sensitive beneficiaries. PPOs often offered lower premiums than Medigap but at the cost of high cost sharing for use of out-of-network providers. By the end of the Demonstration in 2005, a few Demonstration plans had become well established in terms of enrollment, but many still had small enrollment. Other than the out-of-network benefit and lack of referral requirement—which did not seem to be a strong selling point to a large proportion of Medicare beneficiaries—the PPO plans' product was comparable to plan sponsors' HMO offerings.

Not surprisingly, perhaps, the beneficiaries who enrolled in the PPO Demonstration plans tended to look a lot like Medicare HMO enrollees. In particular, they tended to be healthier than the average Medicare FFS enrollee. Compared with existing HMOs, the Demonstration PPOs did not attract disproportionately from FFS, although they did not merely target sponsors' HMO enrollment either.

PPOs—in either their Local or Regional varieties—continue to be an important part of the MA program in 2008. They still have not attracted large enrollment, growing much more slowly, for example, than private FFS plans. But given PPOs' continued dominance of the commercial employer-sponsored insurance market, there may be reason to suspect that PPOs will play an important role in Medicare's future.

SECTION 1
INTRODUCTION AND BACKGROUND

The purpose of this project has been to evaluate the Centers for Medicare and Medicaid Services' (CMS's) Medicare Preferred Provider Organization (PPO) Demonstration, which began offering services to Medicare beneficiaries on January 1, 2003, and ended on December 31, 2005. By initiating this Demonstration project, CMS had the following policy goals (CMS, 2006):

Fulfill the ideals of the Medicare health plan program by expanding the number and types of managed care products available to Medicare beneficiaries as alternatives to traditional Medicare fee-for-service (FFS);
Provide greater access in Medicare to PPOs, which are popular in the employer-sponsored private commercial insurance market; and
Test the impact of enhanced payment and risk-sharing arrangements between CMS and the plans on the range of options and benefits available to beneficiaries.

CMS solicited Demonstration contracts in an April 15, 2002 Federal Register notice, with proposals due by May 30, 2002. CMS awarded contracts to 17 organizations comprising 35 individual Demonstration plans in 23 states. Plans were selected based on evidence of their basic infrastructure, strength of their financial proposal, necessity of offering this project under demonstration conditions (as opposed to through the regular program), special geographic considerations (e.g., offering a PPO plan in an area with limited Medicare health plan options), and the organization's ability to begin enrollment quickly (ideally by January 2003). Of the new PPO options, 14 were located in market areas that Medicare health plans had recently exited. All but one of these Medicare PPOs were products of organizations already offering Medicare health plans.

1.1 History of PPOs in Medicare

PPOs, in general, are created by contractual arrangements between a financial insurer and an organization of health care providers. Unlike the traditional health maintenance organization (HMO) model, PPOs offer enrollees coverage resembling indemnity insurance, using financial incentives rather than strict provider access restrictions, to channel care to network providers. Enrollees do not need referrals to access providers (in or out-of-network), but pay higher cost sharing for services obtained out-of-network. Because PPO network providers are paid based on discounted or otherwise favorable rates, the PPO model gives enrollees incentives to use cost effective providers while not otherwise restricting their provider choice. Particularly during the last 10 years, PPOs have become an increasingly important product in the private sector. The Mercer/Foster Higgins 1999 National Survey of Employer-Sponsored Health Plans reported that PPO options had a 43 percent penetration rate among employees.

As PPOs became more dominant in the private sector, Medicare attempted to attract these options to its health plans program. PPOs offer a model of managed care that is perceived to lie "between" the traditional FFS and HMO options that were available to beneficiaries. Because individuals can have access to a wide range of physician choices without strict gatekeepers and prior approvals, PPOs may appeal to many Medicare beneficiaries enrolled in FFS who are less inclined to enroll in more restricted HMOs. A goal of the Balanced Budget Act (BBA) of 1997 in establishing the Medicare+Choice (M+C) program was to expand the options and penetration of Medicare managed care, but these goals had not been fully realized at the inception of the PPO Demonstration. A PPO program was one step in accomplishing these goals of expanded choice and enrollment in Medicare health plans. Policy makers have also wanted to "modernize" many aspects of the Medicare FFS and health plan programs by having them adopt various strategies more widely used in the private sector. Increased availability of PPOs was one element of this effort to modernize Medicare.

Though PPOs were first formally allowed under mainstream Medicare managed care when the M+C program was created by the BBA in 1997, Medicare has a relatively long history of demonstration experimentation with PPO models. The agency originally solicited for FFS-based PPO demonstrations in the late 1980s. One demonstration resulting from this first PPO demonstration project was known as "CAPP CARE." The Medicare SELECT program represented a slightly later FFS-based PPO Medicare demonstration. Under Medicare SELECT, beneficiaries received their full Medigap supplemental benefits only when they used SELECT network providers. In exchange, SELECT premiums were lower than the same benefit package, offered as a standard non-PPO product by the same insurer.

Medicare's first attempt to solicit for a PPO option under a capitated payment system came in the Medicare Choices Demonstrations. Designed and solicited during 1995 and 1996, the Medicare Choices Demonstrations were conceived of as a wide-scale test of alternative delivery and payment models for the then Medicare risk program. The Medicare Choices project waived existing participation rules and allowed new types of managed care organizations and products such as PPOs and provider-sponsored organizations. One of these Choices sites was the PPO offered by Independence Blue Cross/Blue Shield, which operated at the implementation of the PPO Demonstration under a different managed care demonstration and had roughly 20,000 enrollees.

1.2 The Medicare PPO Demonstration

Despite private sector trends toward PPOs and the expanded managed care participation rules enacted under the BBA, the Medicare health plan program continued to be dominated by traditional HMO contracts. CMS's Medicare PPO Demonstration – evaluated in this project – was a next critical step in the agency's ongoing initiatives to reinvigorate the Medicare health plan program and offer Medicare beneficiaries expanded health care choices. This Demonstration program was modeled after the PPO coverage available in the commercial market. Although all plans were required to offer out-of-network benefits, fewer specific requirements were applied to the benefit design than for most other Medicare health plans. (The requirement that cost sharing not exceed Medicare FFS in actuarial value was relaxed.) Differential cost sharing requirements in and out-of-network were intended to encourage enrollees to use services in a cost effective manner while not providing a disincentive for seeking appropriate care.

The PPO Demonstration included two changes to the standard capitation payment system as an enticement for plans to enter the Demonstration. The Demonstration PPOs were paid using an increased base payment rate (comparable to the old concept of the Average Adjusted Per Capita Cost or AAPCC methodology) of the greater of 99 percent of FFS (compared with the 95 percent of FFS basis of the standard capitation rates) or the prevailing capitation rates found in the county rate book. (Beginning in April 2004, the MA base county payment rate was increased to at least 100 percent of the FFS average expenditure, eliminating the higher PPO Demonstration payment rate.) The second adjustment included a "risk-sharing" option to protect PPO plans against higher than expected medical costs. Risk sharing was thought to be especially important to induce greater participation of PPO plans, because of the risk of out-of-network utilization that PPOs incur.

Risk-sharing arrangements, where applicable, were negotiated between CMS and each plan offering a PPO product, and were symmetrical, meaning that the sharing arrangement between CMS and the plan was the same for both losses and savings. The risk-sharing arrangement specified a targeted medical loss ratio, or medical expense target, reflected as a percentage of total plan revenue. The risk-sharing arrangement was reconciled after the close of the contract year, at which point the actual medical loss ratio was established. To the extent that medical expenses exceed the targeted medical expense by more than a pre-established amount, CMS and the organization shared in the losses. Similarly, if the participating organization experienced savings, CMS shared in the savings. All of the participating organizations that had risk-sharing arrangements with CMS as part of their Demonstration terms and conditions were at full risk below a certain threshold (up to 2–5 percent in 2003). A corridor was established around the medical loss ratio, meaning the first 2–5 percent of any loss or gain in relation to the targeted ratio was assumed by the plan. Beyond the corridor, both CMS and the plan shared gains/losses under various specified arrangements. However, CMS's share of gains/losses was never more than 80 percent.

In addition to higher payment rates in some counties, optional risk sharing, and greater actuarial flexibility in benefit and cost sharing design, the PPO Demonstration offered participating plans $100,000 in start-up funding. Also, the health plan contract application procedure was streamlined (an Adjusted Community Rate proposal was not required) and health plan quality reporting requirements were selected to take into consideration the structural differences in PPOs relative to other managed care products (specifically the implications of provision of service through out-of-network providers).

1.3 After the Demonstration: Local and Regional PPOs

The Medicare Modernization Act (MMA) (Pub. L. 108-173) was passed in late 2003, the first year of the 3-year PPO Demonstration. The MMA created the Medicare Advantage (MA) program and significantly raised payments to Medicare health plans, effective in March 2004. With MA payments raised to at least 100 percent of FFS in 2004, Demonstration plans no longer received a higher capitation rate than non-Demonstration plans in any counties. The MMA created new plan options, including Regional PPOs to be introduced in 2006, and the new Medicare Part D prescription drug benefit, also to be implemented in 2006. Similar in benefits and structure to Local PPOs,⁵ Regional PPOs are required to offer the same benefits at the same premium to an entire MA region, which includes at a minimum an entire state. The MMA, in anticipation of implementation of Regional PPOs, required that all non-regional ("Local") PPOs convert to the MA program by 2006. After that point, no new Local PPOs were allowed to enter for a two year (2006 – 2007) moratorium period. As detailed below, most of the PPO Demonstration plans transitioned to Local PPOs in 2006.

1.4 Overview of Evaluation Design

In summary, this evaluation sought to answer the question: "How well does the PPO model—as implemented in the PPO Demonstration—work for Medicare?" To answer this very general question, the evaluation was constructed in three interrelated parts corresponding to the major stakeholders in the Demonstration. This structure is described below in Figure 1-1.

Below is a figure, click here for the text describing this figure.

Figure 1-1
Structure of the PPO evaluation

Figure 1-1

Beneficiary research topics	PPO organization research topics	Medicare program research topics
Beneficiary awareness of PPOs and reasons for enrolling or not enrolling	Decision to participate in the Demonstration	Plan participation in the Demonstration
PPO enrollee versus non-enrollee characteristics	Benefit design, premium pricing, and marketing	Plan offerings and availability
PPO enrollees drawn from FFS or health plans?	Operational and implementation experience	Biased selection in PPO enrollment
PPO effects on beneficiary out-of-pocket costs	Enrollment and disenrollment trends	Effect of PPOs on Medicare costs – relative to FFS and HMOs
Beneficiary satisfaction with PPOs	Impact of risk sharing, start-up support, and higher capitation payments	Overall assessment of the lessons learned for Medicare in the PPO Demonstration

1.5 Organization of This Report

The remaining sections of the report are as follows. Section 2 analyzes plan participation, availability, and offerings in the PPO Demonstration, and beneficiary enrollment in Demonstration plans. Section 3 discusses findings from our case study site visits with Demonstration plans. Section 4 describes findings from our surveys of beneficiaries enrolled in (or disenrolled from) Demonstration plans. Section 5 summarizes our analysis of biased selection in Demonstration enrollment, and the impact of the Demonstration on Medicare program expenditures. Section 6 provides our conclusions on what was learned from the PPO Demonstration.

A note on terminology: At the inception of the PPO Demonstration in 2003, the Medicare health plan program was known as the "Medicare+Choice" or M+C program. During the Demonstration period, the Medicare Modernization Act of 2003 revised and changed the name of the M+C program to the "Medicare Advantage" or MA program. We generally use the latter (MA) terminology in this report for the entire Demonstration period, unless the context requires the use of the historical "M+C" terminology for accuracy. In addition, in this report, "PPO" should generally be understood to mean "Demonstration PPO" unless specifically indicated otherwise.

SECTION 2
PLAN PARTICIPATION, AVAILABILITY, OFFERINGS, AND ENROLLMENT

A key goal of the Medicare PPO Demonstration was to expand the health plan choices of Medicare beneficiaries, specifically to increase the availability of Medicare PPOs. This section analyzes plan participation in the Demonstration, the resulting availability of PPO plans to beneficiaries, the premiums and benefits of the available plans, and the enrollment in the available Demonstration plans. Section 2.1 documents plan participation and availability. Section 2.2 describes the premiums, benefits, and cost sharing of Demonstration plans. Section 2.3 discusses Demonstration plan use of utilization review and other cost control techniques. Section 2.4 analyzes enrollment in Demonstration plans.

2.1 Plan Participation and Availability

This section first discusses plan participation in the Demonstration, then the availability of Demonstration plans to beneficiaries.

2.1.1 Plan Participation

We analyzed two aspects of plan participation in the PPO Demonstration. First the number of contracts and plans offered in the Demonstration, then the parent companies that sponsored Demonstration plans.

Contracts and Plan Options

Table 2-1 shows the number of PPO Demonstration contracts and plan option- for 2003, 2004, and 2005. Contracts and plan options open to all qualifying Medicare beneficiaries ("open access") and open only to beneficiaries affiliated with specific employers ("employer only") are shown. Non-demonstration PPO contracts and plan options are also shown for comparison.

Table 2-1
PPO Demonstration and non-Demonstration contracts and plan options, 2003–2005
	2003	2004	2005
NOTE: Excludes Puerto Rico and US territories. SOURCE: RTI Analysis of 2003–2005 CMS HPMS Data.
Demonstration
Contracts	31	35	34
Offering employer only plans	7	11	12
Offering open access plans	31	35	34
Plan Options	60	84	125
Employer only	7	23	38
Open access	53	61	87
Non-Demonstration
Contracts	3	8	93
Offering employer only plans	0	1	33
Offering open access plans	3	8	93
Plan Options	4	37	334
Employer only	0	6	55
Open access	4	31	279

The PPO Demonstration attracted substantial plan participation. The Demonstration began on January 1, 2003 with 31 contracts offering 60 plan options. All of the 31 contracts offered at least one open access plan option, and 7 offered at least one employer-only plan option. Of 60 plan options offered, 53 were open access. In 2003, Demonstration offerings dominated available Medicare PPO plans. Only 3 non-Demonstration contracts and 4 plan options were available in 2003.

In 2004, the number of PPO Demonstration contracts rose by 4 to 35, and the number of Demonstration plan options rose from 60 to 84. The majority (16 of 24) of the new plan options were employer-only. Non-Demonstration PPO contracts and plan options increased, but the Demonstration still dominated Medicare PPO offerings in 2004.

In 2005, PPO Demonstration contracts declined by 1, but plan options offered continued to rise, from 84 to 125. Non-Demonstration PPO offerings grew rapidly, from 8 to 93 contracts and from 37 to 334 plans. Factors promoting the dramatic increase in non-Demonstration PPOs in 2005 were higher MMA-mandated capitation payment rates that were implemented in early 2004, the anticipated moratorium on new local PPO plan options in 2006 and 2007, and the anticipated introduction of the Medicare Part D benefit in 2006. Many organizations wanted to "get in the game" by offering in 2005 a plan option type (PPO) that might become significant in Medicare. By the end of 2005, non-Demonstration PPO contracts and plan options outnumbered their Demonstration counterparts by roughly 3 to 1.

Table 2-2 lists the specific contracts participating in the PPO Demonstration, with contract start and end dates and the states in which the contract service areas were located. Thirty-one of 35 Demonstration contracts participated in the Demonstration for its full 3-year extent from January 2003 to December 2005. Two contracts were implemented on September 1, 2003, and 2 more on January 1, 2004. One contract, Health Net in Arizona, withdrew at the end of 2004 prior to the end of the Demonstration.

Table 2-2
PPO Demonstration contracts, duration, and service areas
Contract ID	Legal name	Start date	End date	State(s) of contract service area
SOURCE: RTI Analysis of 2003–2005 CMS HPMS Data.
H0102	UNITED HEALTHCARE INSURANCE COMPANY	1/1/2003	12/31/2005	Alabama
H0103	UNITED HEALTHCARE INS. COMPANY, INC.	1/1/2003	12/31/2005	Alabama
H0313	PACIFICARE OF ARIZONA	1/1/2003	12/31/2005	Arizona
H0314	HEALTH NET LIFE INSURANCE COMPANY	1/1/2003	12/31/2004	Arizona
H0706	AETNA HEALTH INC.	1/1/2004	12/31/2005	New York
H1047	HUMANA INSURANCE COMPANY	1/1/2003	12/31/2005	Florida
H1408	OSF HEALTHPLANS, INC.	1/1/2003	12/31/2005	Illinois
H1412	COVENTRY HEALTH AND LIFE INS. COMPANY	1/1/2003	12/31/2005	Illinois, Missouri
H1413	UNITED HEALTHCARE INS. COMPANY, INC.	1/1/2003	12/31/2005	Illinois, Missouri
H1508	ADVANTAGE HEALTH SOLUTIONS, INC.	1/1/2003	12/31/2005	Indiana
H1715	COVENTRY HEALTH AND LIFE INS. COMPANY	1/1/2004	12/31/2005	Kansas, Missouri
H1805	ANTHEM HEALTH PLANS OF KENTUCKY	9/1/2003	12/31/2005	Kentucky
H1901	TENET CHOICES, INC.	1/1/2003	12/31/2005	Louisiana
H2110	AETNA HEALTH INC.	1/1/2003	12/31/2005	Maryland
H2903	PACIFICARE OF NEVADA, INC.	1/1/2003	12/31/2005	Nevada
H3108	AETNA HEALTH INC.	1/1/2003	12/31/2005	New Jersey
H3109	HORIZON HEALTHCARE OF NJ,INC.	1/1/2003	12/31/2005	New Jersey
H3323	GROUP HEALTH INCORPORATED	1/1/2003	12/31/2005	New York
H3324	HEALTHNOW NEW YORK, INC.	1/1/2003	12/31/2005	New York
H3325	MANAGED HEALTH INC.	1/1/2003	12/31/2005	New York
H3326	UNITED HEALTHCARE INS. COMPANY OF NY, INC.	1/1/2003	12/31/2005	New York
H3403	UNITED HEALTHCARE INS. COMPANY, INC.	1/1/2003	12/31/2005	North Carolina
H3615	COVENTRY HEALTH AND LIFE INS. COMPANY	1/1/2003	12/31/2005	Ohio, West Virginia
H3616	UNITED HEALTHCARE INS. COMPANY, INC.	1/1/2003	12/31/2005	Ohio
H3617	UNITED HEALTHCARE INS. COMPANY, INC.	1/1/2003	12/31/2005	Ohio
H3618	COMMUNITY INSURANCE COMPANY	9/1/2003	12/31/2005	Ohio
H3806	HEALTH NET LIFE INSURANCE COMPANY	1/1/2003	12/31/2005	Oregon, Washington
H3913	UPMC HEALTH BENEFITS, INC.	1/1/2003	12/31/2005	Pennsylvania
H3914	AETNA HEALTH INC.	1/1/2003	12/31/2005	Pennsylvania
H3915	HEALTH ASSURANCE PENNSYLVANIA, INC.	1/1/2003	12/31/2005	Pennsylvania
H4103	UNITED HEALTHCARE INS. COMPANY, INC.	1/1/2003	12/31/2005	Rhode Island
H4403	CARITEN INSURANCE COMPANY	1/1/2003	12/31/2005	Tennessee
H4404	HEALTHSPRING, INC.	1/1/2003	12/31/2005	Tennessee
H5400	UNITED HEALTHCARE INS. COMPANY, INC.	1/1/2003	12/31/2005	Florida
H5401	UNITED HEALTHCARE INSURANCE COMPANY	1/1/2003	12/31/2005	Florida

Table 2-3 indicates the status of PPO Demonstration contracts in 2006, immediately after the Demonstration ended in 2005. Of the 35 PPO Demonstration contracts,

1 was withdrawn previously, in 2005;
2 were consolidated into previously existing MA contracts in 2006;
5 were not offered in 2006 (United and Pacificare accounted for all of these); and
27 were offered as local PPOs, a regular part of the MA program, in 2006.

Table 2-3
Post-Demonstration (2006) status of PPO Demonstration contracts
PPO Demonstration organization	Contract ID or status
PPO Demonstration organization	2005	2006
SOURCE: CMS HPMS Data.
United	H0102	H5500
United	H0103	Not offered
PacifiCare	H0313	Not offered
Health Net (H0314)	Not offered	Not offered
Aetna	H0706	H5531
Humana	H1047	Consolidated into H5415
OSF	H1408	H5525
Coventry	H1412	H5506
United	H1413	H5507
Advantage	H1508	H5508
Coventry	H1715	H5509
Anthem	H1805	H5530
Tenet	H1901	H5501
Aetna	H2110	H5510
Pacificare	H2903	Not offered
Aetna	H3108	H5512
Horizon	H3109	Consolidated into H3154
Group Health, Inc.	H3323	H5528
Healthnow	H3324	H5526
Managed Health, Inc.	H3325	H5514
United	H3326	H5515
United	H3403	H5516
Coventry	H3615	H5517
United	H3616	H5518
United	H3617	Not offered
Anthem	H3618	H5529
Health Net	H3806	H5520
UPMC	H3913	H5533
Aetna	H3914	H5521
Coventry	H3915	H5522
United	H4103	H5527
Cariten	H4403	H5523
Healthspring	H4404	H5524
United	H5400	Not offered
United	H5401	H5532

Plan sponsors found it worthwhile to continue to offer most of the PPO Demonstration contracts by transitioning them to the regular MA program in 2006. By this measure of plan availability beyond the Demonstration period, the Demonstration can be considered a success. However, an additional 5 of the original Demonstration contracts still offered in 2006 were no longer offered in 2007 (see Table 2-12).

Parent Companies

Table 2-4 summarizes the 17 parent companies that participated in the Demonstration, and their characteristics at the time of the Demonstration. All but 1 parent company offered other MA products, although the majority did so on a local basis only. The big national or regional MA players—Aetna, Humana, PacifiCare, and United HealthCare—offered at least one Demonstration PPO, but only United (with 10 offerings) decided to offer products on a large- scale basis. Humana, in contrast, took a cautious approach to the Demonstration, offering a PPO product initially in only one Florida county. Kaiser, one of the largest MA players nationally, was conspicuous by its absence. All of the parent companies were for-profit, with the exception of the 3 New York parent companies. Most of the parent companies were insurers, although 6 of them are owned by provider networks (e.g., the University of Pittsburgh hospital system owns the insurer UPMC). The provider networks typically established the affiliated insurer as a safeguard against other insurers directing business away from them and now see the insurer as a way to cement or expand their Medicare business. Serving publicly insured populations (Medicare, Medicaid) was a core part of their mission and business strategy. The provider-owned parent companies were generally relatively small, local players, and in some cases (e.g., Advantage) were near start-ups. Several of the parent companies offered "point of service" (POS) plans under the Demonstration, rather than the more traditional PPOs. POS plans differ from traditional PPOs in requiring beneficiary selection of a primary care physician, who authorizes in- and out-of-network care. POS plans were easier to implement quickly under existing state HMO insurance licenses and provider payment arrangements.

Table 2-4
PPO Demonstration parent companies
Demonstration parent company	Plan type¹	# of Demonstration contracts	Service area	Profit/ nonprofit	Ownership	Other MA products	Scope of MA products
¹ Both PPO and POS are included in the overall group of Coordinated Care Plans (or CCPs). SOURCE: RTI International.
Advantage	PPO	1	Indiana	For Profit	Provider	Yes	Local
Aetna	POS	4	Maryland, New Jersey, New York, Pennsylvania	For Profit	Insurer	Yes	National
Anthem (includes Community Insurance Company)	PPO	2	Kentucky, Ohio	For Profit	Insurer	Yes	Regional
Cariten	PPO	1	Tennessee	For Profit	Provider	Yes	Local
Coventry (includes Health Assurance)	PPO	4	Illinois, Kansas, Missouri, Ohio, Pennsylvania, West Virginia	For Profit	Insurer	Yes	Regional
Group Health, Inc.	PPO	1	New York	Nonprofit	Insurer	No	Local
HealthFirst (Managed Health Inc.)	PPO	1	New York	Nonprofit	Provider	Yes	Local
Health Net	PPO	2	Arizona, Oregon Washington	For Profit	Insurer	Yes	Regional
HealthNow	PPO	1	New York	Nonprofit	Insurer	Yes	Local
HealthSpring	PPO	1	Tennessee	For Profit	Insurer	Yes	Local
Horizon	POS	1	New Jersey	For Profit	Insurer	Yes	Local
Humana	PPO	1	Florida	For Profit	Insurer	Yes	Regional
OSF Health Plans	PPO	1	Illinois	For Profit	Provider	Yes	Local
PacifiCare	POS	2	Arizona, Nevada	For Profit	Insurer	Yes	Regional
Tenet Choices, Inc.	PPO	1	Louisiana	For Profit	Provider	Yes	Local
United HealthCare	PPO	10	Alabama, Florida, Illinois, Missouri, North Carolina, New York, Ohio, Rhode Island	For Profit	Insurer	Yes	National
UPMC	PPO	1	Pennsylvania	For Profit	Provider	Yes	Local

2.1.2 Plan Availability

Demonstration contracts were located in 21 states in all 4 Census regions and in 9 of the 10 CMS regions but were concentrated in the Mid-Atlantic, Midwest, and Southeast states (29 of 35 contracts). Demonstration open-access plan service areas as of April 2004 are mapped in Figure 2-1. Notably, no Demonstration contracts operated in California, the largest Medicare managed care market. The California market was dominated by successful Medicare HMOs, making entry of PPOs difficult.

Below is a map, click here for the text describing this map.

Figure 2-1
Service areas of Medicare PPO Demonstration contracts, April 2004

Figure 2-1

NOTE: PPO is PPO Demonstration plans. CCP is coordinated care plans. PPOs are one type of CCP. Excludes employer-only plans.

Overall, open-access Demonstration plans were available in only 6 percent of counties in 2003, rising to 10 percent by 2005 (Table 2-5). But Demonstration plans were offered in 22 percent of large urban counties in 2003, rising to 34 percent by 2005. Demonstration plans were less available in small urban and rural areas.

Table 2-5
Number and percentage of counties with one or more open-access
PPO Demonstration contracts, by level of urbanization, 2003–2005
	No. of counties	Number			Percentage
	No. of counties	2003	2004	2005	2003	2004	2005
NOTE: Excludes employer-only plans. Excludes Puerto Rico and U.S. territories. SOURCE: RTI Analysis of 2003–2005 CMS HPMS data.
Total	3,142	185	222	318	5.9	7.1	10.1
Urban	1,089	151	180	236	13.9	16.5	21.7
Large	413	92	111	141	22.3	26.9	34.1
Medium	325	39	46	66	12.0	14.2	20.3
Small	351	20	23	29	5.7	6.6	8.3
Rural	2,053	34	42	82	1.7	2.0	4.0
Urban-adjacent	1,062	34	42	72	3.2	4.0	6.8
Non-urban adjacent	991	0	0	10	0.0	0.0	1.0

By region, open-access Demonstration plans were most available in the Northeast (Table 2-6). A Demonstration plan was available in one-third of Northeast counties in 2003 and 2004, rising to over 40 percent in 2005. Availability was markedly lower in other regions, although it grew rapidly in the Midwest over the course of the Demonstration.

Table 2-6
Number and percentage of counties with one or more PPO Demonstration
contracts, by census region, 2003–2005
	No. of counties	Number			Percentage
	No. of counties	2003	2004	2005	2003	2004	2005
NOTE: Excludes employer-only plans. Excludes Puerto Rico and U.S. territories. SOURCE: RTI Analysis of 2003–2005 CMS HPMS data.
Total	3,142	185	222	318	5.9	7.1	10.1
Northeast	217	70	72	90	32.3	33.2	41.5
Midwest	1,055	27	42	105	2.6	4.0	10.0
South	1,425	66	86	105	4.6	6.0	7.4
West	445	22	22	18	4.9	4.9	4.0

PPO Demonstration plans were primarily an urban phenomenon. In 2003, more than 80 percent of counties (151 of 185—see Table 2-5) in which open-access Demonstration plans were offered were urban, and the only rural counties in which plans were offered were adjacent to urban areas. The Demonstration provided no evidence that local PPO plans were more likely than other MA plan types requiring a contracted provider network (primarily HMOs) to expand Medicare managed care options in rural areas. Demonstration PPOs were more likely to serve larger urban counties than existing Medicare coordinated care plans (CCPs, which are mostly HMOs)—see Figure 2-2.

Below is a bar graph, click here for the text describing this graph.

Figure 2-2
Distribution of PPO and CCP counties by urbanicity, 2004

Figure 2-2

NOTE: PPO is PPO Demonstration plans. CCP is coordinated care plans. Excludes Part B only and employer-only plans.
SOURCE: RTI analysis of CMS HPMS April 2004 file.

Descriptive and multivariate analysis (Pope et al., 2005) showed that the most powerful predictor of PPO plan entry was greater existing Medicare or commercial managed care presence in an area. Demonstration plans located mostly where other Medicare coordinated care options were available and relatively successful. Two reasons for this were the aggressive start date of the Demonstration (less than a year after the Demonstration was announced and applications were due) and the difficulty of building provider networks in new service areas, as opposed to executing an addendum to existing provider HMO contracts. But Demonstration plans did increase beneficiary choice of Medicare coordinated care options, and in a modest number of counties (21 in 2004) Demonstration PPOs were the only coordinated care option available to Medicare beneficiaries. We found no evidence that the higher capitation payment rates (100 percent of fee-for-service expenditures) offered under the Demonstration in some counties in 2003 induced greater plan entry in those counties.

2.2 Plan Premiums, Benefits, and Cost Sharing

This section documents the premiums, benefits, and cost sharing of Demonstration plans. These aspects of the plans determine their relative attractiveness and value to beneficiaries.

2.2.1 Premiums

Table 2-7 shows the distribution of Demonstration PPO open-access plan monthly premiums for 2003, 2004, and 2005. The mean plan premium was $90.38 in 2003, then declined to $75.63 in 2004 and $60.69 in 2005. These premium reductions mirror an overall trend for MA plans following the MMA-mandated increase in MA capitation payment rates that took effect in early 2004 (Pope et al., 2006). Most commonly, Demonstration plan premiums were between $50 and $75 in 2003 and 2004, and between $25 and $50 in 2005. In each year, a few Demonstration plans charged no premium; the maximum premium charged by any Demonstration plan varied from $170 in 2005 to $227 in 2004.

Table 2-7
Distribution of PPO Demonstration open-access plan
monthly premium amounts, 2003–2005
	2003	2004	2005
NOTE: Excludes employer-only plans. SOURCE: RTI Analysis of 2003–2005 CMS HPMS Data.
N (plans)	53	61	87

Mean	$90.38	$75.63	$60.69
Minimum	0.00	0.00	0.00
Median	86.82	69.00	59.00
Maximum	184.00	227.00	170.00

Premium range	Number of plans
$0	2	2	7
1 – 25	0	2	7
>25 – 50	3	10	23
>50 – 75	18	24	21
>75 – 100	12	11	20
>100 – 150	13	8	8
>150	5	4	1

PPO premiums were generally higher than competing HMO/CCP options, but lower than the most popular Medigap plan (Pope et al., 2005). The distribution of Demonstration PPO, HMO/CCP, and Medigap premiums for April 2004 is displayed in Figure 2-3. PPOs are therefore a mid-range product, offering higher premiums but more provider access than HMOs, but lower premiums and less provider access than Medigap.

Below is a bar graph, click here for the text describing this graph.

Figure 2-3
PPO and competing CCP and Medigap Plan F monthly premiums, 2004

Figure 2-3

NOTES: PPO is PPO Demonstration plans. CCP is coordinated care plans. Excludes Part B only and employer-only plans. Competing plans are available in at least one PPO service area county.
SOURCE: RTI analysis of CMS HPMS April 2004 file and AARP Medigap premiums.

2.2.2 Benefits

We analyzed the benefits and cost sharing of PPO Demonstration open-access plans in April 2004, after the MMA-mandated capitation payment increases took effect. We compared benefits and cost sharing with other Medicare CCPs/HMOs—and, in some analyses, with original Medicare fee-for-service (FFS) and with Medicare supplemental insurance (Medigap)—that were competing in the same service areas as the Demonstration PPOs. Medigap standard Plan F covers the following: hospital and Part B coinsurance, hospital and Part B deductibles, skilled nursing facility co-insurance, Part B excess charges, emergency care outside of the United States, and preventative care. Medigap Plan F does not offer coverage for prescription drugs. We compared to Medigap Plan F because it is the most popular Medigap plan nationally.

Drug coverage was an important aspect of MA plan benefits during the Demonstration period, which was before the 2006 implementation of the Medicare Part D benefit. Starting in 2006, most MA contracts were required to offer at least one plan with at least the standard Part D drug benefit. Before 2006, drug benefits were not required in MA plans, although CMS strongly encouraged PPO Demonstration contracts to offer drug benefits.

Most PPO contracts (91 percent) offered a plan with an outpatient prescription drug benefit in 2004 (Table 2-8). PPOs were more likely than competing CCPs to offer a drug benefit (91 versus 79 percent). However, when offered, the PPO drug benefit was less generous on average than that of competing CCPs. Only 42 percent of PPO drug benefit plans covered brand drugs, compared with 53 percent of CCP benefit plans. Only 20 percent of PPO drug benefit plans offered unlimited generics, compared with 31 percent of CCP plans, and when there was a maximum benefit, it was typically $500 in PPO plans compared with $800 in CCPs. PPOs told us it was important to have a drug benefit to attract enrollment (Greenwald et al., 2004), but may have limited it to keep their premiums down or fund their out-of-network benefit.

Table 2-8
Prescription drug benefits of PPOs and competing coordinated care plans, 2004
	PPO	CCP
NOTES: Includes "Part A and Part B" plans only. Employer-only plans are excluded. PPO is PPO Demonstration plans. Competing CCP plans are defined by those offered in at least one PPO service area county. SOURCE: RTI International analysis of CMS HPMS April 2004 file.
% of contracts with drug benefit	91%	79%
% of plans with drug benefit	82%	70%
Plans with a drug benefit
Generic coverage only	58%	47%
Unlimited	20%	31%
Maximum benefit	38%	16%
Median annualized maximum	$500	$800
Brand drug coverage	42%	53%
Unlimited	6%	6%
Brand benefit maximum, unlimited generics	22%	29%
Median annualized maximum	$600	$900
Brand and generic combination maximum	14%	19%
Median annualized maximum	$1,000	$1,000

All Demonstration plans provided out-of-network benefits, the key feature distinguishing PPOs from HMOs. In 2003 and 2004, all Demonstration plans covered a core set of services out-of-network, including acute hospitalization, outpatient hospital services, and primary care and specialist physician services. Other standard Medicare benefits—such as skilled nursing facility stays, home health visits, and durable medical equipment—were covered by most, but not all, Demonstration plans out-of-network. After a U.S. Government Accountability Office report (U.S. GAO, 2004) noted that according to their contracts, Demonstration plans should provide all covered benefits out-of-network, CMS worked with the plans to make all covered benefits available out-of-network in 2005.

In addition to out-of-network benefits, PPOs provided richer in-network benefits than the standard Medicare fee-for-service benefit package. These supplemental benefits took the form of either enhancements to a Medicare-fee-for-service covered benefit, such as covering skilled nursing stays without the Medicare-required prior hospital stay, or providing a benefit that was not part of the standard Medicare fee-for-service package, such as dental benefits. Supplemental in-network benefits of Demonstration PPOs and competing CCPs are shown in Table 2-9.

Table 2-9
Selected in-network supplemental benefits provided by Demonstration PPOs and
competing coordinated care plans (CCPs), 2004
	Percentage of plans
	PPO	CCP
NOTES: PPO is PPO Demonstration plan. CCP is coordinated care plan. Competing CCPs are offered in at least one PPO service area county. Includes "additional" and "mandatory" benefits, excludes "optional" benefits (they require an extra premium). SOURCE: RTI analysis of CMS HPMS April 2004 file.
Enhanced Benefit
Skilled nursing facility	96.7	89.2
Emergency/urgent care, worldwide coverage	100.0	82.8
Chiropractic, routine care	18.0	9.5
Podiatry, routine foot care	55.7	54.7
Psychiatry	98.4	92.7
Disease management	70.5	71.1
Routine physical	100.0	100.0
Smoking cessation	21.3	37.1
Health club membership	8.2	30.2
Immunizations	24.6	25.4
Benefit Not Covered by Traditional Medicare
Acupuncture	1.6	5.2
Dental	32.8	43.1
Eye exams	82.0	92.2
Eye wear	50.8	75.0
Hearing exams	65.6	78.0
Hearing aids	23.0	55.6

The enhanced benefits profile of PPOs and competing CCPs was similar. Popular enhanced PPO and CCP benefits included non-Medicare-covered skilled nursing stays, psychiatry, disease management, and routine physical examinations. Consistent with their greater out-of-network coverage, PPOs were more likely to provide worldwide emergency/urgent care coverage. PPOs were also more likely than other CCPs to cover routine chiropractic care, although it was not commonly covered in either PPOs or CCPs. CCPs were more likely to provide health club/fitness classes and smoking cessation.

Among benefits not covered by Medicare, most PPOs provided a vision benefit (eye exams/wear), a majority offered a hearing benefit, and one third provided a dental benefit. But a lower proportion of PPOs than competing CCPs provided vision, hearing, and dental benefits. For example, less than one quarter of PPOs covered hearing aids, compared with more than half of competing CCPs. Offering richer benefits in addition to their out-of-network coverage does not appear to have been part of PPOs' strategy to attract enrollees. Instead, they might have restrained other benefits to keep down premiums or fund the costs associated with their out-of-network benefit. The aggressive implementation time frame may also have limited the range of benefits that could be developed for the new products.

PPOs provided less restrictive access to network physician specialists than CCPs. Seventy-two percent of competing CCPs required referrals for a specialist visit compared with only 10 percent of Demonstration PPOs. A large majority of the Demonstration plan sponsors offered a traditional PPO product, but three parent companies offered a point of service (POS) product (Aetna, Horizon, and PacifiCare). These three sponsors offering a POS varied in how they differed from the traditional PPO model. For instance, Aetna did not require a gatekeeper PCP, but the product also did not offer out-of-network options for all services in 2003 and 2004. PacifiCare of Nevada, on the other hand, for in-network services required the selection of a primary care physician and referrals. Horizon offered a POS with out-of-network coverage but with high cost sharing. Additionally, the Horizon plan retained many of the HMO referral requirements.

2.2.3 Cost Sharing

Table 2-10 reports typical (median) cost sharing for selected services for PPOs (in-network and out-of-network), competing CCPs, and original Medicare FFS. Office visit co-payments to see a primary care physician (PCP) or specialist did not vary substantially among the plans, but inpatient hospital co-payments and coinsurance rates for the out-of-network benefit did. A typical co-payment to see a PCP was $10, whereas a typical specialist office visit co-payment was $20. Some plans charged a flat co-payment for a hospital stay, regardless of the length of stay (Advantage, Aetna, Cariten, and OSF). Other plans charged a daily co-payment only for up to 5 days. Four plans had at least one plan option in 2004 with no inpatient hospital co-payment (Group Health, PacifiCare, Tenet, UPMC), but they offset this benefit by charging a higher premium or not offering a drug benefit.

Table 2-10
Cost sharing in PPOs, competing coordinated care plans,
and Medicare fee-for-service typical (median) co-payment ($),
coinsurance (%), or deductible ($) for selected services, 2004
Service	PPO		CCP	FFS
Service	In-network	Out-of-network	CCP	FFS
NOTES: Includes "Part A and Part B" plans only. Employer-only plans are excluded. PPO is PPO Demonstration plans. Competing CCP plans are defined by those offered in at least one PPO service area county. FFS is original Medicare fee-for-service. ¹ Co-payments per day are often limited to the first days of a stay, for example, the first 5 days. Co-payments may vary for different days of a stay. ² For FFS, this refers to initial deductible per benefit period. Beyond day 60, additional cost sharing applies. ³ Co-payments vary across outpatient services. For CCPs, the median minimum co-payment is $50 and the median maximum co-payment is $100. ⁴ 30-day supply at designated retail pharmacy. SOURCE: RTI analysis of CMS HPMS April 2004 file.
Primary care physician visit
Co-payment	$10	Rare	$10	—
Coinsurance	—	20%	—	20%
Specialist physician visit
Co-payment	$20	Rare	$20	—
Coinsurance	—	20%	—	20%
Hospital inpatient stay
Co-payment per day¹	$100	Rare	$175	—
Co-payment per stay²	$250	$750	$250	$876
Coinsurance	Rare	20%	Rare	—
No cost sharing (% of plans)	13%	0%	19%	—
Hospital outpatient
Co-payment per visit³	$50	Rare	$50–100	—
Coinsurance	10%	20%	20%	20%
No cost sharing (% of plans)	33%	0%	29%	—
Global deductible	Rare	$250	Rare	$110
Global deductible	Rare	$250	Rare	(Part B)
Prescription drugs⁴
Generic-only drug tiers	$10	—	$10	—
Some or all brand drug tiers	$37.50	—	$30	—

Thirty-nine percent of Demonstration plans had in-network out-of-pocket maximums whereas only 23 percent had out-of-network out-of-pocket maximums. Among PPOs that had a maximum, the in-network out-of-pocket maximum was typically about $1,800. When there was an out-of-network out-of-pocket maximum, it was typically about $3,250. A smaller percentage of competing CCPs than PPOs offered an in-network out-of-pocket maximum (30 versus 39 percent), which was typically somewhat greater ($2,560 versus $1,800). Some PPOs were concerned that adding catastrophic coverage (an out-of-pocket maximum) might result in adverse selection, attracting beneficiaries with high expected medical expenses (Greenwald et al., 2004).

Figure 2-4 shows the implications of these cost sharing rules for total beneficiary out-of-pocket expenses. In addition to cost sharing (which includes expenses for uncovered services), the total out-of-pocket costs shown in Figure 2-4 include premiums (Part B and health plan) and prescription drug expenses (assuming no drug coverage beyond what is offered by the health plan). For all types of plans, in-network cost-sharing levels were assumed to determine total out-of-pocket costs. In our analysis, we summarized total out-of-pocket costs by plan type for enrollees in excellent, good, and poor health, aged 70 to 74.⁶ Plan types are Demonstration PPOs, competing CCPs/HMOs, original Medicare FFS, and original Medicare plus competing Medigap Plan F. Plan type costs are unweighted averages across plans of a given type; for example, an average of the 61 PPO Demonstration open-access plans.

Below is a bar graph, click here for the text describing this graph.

Figure 2-4
Predicted out-of-pocket cost by plan type: Beneficiaries aged 70–74

Figure 2-4

NOTE: PPO is PPO Demonstration plan. MGAP is competing Medigap Plan F. CCP is competing coordinated care plan. Rx is prescription drug costs. Total Premiums includes health plan and Medicare Part B premiums. Cost sharing includes costs for noncovered services. Assumes in-network cost-sharing levels. Plan type costs are unweighted averages across plans of a given type. Excludes institutionalized beneficiaries.
SOURCE: RTI analysis of CMS 2004 out-of-pocket cost data.

As shown in Figure 2-4, a beneficiary was predicted to have higher out-of-pocket costs in a Demonstration PPO than in a competing CCP at each health status level, due to the higher PPO premium. This was true even if no out-of-network providers were patronized. But the difference between PPOs and CCPs narrowed as health worsened because of lower PPO cost sharing for inpatient services. PPOs, of course, offer an out-of-network benefit that HMOs lack, which is a reason for the higher PPO premium.

PPOs (and CCPs) occupied an intermediate position between FFS and Medigap in terms of out-of-pocket costs and risk protection. PPOs were less expensive than Medigap F for beneficiaries in excellent and good health status, but more expensive for beneficiaries in poor health status. PPO premiums and drug costs were lower than Medigap's at each health status level, but cost sharing was higher and grew more rapidly, even if only in-network providers were used. On the other hand, PPOs were more expensive than FFS for excellent and good health statuses, but less expensive for poor health status. PPO premiums were always higher, but drug costs and cost sharing were lower and grew less rapidly as health and utilization worsened, gradually offsetting higher PPO premiums. PPOs exposed enrollees to more financial risk than did Medigap F (a difference in total out-of-pocket costs between excellent and poor health statuses of $310 versus $265 for Medigap), but less than FFS alone ($310 versus $472).

2.3 Utilization Review, Case Management, and Other Cost Management Techniques

Table 2-11 summarizes various cost management techniques employed by the Demonstration plans, collected from our 2003 case study visit interviews and from plan benefit documents. Many of the Demonstration plans asked for notification of use of out-of-network services. The plans tended to use this notification in an attempt to "know what's going on" with enrollees rather than to try to steer them back into the provider network. On the other hand, plans used this as an opportunity to remind enrollees of the out-of-network cost sharing. A number of plans required prior authorization for some services, most commonly inpatient hospitalizations. In such cases, a beneficiary who did not pre-authorize could be charged a penalty. For example, PacifiCare charged $500 for unauthorized admissions, in addition to any other applicable cost sharing. One plan, Horizon, had retained its HMO referral requirements in the Demonstration POS product.

Table 2-11
Cost management techniques employed in PPO Demonstration plans
Demonstration plan	Preauthorization required	Prenotification requested	Disease management/ case management	Physician profiling
SOURCE: RTI 2003 site visits with plans; plan benefit documents.
Advantage		●	●	●
Aetna	●		●	●
Cariten			●	●
Coventry
St. Louis			●	●
Pennsylvania			●	●
West Virginia	●		●	●
Group Health, Inc.			●
HealthFirst	●
Health Net		●	●	●
HealthNow	●		●
HealthSpring			●
Horizon	●		●	●
Humana	●		●
OSF Health Plans	●		●	●
PacifiCare	●		●
Tenet Choices, Inc.			●	●
United HealthCare		●	●	●
UPMC	●		●	●

All participating plans instituted some degree of utilization review and/or case management for their Demonstration products. In general, the PPO plans tended to use the same basic utilization management/case management protocols found in their other managed care products. Many of the Demonstration plans used an initial assessment visit to assess enrollee health needs and identify individuals who should be targeted for disease management. Disease management protocols tended to focus on the high cost, high prevalence diseases among the Medicare population: diabetes, congestive heart failure, and chronic obstructive pulmonary disease. Some of the plans conducted these programs internally, whereas others hired independent disease management vendors.

2.4 Enrollment

This section documents PPO Demonstration plans' enrollment of Medicare beneficiaries. Section 2.4.1 discusses trends in Demonstration enrollment, overall and by individual contract. Section 2.4.2 analyzes the market share of Demonstration plans in their service areas. Section 2.4.3 discusses the prior enrollment status (FFS versus HMO) of Demonstration enrollees.

2.4.1 Enrollment Trends, Overall and by Contract

Table 2-12 reports enrollment in each PPO Demonstration contract from 2003 to 2005, and in successor contracts (where applicable) in 2006 and 2007. Enrollment is for July of each year, except for 2007, when it is for February. Enrollment in "employer only" plans is included.⁷ Figure 2-5 graphs Demonstration enrollment from 2003 to 2005, total, and for the Horizon New Jersey contract versus all other Demonstration contracts (non-Horizon).

Table 2-12
Enrollment in PPO Demonstration or successor contracts, 2003–2007
Contract ID	Legal name	State(s) of contract service area	Number of enrollees
			PPO Demonstration contracts			Successor contracts
			2003	2004	2005	2006	2007
¹ In 2006, PPO Demonstration contract was consolidated into another Medicare Advantage contract already in existence (see *Table 2-3*). Enrollment for 2006 and 2007 represents combined enrollment of these two contracts. NOTES: Enrollment is for July of each year, except for 2007, when it is for February. Includes enrollment in both open-access and employer-only Demonstration plans and enrollees residing in all locations. SOURCE: RTI Analysis of 2003–2005 Medicare Enrollment Database.
—	Total	—	70,653	104,361	124,073	—	—
H3109	Horizon Healthcare of NJ	New Jersey	47,592	49,289	50,144	—	—
—	All non-Horizon contracts	—	23,061	55,072	73,929	—	—

H0102	UNITED HEALTHCARE INSURANCE COMPANY	Alabama	337	689	494	472	421
H0103	UNITED HEALTHCARE INS. COMPANY, INC.	Alabama	38	21	20	Not offered	Not offered
H0313	PACIFICARE OF ARIZONA	Arizona	346	758	716	Not offered	Not offered
H0314	HEALTH NET LIFE INSURANCE COMPANY	Arizona	853	2,377	Not offered	Not offered	Not offered
H0706	AETNA HEALTH INC.	New York	Not offered	181	282	683	677
H1047	HUMANA INSURANCE COMPANY	Florida	44	1,087	1,562	8,257¹	1,337¹
H1408	OSF HEALTHPLANS, INC.	Illinois	1,484	4,919	7,313	9,483	10,282
H1412	COVENTRY HEALTH AND LIFE INS. COMPANY	Illinois, Missouri	312	569	581	175	Not offered
H1413	UNITED HEALTHCARE INS. COMPANY, INC.	Illinois, Missouri	1,609	3,139	3,203	4,286	4,237
H1508	ADVANTAGE HEALTH SOLUTIONS, INC.	Indiana	88	736	1,555	2,147	2,013
H1715	COVENTRY HEALTH AND LIFE INS. COMPANY	Kansas, Missouri	Not offered	52	768	4,224	7,003
H1805	ANTHEM HEALTH PLANS OF KENTUCKY	Kentucky	Not offered	55	160	675	835
H1901	TENET CHOICES, INC.	Louisiana	225	396	449	579	579
H2110	AETNA HEALTH INC.	Maryland	2,741	3,635	4,516	4,404	3,928
H2903	PACIFICARE OF NEVADA, INC.	Nevada	28	121	113	Not offered	Not offered
H3108	AETNA HEALTH INC.	New Jersey	5,891	6,598	6,515	7,703	7,966
H3109	HORIZON HEALTHCARE OF NJ,INC.	New Jersey	47,592	49,289	50,144	58,543¹	55,348¹
H3323	GROUP HEALTH INCORPORATED	New York	994	4,169	5,048	6,782	7,954
H3324	HEALTHNOW NEW YORK, INC.	New York	148	1,753	3,692	10,052	10,704
H3325	MANAGED HEALTH INC.	New York	26	83	40	25	Not offered
H3326	UNITED HEALTHCARE INS. COMPANY OF NY, INC.	New York	65	830	849	721	Not offered
H3403	UNITED HEALTHCARE INS. COMPANY, INC.	North Carolina	702	2,658	2,917	2,627	2,401
H3615	COVENTRY HEALTH AND LIFE INS. COMPANY	Ohio, West Virginia	20	3,168	6,053	7,130	5,440
H3616	UNITED HEALTHCARE INS. COMPANY, INC.	Ohio	513	951	807	84	Not offered
H3617	UNITED HEALTHCARE INS. COMPANY, INC.	Ohio	76	190	168	Not offered	Not offered
H3618	COMMUNITY INSURANCE COMPANY	Ohio	Not offered	344	809	3,958	5,323
H3806	HEALTH NET LIFE INSURANCE COMPANY	Oregon, Washington	165	5,180	14,726	18,879	18,561
H3913	UPMC HEALTH BENEFITS, INC.	Pennsylvania	274	467	756	1,213	850
H3914	AETNA HEALTH INC.	Pennsylvania	2,676	3,707	4,190	6,052	6,322
H3915	HEALTH ASSURANCE PENNSYLVANIA, INC.	Pennsylvania	1	11	38	2,803	2,682
H4103	UNITED HEALTHCARE INS. COMPANY, INC.	Rhode Island	216	687	539	489	473
H4403	CARITEN INSURANCE COMPANY	Tennessee	17	26	40	54	45
H4404	HEALTHSPRING, INC.	Tennessee	364	624	606	461	Not offered
H5400	UNITED HEALTHCARE INS. COMPANY, INC.	Florida	435	1,749	1,859	Not offered	Not offered
H5401	UNITED HEALTHCARE INSURANCE COMPANY	Florida	2,373	3,142	2,545	1,740	1,479

Below is a line graph, click here for the text describing this graph.

Figure 2-5
Enrollment in the PPO Demonstration, 2003–2005

Figure 2-5

NOTES: Includes enrollees in both open-access and employer-specific plans, and enrollees residing in all locations. Enrollment is as of July of each year. "Horizon" is the Horizon Healthcare of New Jersey Demonstration contract. "Non-Horizon" is all other Demonstration contracts.
SOURCE: RTI analysis of Medicare's Enrollment Database, 2003–2005.

Initial enrollment (2003) in the Demonstration was dominated by one plan, Horizon Healthcare of New Jersey, contract number H3109. Nearly 50,000 of the approximately 70,000 2003 Demonstration enrollees were in Horizon. Almost all of the initial Horizon enrollees transferred from a 2002 Horizon HMO product that was discontinued effective contract year 2003. For this reason, initial Horizon Demonstration enrollment is more a continuation of the earlier HMO enrollment than it is new enrollment attracted to a PPO product.

Although much smaller than Horizon, several other Demonstration plans attracted significant 2003 enrollment. The second-largest enrollment plan, also in New Jersey and also with antecedents in a discontinued product, was Aetna contract H3108. This plan enrolled nearly 6,000 beneficiaries in 2003, and 5 other contracts enrolled more than 1,000 beneficiaries. The other 24 Demonstration contracts available in July 2003 had relatively modest initial enrollments of less than 1,000.

Horizon enrollment was relatively flat throughout the Demonstration (2004 and 2005), but non-Horizon enrollment rose steadily, if not dramatically (Figure 2-5). In 2004, total Demonstration enrollment exceeded 100,000 beneficiaries, more than half of whom were enrolled in non-Horizon contracts. In 2005, the Demonstration's last year, enrollment was nearly 125,000, with more than 70,000 of the total in non-Horizon plans.

By 2005:

Two contracts enrolled more than 10,000 beneficiaries (Horizon in New Jersey, Health Net in Oregon/Washington);
Four contracts enrolled between 5,000 and 10,000 beneficiaries (OSF in Illinois, Aetna in New Jersey, Group Health Incorporated in New York, Coventry in Ohio/West Virginia);
Six contracts enrolled between 2,500 and 5,000 (United in Illinois/Missouri, Aetna in Maryland, HealthNow in New York, United in North Carolina, Aetna in Pennsylvania, United in Florida);
Three contracts enrolled between 1,000 and 2,500 (Humana in Florida, Advantage in Indiana, United in Florida); and
The remaining 19 contracts offered in 2005 enrolled fewer than 1,000 beneficiaries.

Several of the Demonstration contracts with large 2005 enrollments began the Demonstration in 2003 with large enrollments, drawing from discontinued predecessor plans of the same plan sponsor (for example, Horizon in New Jersey, and Aetna in New Jersey and Maryland). Other Demonstration contracts with large 2005 enrollments began with limited enrollment in 2003, but grew rapidly. Contracts that experienced rapid enrollment growth over the Demonstration period included Health Net in Oregon/Washington, OSF in Illinois, Group Health Incorporated in New York, and Coventry in Ohio/West Virginia. There is no obvious similarity in these "successful" contracts, as they ranged from the West Coast to the East Coast, from urban to rural service areas, from large national/regional MA plan sponsors to small local plan sponsors, and from plans facing little MA competition in their service area to plans facing substantial competition. All of these contracts did offer a traditional PPO product as opposed to a POS product.⁸

Total enrollment is not comparable for post-Demonstration in 2006 and 2007 because some contracts were not offered after 2005 and others were combined with preexisting plans with much larger enrollment. Some contracts that had been growing during the Demonstration period continued to gain enrollment in 2006 and 2007; for example, OSF in Illinois, Group Health in New York, HealthNow in New York, and Health Net in Oregon and Washington. Other contracts that were not gaining enrollment during the Demonstration grew significantly for the first time post-Demonstration, including Coventry in Kansas and Missouri, Community Insurance (Anthem) in Ohio, and Health Assurance (Coventry) in Pennsylvania. Still other contracts did not participate in the overall growth of MA enrollment in 2006 and 2007, experienced relatively stable enrollment in those years. Only 5 of the original 10 United Healthcare Demonstration contracts were still offered in 2007.

2.4.2 Market Share

As of March 2004, Demonstration PPOs' enrollment share of all Medicare beneficiaries in their combined service areas was 1.0 percent, and their share of total Medicare health plan enrollment in their service areas was 4.6 percent (Table 2-13). Only four Demonstration contracts had more than a 1 percent enrollment market share in their service areas: Horizon in New Jersey (4.1 percent), Coventry in Ohio/West Virginia (3.5 percent), OSF in Illinois (2.8 percent), and Aetna in Maryland (1.1 percent). Horizon had a slightly larger market share than competing coordinated care plans in its service area, and OSF and Aetna in Maryland had about the same market shares as competing CCPs. In all other Demonstration market areas, the CCP market share was at least several times the PPO market share, and in most cases it was much larger. FFS had more than half the Medicare enrollees in all service areas. Despite the growth in Demonstration enrollment from 2003 to 2005, it is clear from the low market share numbers that Demonstration PPOs were the preferred choice of only a small minority of Medicare beneficiaries during the Demonstration period.

Table 2-13
Market share by plan type, by PPO service area, 2004 (in descending order of PPO market share)
Contract	Organization name	State	Service area market share (%)
Contract	Organization name	State	Demo PPO	CCP	Other plan	Other demo PPO	FFS
NOTE: Includes beneficiaries with Part A and Part B coverage as of March 2004. Market shares for each Demonstration PPO calculated for beneficiaries residing in that contract's service area counties. CCP is coordinated care plan. FFS is fee-for-service. SOURCE: RTI analysis of the March 28, 2004 Medicare Enrollment Database.
TOTAL	—	—	1.01	19.24	1.79	NA	77.96
H3109	HORIZON HEALTHCARE OF NJ,INC.	New Jersey	4.12	3.10	0.10	0.56	92.11
H3615	COVENTRY HEALTH AND LIFE INS. COMPANY	Ohio, West Virginia	3.52	15.95	3.94	0.07	76.52
H1408	OSF HEALTHPLANS, INC.	Illinois	2.77	2.97	0.52	0.00	93.74
H2110	AETNA HEALTH INC.	Maryland	1.08	1.03	2.24	0.00	95.65
H3108	AETNA HEALTH INC.	New Jersey	0.84	3.05	0.13	3.40	92.59
H3914	AETNA HEALTH INC.	Pennsylvania	0.82	24.53	3.18	0.01	71.46
H1413	UNITED HEALTHCARE INS. COMPANY, INC.	Illinois, Missouri	0.81	21.70	0.74	0.15	76.59
H3806	HEALTH NET LIFE INSURANCE COMPANY	Oregon, Washington	0.72	32.07	10.00	0.00	57.20
H5401	UNITED HEALTHCARE INSURANCE COMPANY	Florida	0.53	18.59	0.31	0.07	80.50
H4103	UNITED HEALTHCARE INS. COMPANY, INC.	Rhode Island	0.53	37.32	0.02	0.00	62.14
H0102	UNITED HEALTHCARE INSURANCE COMPANY	Alabama	0.45	22.78	0.92	0.00	75.85
H3403	UNITED HEALTHCARE INS. COMPANY, INC.	North Carolina	0.44	10.88	0.05	0.00	88.63
H3324	HEALTHNOW NEW YORK, INC.	New York	0.38	24.74	0.04	0.00	74.84
H3323	GROUP HEALTH INCORPORATED	New York	0.35	23.25	1.93	0.03	74.43
H3616	UNITED HEALTHCARE INS. COMPANY, INC.	Ohio	0.31	15.17	0.16	0.07	84.29
H0314	HEALTH NET LIFE INSURANCE COMPANY	Arizona	0.30	31.67	0.45	0.10	67.48
H5400	UNITED HEALTHCARE INS. COMPANY, INC.	Florida	0.30	34.78	0.06	0.01	64.86
H4404	HEALTHSPRING, INC.	Tennessee	0.27	12.48	0.18	0.00	87.06
H1901	TENET CHOICES, INC.	Louisiana	0.21	32.49	0.13	0.00	67.18
H1508	ADVANTAGE HEALTH SOLUTIONS, INC.	Indiana	0.18	2.98	0.19	0.00	96.65
H1412	COVENTRY HEALTH AND LIFE INS. COMPANY	Illinois, Missouri	0.15	21.70	0.74	0.81	76.59
H0313	PACIFICARE OF ARIZONA	Arizona	0.12	36.31	0.41	0.26	62.91
H1805	ANTHEM HEALTH PLANS OF KENTUCKY, INC.	Kentucky	0.10	11.00	0.41	0.00	88.49
H1047	HUMANA INSURANCE COMPANY	Florida	0.09	23.42	0.41	0.62	75.46
H0706	AETNA HEALTH INC.	New York	0.09	12.76	0.26	0.40	86.50
H3618	COMMUNITY INSURANCE COMPANY	Ohio	0.08	15.40	0.16	0.29	84.07
H3617	UNITED HEALTHCARE INS. COMPANY, INC.	Ohio	0.07	15.14	6.29	0.00	78.50
H3913	UPMC HEALTH BENEFITS, INC.	Pennsylvania	0.07	36.83	1.60	0.00	61.51
H3326	UNITED HEALTHCARE INS. COMPANY OF NY, INC.	New York	0.05	22.40	2.03	0.31	75.21
H2903	PACIFICARE OF NEVADA, INC.	Nevada	0.05	14.56	23.26	0.00	62.12
H0103	UNITED HEALTHCARE INS. COMPANY, INC.	Alabama	0.04	19.58	0.25	0.00	80.14
H4403	CARITEN INSURANCE COMPANY	Tennessee	0.01	12.64	0.34	0.00	87.01
H1715	COVENTRY HEALTH AND LIFE INS. COMPANY	Kansas, Missouri	0.01	18.22	0.30	0.00	81.48
H3325	MANAGED HEALTH INC.	New York	0.01	25.33	2.26	0.36	72.04
H3915	HEALTH ASSURANCE PENNSYLVANIA, INC.	Pennsylvania	0.00	39.94	1.38	0.05	58.63

2.4.3 Prior Enrollment Status

Aside from tracking overall enrollment trends, we were interested in learning whether Medicare PPO Demonstration enrollees came from FFS—suggesting that PPOs may in fact be a more attractive managed care option to beneficiaries historically wary of leaving FFS. Or, were PPO enrollees primarily switchers from other CCPs/HMOs? Table 2-14 presents the prior enrollment status of March 2004 enrollees in PPO Demonstrations and recent enrollees in competing CCPs in Demonstration service areas. Enrollees in the Horizon PPO Demonstration contract who were previously enrolled in the Horizon HMO are excluded from these data.

Table 2-14
Prior enrollment status of PPO and recent CCP enrollees, 2004
Prior enrollment	Current enrollment
Prior enrollment	PPO	CCP
NOTES: Includes beneficiaries with Part A and Part B coverage as of March 2004, residing in the open enrollment service area counties of any PPO Demonstration contract. Includes beneficiaries who enrolled in their current plan January 1, 2003 or after. Excludes Horizon PPO Demonstration enrollees previously enrolled in the Horizon HMO. ¹ Beneficiaries who newly enrolled in the Medicare program January 2003 or after. ² Prior plan has a different parent company than the current plan. ³ Prior plan has the same parent company as the current plan. SOURCE: RTI analysis of the March 28, 2004 Medicare Enrollment Database.
Recent Medicare enrollee¹	14.8%	23.2%
Fee-for-service Medicare	41.9	39.6
Medicare health plan	43.4	37.3
Unaffiliated²	27.8	—
Affiliated³	15.5	—

There was some expectation that PPOs would be more attractive to FFS beneficiaries than were other CCPs (mostly HMOs) because of PPOs' greater freedom of provider choice. But PPOs drew about the same proportion of their enrollees from FFS as CCPs (42 versus 40 percent—see Table 2-14). Also, compared with CCPs, PPOs drew a somewhat lower proportion of their enrollees (15 versus 23 percent) from recent Medicare program enrollees (beneficiaries new to the Medicare program during the Demonstration period), which is inconsistent with the hypothesis that PPOs were especially attractive to Medicare "age ins"—those joining the program when they become eligible at age 65.

Among the 43 percent of PPO enrollees previously in Medicare health plans, about two thirds (64 percent) were previously enrolled in unaffiliated plans and about one third (36 percent) were previously enrolled in affiliated plans. An affiliated plan was a plan (typically an HMO) offered in the same market area by the same parent company that was sponsoring the Demonstration PPO; for example, a United Healthcare Medicare HMO offered in the same service area as the United Demonstration PPO. Thus, the Demonstration PPOs were not simply siphoning affiliated HMO enrollment—only about 15 percent of total PPO enrollment came from this source.⁹

SECTION 3
FINDINGS FROM CASE STUDY INTERVIEWS OF PARTICIPATING ORGANIZATIONS

This section provides an overview of the early Demonstration experience from the perspective of participating organizations (plans), summarizing the key themes of the Demonstration's implementation and operations through mid 2003. We conducted site visits with each of the Demonstration parent company organizations to discuss in more detail the characteristics of their PPO Demonstration products and their experiences in implementing their plans.¹⁰ Site visits were conducted between April and July 2003. The plan personnel we interviewed varied across sites but typically included the Medicare product manager, government relations specialist, actuarial/financial personnel, marketing directors, and top management. We also spoke with representatives from the CMS Central Office and CMS Regional Offices regarding their perspectives on the implementation of this Demonstration.

3.1 Factors Influencing Organizations' Participation in the Demonstration

Table 3-1 summarizes the reasons reported by the Demonstration organizations for offering a PPO/POS product through the Demonstration. The most common reasons organizations cited for offering a PPO under this Demonstration were as follows:

Table 3-1
Organizations' reasons for joining the PPO Demonstration
Demonstration organization	Availability of risk sharing	Higher county payment rates	Streamlined administrative process	$100,000 start-up funding	PPO product appeal	CMS exposure on the PPO Demonstration	Other reasons
NOTES: In this table, we note factors that were important to organizations in two ways. Factors noted by organizations as especially important are identified as "very important." Other factors noted are indicated by "●" SOURCE: RTI 2003 site visits with plans.
Advantage	Very Important			●	●		Large employer approached the organization
Aetna	●	●	●	●
Cariten				●	●
Coventry	Very important	Very important	●	●
Group Health, Inc.	●	●		●	●
HealthFirst	●			●	●
Health Net	Very important	Very important	●	●	●	●
HealthNow					●		PPO an opportunity to increase Medicare business
HealthSpring				●			PPO an opportunity to increase Medicare business
Horizon		Very important	●	●			Waiver of FFS cost sharing Limits
Humana	●	●	●				Ability to test PPOs on a limited basis
OSF Health Plans		Very important	●	●
PacifiCare	●	●		●	●		PPO an opportunity to increase Medicare products
Tenet Choices, Inc.		Very important			●	●
United HealthCare	Very important	●	●	●	●
UPMC		●			●		PPO an opportunity to increase Medicare products

Higher County Payment Rates: 11 of 16 organizations were attracted to the Demonstration by the higher county payment rates (based on 99 percent of FFS rather than the existing rate book), with 5 rating it a very important reason for their participation in the Demonstration. Organizations also told us that the prospect for payment growth equal to FFS cost growth (rather than the 2 percent minimum growth rates for Medicare health plans in effect at the start of the Demonstration) was as important as higher initial rates. Some organizations, however, were disappointed that the actual Demonstration payment rates in the first year were not as much above the standard Medicare rates as they had expected.
Availability of Risk Sharing: The ability to share risk with CMS was mentioned by 9 organizations as an appealing feature of the Demonstration, with 4 rating it as very important for their participation decision. This option allowed organizations to try out a new Medicare option without bearing the full financial risk, a particular concern because of PPOs' coverage of out-of-network providers.

These two payment enhancements (higher capitation rates and risk sharing) were the only reasons for Demonstration participation rated as very important by any of the organizations.

$100,000 Start-up Funding: 12 Demonstration organizations mentioned the $100,000 in start-up funding offered by CMS, not available under the regular program, as a factor in their decision to join the Demonstration.
PPO Product Appeal: Ten organizations reported that they found the PPO an appealing product in the current Medicare marketplace. Many noted that the PPO/POS option is a complement to, not competitor for, HMO products.
Streamlined Administrative Process: The Demonstration application process was streamlined relative to the regular health plan contract application, an attractive feature to 7 organizations. In particular, under the Demonstration organizations did not have to file an adjusted community rate (ACR) proposal.
CMS Exposure: 2 Demonstration organizations joined, in part, because they expected to take advantage of CMS-sponsored visibility for the project.

Other reasons for Demonstration participation mentioned by one or more organizations included desire to offer a full range of Medicare products to attract as many Medicare enrollees as possible, desire to test the PPO model in Medicare, greater actuarial freedom in setting cost sharing, PPOs being part of an organizational strategic direction to increase Medicare business, getting in the market first with a new product and distinguishing our organization from competitors, and a large employer was interested in Medicare PPOs for its retirees and contacted the plan sponsor.

During the site visits, we learned that most Demonstration organizations were able to achieve the January 1, 2003, start date only because of their past experience with the Medicare health plans program. They chose service areas for the Demonstration that relied on existing provider networks and the ability to enact contract amendments (rather than establish new provider contracts). These organizations told us that it would have been very difficult to offer the PPO product in this Demonstration without past experience working with CMS and an existing provider network. Also, many private sector PPOs do not bear risk, and even with CMS sharing risk, they would have to be licensed as risk-bearing entities by their states. The very aggressive time frames for the Demonstration implementation—less than 1 year between Demonstration application due date and first enrollments—may have all but eliminated the chance that organizations new to Medicare would participate in the Demonstration.

Although the PPO Demonstration did not attract many new organizations to Medicare, the features of the Demonstration—primarily the favorable payment rates—did keep some organizations in the Demonstration when they might otherwise have left Medicare altogether. The best example of this was Horizon in New Jersey, which substituted its existing HMO product with the new PPO Demonstration product because of the higher payment rates available under the Demonstration, and because Horizon was finding it increasingly difficult to offer its former product under the regular health plans program. Horizon had, by far, the largest enrollment of the PPO Demonstration with more than 45,000 enrollees during the Demonstration's first year. In the absence of the PPO Demonstration, many of these beneficiaries may have been left with no Medicare health plan option, because Horizon was the only remaining statewide Medicare managed care option at the time.

In addition, some organizations either reentered Medicare markets from which they had previously withdrawn (for example, Aetna in Maryland) or expanded into areas where they did not offer a plan (for example, Health Net had contracts in Arizona but expanded Medicare business into its Oregon organization). In these cases, expanded choices were offered to Medicare beneficiaries, although not necessarily by new organizations.

We also found that, aside from the more favorable financial arrangements offered under the Demonstration, a number of organizations were genuinely interested in testing out the PPO product for Medicare. We asked why they had not done so under the regular program—PPOs had been allowed since the Balanced Budget Act of 1997 (BBA) was implemented. The organizations told us that, although allowed under BBA rules, it was hard to commit to a new product at a time when the decision to continue to participate with Medicare was made year to year based on the new payment rates. So although many organizations liked the PPO model and perceived it as either the "up and coming product" for Medicare or as a logical companion to HMO products, financial considerations simply made new Medicare products difficult to launch. The Demonstration, with potentially more favorable payment rates and the availability of risk sharing, gave plans the opportunity to try out a PPO product.

Related to understanding why PPOs have not been more widely offered in the Medicare program prior to the Demonstration, we asked organizations why PPOs have not dominated the market in Medicare the way they have in the commercial market. A number of organizations told us that employers, faced with both rapidly rising health care costs and with some backlash from employees about traditional HMO coverage, find PPO options something of a middle ground. PPO options tend to cost less than traditional indemnity insurance plans, which are increasingly rare. Under most employer contracts, choices are often limited to one or two options, so minimum enrollments are more likely to be achieved for organizations bidding for employer contracts. Under Medicare, however, most organizations told us that they struggled year to year with difficult financial decisions in participating with Medicare at all. They argued that under difficult financial constraints, they would be unlikely to offer a new, more risky Medicare product. The PPO Demonstration, however, provided many of them the opportunity to experiment with PPOs and Medicare under more favorable payment rates and with the security of risk sharing with CMS.

Finally, we asked the Demonstration organizations about their willingness to offer the PPO, or any managed care product, in traditionally underserved rural areas. The Demonstration organizations cited several factors that limited the viability of Medicare health plans, especially those requiring contracted provider networks, in rural areas. CMS currently maintains the same provider network requirements for rural areas as for urban areas. However, meeting these requirements in many rural areas can be nearly impossible, the organizations told us. The supply of physicians and hospitals, as well as other providers, can be very limited in rural areas. Some organizations told us that they could not meet CMS network requirements even if they signed up every provider in an area. Because of this relative scarcity, rural providers have market power and sometimes require more than 100 percent of FFS Medicare payments in order to participate in managed care networks (we heard figures of up to 110 percent). Some organizations noted that rural providers were "not organized," and that contracting with individual providers in rural areas was expensive and labor intensive. Also, it is difficult to achieve the necessary scale for plan viability of 5,000 to 10,000 enrollees in rural areas. As a result, many organizations view rural areas as less attractive potential markets.

3.2 Product Pricing and Design

Many organizations wanted their Demonstration plans to be competitive with Medigap alternatives and priced them accordingly (i.e., at prices below competing Medigap products). Initial premiums were often set conservatively (i.e., high) because of the lack of historical utilization data on Medicare PPOs. Other considerations in setting premiums were to

be competitive with other Medicare health plans offered in the same market (including other Demonstration plans in markets where such plans were offered by more than one participating organization);
pay for PPO enrollees' out-of-network benefit;
account for the higher utilization of PPO enrollees, because utilization is less tightly controlled in PPOs than in HMOs;
account for some anticipated adverse selection in enrollment;
account for variations in the Medicare county capitation rate; and
achieve profitability.

Cost sharing was set to be able to attain the desired premium (e.g., to underprice Medigap). Organizations typically set in- versus out-of-network cost sharing to give enrollees strong incentives to use in-network providers. Several organizations said that they wanted to make their cost the same whether the enrollee went in- or out-of-network. To do this, they raised the enrollee out-of-network cost sharing (as compared with in-network cost sharing) enough to offset the higher out-of-network provider payment rates.

Many organizations believed that drug benefits were felt to be necessary to attract enrollees to the PPOs from Medicare supplements, and to be competitive with other Medicare health plan offerings. However, several noted that beneficiaries may have drug coverage from other sources—a state drug program, a former employer, etc.—so that it also made sense to offer plans without drug coverage.

Plans did not have a uniform perspective on expected risk selection in PPOs. Some thought that sicker beneficiaries who wanted to use out-of-network providers would be attracted to PPOs, creating adverse selection. Others thought that beneficiaries who were healthy and wanted to travel would be especially attracted to the PPO out-of-network benefit. In general, organizations were more concerned about adverse selection from plans' drug benefits than from the PPO out-of-network benefit or plan type per se. As a result, some plans indicated that they kept the drug benefits similar to other Medicare health plan drug benefits to avoid adverse selection. Some organizations thought there might be adverse selection because of the inability to vary premiums by age or to medically underwrite applicants, but these practices are proscribed for all Medicare health plans, not just PPOs.

In general, plans did not expect Demonstration plan enrollees to heavily use out-of-network providers—maybe for 5 percent to 10 percent of their use—and early Demonstration experience was in line with this. Some organizations experienced implementation/operational problems with non-network providers not accepting Demonstration plan coverage for enrollees who attempted to use it. These early issues seemed to improve as time went on.

Offering employer group-only plans was attractive to some organizations because group rather than individual enrollment resulted in lower per member acquisition costs and a lesser possibility of adverse selection. Employer group-only plans offered under the Demonstration might be (1) contributory, in which the employer paid all or some of premium or enhanced the plan benefits; or (2) sponsored, in which the employer arranged to offer the plan, but did not contribute to it financially. The latter type of plan could still have a lower premium for the employer's beneficiaries because of its group discount or risk rating. Employer plans might have a standard drug benefit, or an experience-rated employer-specific drug rider. Some organizations also had substantial employer-sponsored enrollment in their open access plans.

3.3 Factors Influencing Beneficiary Enrollment in PPOs

For most organizations, with the exception of Horizon and Aetna (whose enrollments were high during the initial year of the Demonstration), Demonstration enrollment was slower than projected. Organizations did not seem discouraged by this at the time, nor did they feel the PPO was not a viable Medicare product. Instead, most organizations with disappointing enrollment concluded that Medicare beneficiaries simply needed time to "get used to a new product." Organizations argued that CMS's aggressive timeline for enrollments may have been unrealistic. Most organizations told us that any new product for Medicare beneficiaries needs time in the marketplace to establish positive word of mouth.

In addition, most organizations marketing the PPOs to Medicare beneficiaries found that getting beneficiaries to even consider changing insurance options was a major hurdle. This is particularly the case since media coverage of the PPO Demonstration, expected by some organizations to be greater, was almost nonexistent. As a result, beneficiaries were hearing about the new PPO options almost exclusively from the organizations offering the option. Organizations reported to us that many beneficiaries (not unreasonably) were wary of information they received from "someone trying to sell them something." At the time, however, no other prominent source of information on the PPO option existed. Our discussions with the CMS Regional Offices did not uncover any known educational efforts about PPOs. Beginning in 2004, the Medicare & You handbook did include some information about PPO options.

Price appeared to be the factor that most influenced Medicare beneficiaries' willingness to consider PPOs, a new Medicare option. The organizations we spoke with observed that many beneficiaries are willing to pay sometimes high Medigap premiums to maintain freedom of provider choice and access to all services without referral. That said, many organizations also noted that beneficiaries are sometimes willing to consider other options if they can save money, particularly among beneficiaries who may be feeling increased financial pressures. Beneficiaries have not seemed willing, so far, to pay higher premiums for a PPO/POS than for available HMOs or to switch from Medigap without substantial savings. The Demonstration organizations that were most successful tended to offer a well-priced PPO option.

Some organizations felt that their initial Demonstration PPO offerings were not sufficiently differentiated from their HMO product(s), and that this lack of differentiation might have hurt first-year PPO enrollment. For example, the PPO and HMO provider networks and benefits were the same, except for the PPO's out-of-network benefit. These organizations planned to better differentiate their PPO from their HMO products in the future. For example, the future PPO product might have a wider provider network or richer benefits than the HMO.

The Demonstration organizations reported that the PPO coverage of out-of-network providers did not attract large numbers of beneficiaries; the out-of-network benefit per se was not valued enough by most beneficiaries to draw in large enrollments or to command a large price premium over HMOs in the marketplace. It is also true that organizations required substantial out-of-network cost sharing, 20 percent or 30 percent coinsurance, often with a deductible and. a high (or no) out-of-pocket maximum. Although this serves to give enrollees strong incentives to use in-network providers, it meant that utilizing the PPO out-of-network benefit was very expensive for them, limiting its attractiveness. In our site visits, several organizations told us that beneficiary concerns over possibly high out-of-pocket costs, or uncertainty about what those costs might be—particularly if they decided to use out-of-network services—was a major reason beneficiaries cited for choosing not to enroll in the PPO (after initially considering the product). Because of this, several organizations told us that they planned to lower out-of-pocket costs for in- and out-of-network services to improve the appeal of the PPO. This may be why a number of organizations indicated an intent to add an out-of-pocket maximum, standard in commercial PPOs. Although many organizations used the availability of out-of-network coverage as a primary selling point of PPOs, beneficiaries seemed to focus on total monthly premiums.

3.4 Impediments for Medicare PPOs

A major challenge for Medicare PPOs was that their low enrollments limit their bargaining power with providers. The PPO model is based on insurers obtaining price discounts from providers in return for in-network designation. These discounts allow the sponsoring insurer to lower the premium charged to employers (in the private sector) or to Medicare beneficiaries (in Medicare). PPOs can thus be priced lower than traditional indemnity products with no networks (e.g., Medigap). But Medicare PPOs had such low enrollments that they had limited bargaining power to obtain discounts from providers—we learned that discounts were typically 5 to 10 percent or less. Inability to obtain discounts from providers made offering competitive benefits to attract enrollees to Medicare PPOs more difficult. In the private sector, in contrast, a large employer or insurer has considerable enrollment to use as leverage to bargain for lower rates from providers. Medicare PPOs suffer from a chicken and egg problem—they need larger enrollments to get provider discounts and offer a lower price to beneficiaries, but they need a lower price to get larger enrollments.

Another way in which PPOs could offer a lower price than Medigap is through managing beneficiary utilization to eliminate unnecessary or low-value care. But the organizations we interviewed did not expect significant savings relative to Medicare FFS through care management. Virtually all of the PPOs engaged in some combination of disease management, utilization review, physician profiling, or prior authorization. But the lack of PCP gatekeeping, the out-of-network benefit, little risk sharing with providers, and the sometimes broader PPO provider networks limited the ability to manage care in a PPO compared with an HMO. The PPO is more an "open access" model that focuses on beneficiary choice of provider, not the tight utilization controls, provider risk sharing, and narrow networks of the most efficient providers that may exist in HMOs. Even in HMOs, there is a general trend away from tight utilization management and putting providers at risk because this has not proven successful in the marketplace.

The competition faced by PPOs relative to other insurance options may also limit their widespread appeal to Medicare beneficiaries. It is difficult for PPOs to compete with subsidized employer-supplements (although some employers may decide to offer PPOs to their retirees as a cost control measure, and employer-sponsored retiree coverage is projected to erode over time). Lower income beneficiaries care mostly about price and do not value the PPO out-of-network benefit highly—they tend to prefer Medicare HMOs or may have Original Medicare only. This left Demonstration PPOs competing with Medigap insurers, which account for about one-third of the total Medicare market. But Medigap has certain advantages over PPOs. It offers complete freedom of choice like the PPO and has limited cost sharing for all providers, not just network providers. Limited provider discounts and savings from utilization management constrain how much PPOs can underprice Medigap. Also, Medigap is typically allowed to age-rate premiums whereas PPOs cannot, further eroding PPOs' price competitiveness for the younger elderly, a natural target market for PPOs.

3.5 Marketing of Medicare PPOs

Many of the participating organizations sought to market the PPO not as something "new" but as an expansion of their overall product line. Aetna and United HealthCare, for instance, were careful to not market the PPO as something "new and different," for fear that this message portrayed a temporary product that would eventually leave the market. Most organizations already offered a Medicare HMO product, and the PPO gave the organizations an opportunity to offer a higher-end plan with greater freedom and out-of-network opportunities.

For those organizations already offering a Medicare HMO product, the target population was typically younger beneficiaries with middle or higher incomes. PPOs were expected to appeal to a more affluent clientele than HMO enrollees, and to beneficiaries who were willing to accept some financial risk in the form of higher and less predictable out-of-pocket costs than in HMOs or in Medicare supplements.

Many organizations targeted beneficiaries who travel a lot or live as "snowbirds" for half the year and therefore value an out-of-network benefit. Group Health, for instance, recognized that many of its potential enrollees traveled to Florida and therefore had plans under way to offer a network of providers in southern Florida for its New York enrollees who traveled or lived there on a seasonal basis. Similarly, Health Net offered provider networks in Arizona and Oregon to snowbirds who lived in both service areas throughout the year.

Other organizations directly targeted Medigap purchasers to the PPO, particularly in markets where Medigap premiums were high and are potentially pricing beneficiaries out of that product. For instance, Advantage shared with us that beneficiaries coming to marketing meetings experienced rapidly rising Medigap premiums, and they saw this as their niche for offering cheaper premiums than the Medigap F and G products for comparable coverage. Likewise, Coventry priced its PPO in the St. Louis market at about one-third of the Medigap Plan F premium to attract Medigap "types." Although several of the Demonstration organizations conceded that it was difficult to get Medicare beneficiaries to give up their Medigap plans, organizations also saw an opportunity when beneficiaries were looking for cheaper alternatives. Due to rapidly rising health care costs, increasing numbers of Medicare beneficiaries faced the loss or reduction of retiree benefits. Medigap premiums for individually purchased policies rise as beneficiaries age. Because of these factors, a number of Demonstration organizations positioned their PPO/POS products to attract these price-sensitive beneficiaries.

A final target group many of the organizations marketed to were age-ins--those who aged into the Medicare population at age 65. This group is seen as having the largest long-term potential, largely because these younger groups were more familiar with the PPO concept and with managed care in general. Many of the organizations told us that they had a large commercial PPO business with a working-age population that began aging into Medicare, and the PPO product allowed these age-ins to continue having similar coverage upon Medicare eligibility.

As Table 3-2 shows, all organizations gave the product a marketing name for their advertising and information dissemination. There was considerable variation in whether the organizations used the acronym PPO or something else to market the product. Nine of the 16 organizations used the acronyms "PPO" or "POS" in their plan's marketing name. Most organizations told us that most Medicare beneficiaries did not have a good understanding in the abstract of the differences between the acronyms "HMOs" and "PPOs" or "POS," and that the organizations did not emphasize the acronyms in their marketing. In a few markets (e.g., New York City), some organizations felt that "PPO" was understood well enough to be emphasized in their marketing.

Table 3-2
Demonstration organizations' product marketing names
Demonstration organization	Trade name
SOURCE: RTI 2003 site visits with plans.
Advantage	Advantage Preferred Plus
Aetna	Aetna Golden Choice
Cariten	Cariten Senior Health PPO
Coventry
Pittsburg market	Health Assurance Advantra M+C PPO
West Virginia & Ohio area markets	Advantra PPO
St. Louis area market	GHP Advantra PPO
Group Health, Inc.	GHI Medicare Choice PPO
HealthFirst
Option 1	HealthFirst PPO Select Plan
Option 2	HealthFirst PPO Complete
Health Net	SeniorCare Options Plus
HealthNow	Traditional Blue Medicare PPO
HealthSpring	HealthSpring Medicare + Choice PPO Plan
Horizon
Option 1	Medicare Horizon Blue
Option 2	Medicare Horizon Blue Plus
Humana	Humana Gold PPO
OSF Health Plans	OSF Care Preferred
PacifiCare	Secure Horizons Medicare POS
Tenet Choices, Inc.	Health Care Select
United HealthCare	Medicare Complete Choice
UPMC	UPMC for Life PPO

All the organizations conducted advertising campaigns to some degree, recognizing that eligible beneficiaries are largely unaware of the new Medicare PPO option. Although the organizations differed in their focus of marketing techniques, the large majority of them at a minimum did direct mailings to eligible groups they had targeted as potential enrollees. United HealthCare, for instance, looked at specific demographics of the area—age and income—in placing direct mail. Relatively few organizations used television or radio advertising, although most ran newspaper ads at the onset of the Demonstration enrollment period to create a publicity "blitz" for the product.

The following were some of the key marketing themes among the organizations:

More freedom of choice to see providers and go out-of-network
No referrals or authorizations required
No requirement to choose a primary care physician
Lower premiums than most Medigap plans
Savings when seeing a doctor in-network
Easier and less paperwork than FFS or supplemental insurance
Prescription drug coverage (if applicable)

In sum, key marketing themes were "choice," "value," "freedom," and "convenience." Disease management, extensive provider networks, and out-of-pocket limits (for applicable plans) were marketing themes used to a lesser extent but were still considered by some organizations as important selling points when presenting the product.

The majority of Demonstration organizations reported that they relied primarily on direct mail with in-person follow-up to attract most of their enrollees. Organizations conceded that this approach can be very costly, but they also reported that it was necessary given all of the education they needed to do about Medicare managed care and PPOs specifically. Some plans also conducted seminars and other group sessions. These were less successful in some areas, however, since new restrictions on gathering attendee names and phone numbers have been put into place. In general, the organizations found that PPO marketing cost per member enrolled—"member acquisition cost"—was quite high.

3.6 Provider Networks and Reimbursement

Because of the quick timing involved in implementing the PPO Demonstration, the organizations relied on their existing provider networks in establishing their Demonstration service areas. The goal of most organizations was to either use already established provider networks (which may have included language to cover all the insurer's products) or, if necessary, employ provider contract amendments to add the Medicare PPO product. In most cases, the Demonstration organizations told us that they had few difficulties with this approach. However, in a few cases, organizations were not able to sign up either a specific physician group practice or a hospital system. The reasons for resistance among these provider organizations were most often related to unwillingness to participate with a Medicare product or accept the payment rates offered by the organization. In instances where Demonstration organizations attempted to add rural providers in expansion counties around the established network service areas, unwillingness to accept organization payment rates and/or participate in managed care were barriers. For example, Health Net of Arizona told us that they might have included the semi-rural area between their current PPO service areas around Tucson/Phoenix in southern Arizona and Flagstaff in northern Arizona, but providers in this area required reimbursement of more than Medicare FFS rates—not a feasible payment level for this Demonstration. One organization, PacifiCare, was unable to offer a PPO product in Southern California because of difficulties in recruiting a provider network for the product. PacifiCare's HMO provider network was accustomed to accepting capitated risk and engaging in extensive care management and was not very amenable to shifting to the more open PPO model.

Even in the longer run, several organizations saw HMOs and PPOs as complementary offerings that would be viable in the same types of markets, and planned to offer PPOs in areas where they also offered HMOs. Organizations could amortize the fixed costs of network development and contracting across multiple product types. Offering Medicare health plans—including PPOs—was most attractive in areas where competitive provider networks could be contracted at sufficient discounts relative to the Medicare county rate to make the product financially viable. These areas were generally urban areas where competition among providers could be used to obtain discounted payment rates from providers.

Almost universally among the Demonstration organizations, in-network providers were reimbursed at discounted Medicare FFS rates. For physicians, this translated into a discounted Medicare Fee Schedule amount. For hospitals, payments were generally based on either Medicare diagnosis-related group (DRGs) rates or a per diem amount. Very few organizations reported to us any risk-bearing payment arrangements among providers; in these few cases, only primary care physician groups were paid on an at-risk basis.

3.7 Overall Perceptions and Comments About PPOs

In general, the Demonstration organizations had a favorable view of the PPO product for Medicare, despite slower than expected enrollments. These organizations remained committed to the PPO Demonstration product and were willing to give it some time. Many told us that launching a new product for Medicare was often a slow process, as beneficiaries were often slow to respond to anything unfamiliar. A number of Demonstration organizations also told us that, in general, most Medicare beneficiaries had little interest in shopping around for new Medicare coverage. If they already had a Medigap policy and could continue to afford the premiums, getting beneficiaries to consider something else was a challenge. That said, with Medigap premiums rising as Medicare beneficiaries get older (a process sometimes referred to as higher age attained rate bands), many of the Demonstration organizations positioned their PPO/POS product to be attractive to beneficiaries who may need to make a switch for financial reasons.

To make the PPO product viable for organizations, we heard most often that payment rates were a critical factor in the success of the PPO (as with any of the Medicare managed care products). If the reimbursement rates for Medicare PPOs were not adequate, then organizations would not be able to offer products—particularly a more expensive product like the PPO. We also heard a number of times that encouraging increased participation from nonparticipating providers (e.g., noncontracting providers being willing to provide services to PPO members) was critical to the long-term success of the PPO product.

Finally, many Demonstration organizations told us that education was the key to success for the Medicare PPO product, and there had been limited education and exposure of PPOs by CMS. Organizations believed that beneficiary and provider understanding of PPOs, and how they differ from HMOs and other managed care products, could be very limited in some areas. Therefore, organizations stressed the importance of a more proactive education campaign by CMS for both beneficiaries and providers about PPOs in improving acceptance of PPOs on a wider scale.¹¹

SECTION 4
FINDINGS FROM BENEFICIARY SURVEYS

Beneficiary surveys were used to address the following important Demonstration evaluation questions:

the reasons that beneficiaries enrolled, or chose not to enroll, in a Demonstration PPO;
the characteristics of PPO Demonstration enrollees versus enrollees in FFS or MA plans;
the experiences, access to care, and satisfaction of PPO Demonstration versus FFS and HMO enrollees; and
the reasons for disenrollment given by PPO disenrollees versus other plan type disenrollees.

To examine beneficiaries' rationale for enrollment (or non-enrollment), RTI conducted a survey of beneficiaries enrolled in the Demonstration plans and a comparison group of beneficiaries enrolled in HMOs and Medicare FFS drawn from the same markets.

To examine reasons that Demonstration enrollees chose to disenroll from these plans, we analyzed the CMS-sponsored Medicare Consumer Assessment of Healthcare Providers and Systems (CAHPS) Disenrollment Reasons Survey, conducted on an annual basis to determine the reasons beneficiaries voluntarily leave their MA plans. Rather than field a separate disenrollment survey for this evaluation, CMS decided to include a sample of Demonstration disenrollees in the 2004 Medicare CAHPS Disenrollment Reasons Survey—hereafter referred to as the "Reasons Survey." The survey is administered annually to determine reasons beneficiaries leave their MA plans, and because it included all Medicare plans, allowed us to compare reasons for disenrollment between Demonstration and non-Demonstration plans. In this section, we present our findings from both survey analyses.

4.1 Enrollee/Non-Enrollee Survey

For this evaluation, RTI conducted a Medicare Health Insurance Options Survey to gather data to analyze why Medicare beneficiaries chose to enroll, or not enroll, in the Demonstration PPOs available in their area. Participants in this survey included beneficiaries enrolled in a Medicare Demonstration PPO, a Medicare HMO,¹² or FFS in the 210 counties with at least one Medicare PPO Demonstration plan. The mail survey with telephone follow-up of nonrespondents was fielded from July through October 2004. From among the eligible beneficiaries, 31,181 individuals were randomly selected. The overall response rate was 67.6 percent, with PPO Demonstration enrollees having the highest response rate (75.3 percent), followed by HMO enrollees (68.7 percent), and FFS beneficiaries (59.7 percent). The results were weighted and adjusted for survey design to generalize the effects of the Demonstration to the entire Medicare population residing in the Demonstration areas.

The survey analysis focused on three main questions central to understanding the Demonstration:

Did beneficiary characteristics vary by plan type?
What factors affected beneficiary plan choice?
How did beneficiary experience and rating of health care vary by plan type?

4.1.1 Beneficiary Characteristics by Plan Type

Our findings are summarized in Table 4-1, which is discussed below.

Table 4-1
Socioeconomic and health status characteristics of beneficiaries enrolled in Medicare PPO
Demonstration compared with those enrolled in HMOs and FFS
	Sample size	Percentage of enrollees
	N	PPO	HMO	FFS
NOTE: Differences statistically significant at p ≤ .01 are noted with an asterisk (*). SOURCE: RTI analysis of the Medicare Health Insurance Options Survey.
Mean age (in years)	20,304	71.4	73.6*	72.4*
Female	20,304	56.3	59.5	57.3
Race	19,755
White		92.3	87.4*	88.6*
Black		4.5	11.1*	10.0*
Other		3.1	1.5*	1.5*
Hispanic	20,301	3.4	4.8*	3.9*
Education	19,409
Some high school or less		28.8	30.1*	23.4*
High school grad/some college		56.9	56.6*	55.7*
College grad or more		14.3	13.3*	20.9*
Household income	17,522
Less than $10,000		10.1	16.9*	16.5*
$10,000–$19,999		30.4	34.2*	26.0*
$20,000–$29,999		25.9	25.2*	18.6*
$30,000–$49,999		22.1	16.6*	21.3*
$50,000 and over		11.5	7.1*	17.7*
Dually eligible (Medicaid)	20,304	2.1	7.0*	11.9*
Self-rated general health status	19,919
Excellent		7.4	7.0*	6.2*
Very good		24.5	24.8*	23.1*
Good		40.3	39.5*	35.9*
Fair		23.0	23.4*	26.0*
Poor		4.9	5.3*	8.9*
Activity of daily living limitations	20,190
None		64.1	65.2	59.4*
Mild (1–2)		26.9	24.8	25.9*
Severe (3+)		8.9	10.0	14.7*

Socioeconomic Characteristics. We found that PPO Demonstration enrollees were more likely to be white and less likely to be black or Hispanic. There were no meaningful differences by plan type in age or gender. PPO and HMO enrollees had fewer years of education compared with FFS beneficiaries. PPO enrollees reported higher incomes than HMO enrollees. One-third (34 percent) of PPO enrollees had an income greater than $30,000, compared with only 24 percent of HMO enrollees. On the other hand, PPOs enrolled a smaller proportion of both very-low-income and very-high-income beneficiaries than did FFS. This difference can be explained in part by the more heterogeneous group of beneficiaries in FFS. A larger proportion of FFS beneficiaries were poor, as indicated by the proportion dually eligible for Medicare and Medicaid (12 percent FFS versus 2 percent PPO). Poor beneficiaries are unlikely to be able to afford PPO premiums. On the other hand, high-income beneficiaries are more likely to stay in FFS because they are more likely to have employer-sponsored supplemental coverage, or can afford the higher premiums of individually purchased Medigap policies.

Health Status We found no meaningful differences in health between PPO Demonstration and HMO enrollees. However, FFS beneficiaries were sicker than PPO and HMO enrollees. For example, only 5 percent of PPO and HMO enrollees reported "poor" general health compared with 9 percent of FFS beneficiaries (Table 4-1). Respondents were also asked whether they had difficulty performing each of six activities of daily living (ADLs): bathing, dressing, eating, getting in or out of bed or chairs, walking, and toileting. PPO enrollees. were considerably less likely (9 percent) than FFS beneficiaries (15 percent) to have severe ADL limitations (three or more). We found no meaningful differences between PPO enrollees and HMO enrollees in ADL functioning.

4.1.2 Beneficiary Choice of Plan

The second issue we evaluated was beneficiary choice of health plan. In this part of the analysis, we wanted to understand what beneficiaries knew about PPOs and the factors that influenced them to choose or not choose the new PPO Demonstration option. Using the survey results, we first examined how aware beneficiaries were of PPO plans, and what they knew about their features. Then we looked at the reasons beneficiaries reported for choosing their current Medicare health insurance option— including the Demonstration PPOs, Medicare HMOs, or FFS. Finally, we also examined reasons beneficiaries gave for not joining a new PPO plan.

Awareness and Knowledge of PPOs. We found that beneficiary awareness and knowledge of PPO Demonstration plans was generally low, even among those enrolled in PPO plans. Only approximately one-third of PPO enrollees (31.5 percent) and even a lower proportion of HMO enrollees (20.2 percent) and FFS beneficiaries (27.4 percent) recognized the PPO or POS term. For all groups the most frequently cited source of information about PPO or POS plans was the Medicare program or other government agency, followed by advertisements and meetings with a health plan or insurance agent.

Although these awareness rates seem low—particularly among beneficiaries who had in fact enrolled in a Demonstration PPO—awareness of the PPO label appears unnecessary for enrollment, and lack of awareness of the term "PPO" does not necessarily imply a lack of awareness of the features of a specific PPO plan. We found in our case studies of the Demonstration PPOs that many Medicare PPOs do not market themselves as PPOs but rather by a plan name. In enrollment decisions, beneficiaries apparently focus on specific plan features that are of interest to them, as opposed to choosing a generic insurance plan type. Therefore, beneficiaries may not have widespread knowledge of the term PPO and the generic characteristics of PPOs, but they may still have some basic understanding of their specific plan features.

To test whether Medicare beneficiaries had basic knowledge of the common features of PPOs, we included two questions in the survey to assess beneficiary knowledge about specific attributes of plan options. These questions were asked only of the respondents who indicated that they were aware of the term PPO. One question asked beneficiaries whether Medicare PPO plans have more, less, or about the same amount of freedom to choose their doctors and hospitals as Medicare HMOs. Of beneficiaries who reported awareness of PPO plans (n = 6,385), 35 percent of PPO enrollees answered the choice question correctly, compared with only 19 percent of HMO enrollees and 14 percent of FFS beneficiaries. The majority of HMO and FFS respondents (54 percent of HMO enrollees and 58 percent of FFS beneficiaries) responded "Don't Know" to this question, whereas only 27 percent of PPO enrollees responded similarly. The second question asked beneficiaries whether enrollees in PPO plans have to pay more to see doctors who are not on the plan's list of preferred providers. A higher proportion of PPO enrollees (62 percent) answered the second question correctly ("yes"), as compared with FFS beneficiaries (39 percent) and HMO enrollees (36 percent). The majority of HMO and FFS respondents (58 percent of HMO enrollees versus 56 percent, respectively) responded "Don't Know" to this question, whereas only 31 percent of PPO enrollees responded similarly.

Most Important Reason for Joining Plan. Next we asked beneficiaries who had recently switched plans to cite the most important reason for choosing their current option. Among beneficiaries that offered a reason, PPO and HMO enrollees cited most often their desire to control or reduce their costs. By comparison, FFS enrollees most often cited their desire to choose a doctor or hospital (Figure 4-1).

Below is a bar graph, click here for the text describing this graph.

Figure 4-1
Main reason for choosing current plan, by plan type

Figure 4-1

SOURCE: RTI analysis of the Medicare Health Insurance Options Survey.

Reasons for Not Enrolling in a PPO Plan. We analyzed the reasons for not joining a PPO Demonstration plan that were expressed by the 3,009 HMO and FFS enrollees who said they had at least heard of the PPO terminology and had considered joining one. From these responses, we found that a larger proportion of HMO than FFS enrollees (24 percent versus 16 percent, respectively) had considered joining a PPO.

Satisfaction with their current insurance and lack of understanding about PPO Demonstration options were the two leading reasons cited by beneficiaries for not joining a PPO (Figure 4-2), with little difference between the HMO and FFS comparison groups in the frequency cited. These reasons were followed by prescription drug coverage or other benefits not being good enough, and PPO plans costing too much; cost was more frequently cited as a reason for the HMO enrollees and lack of desired benefits, by the FFS beneficiaries.

Below is a bar graph, click here for the text describing this graph.

Figure 4-2
Reasons for not joining a PPO plan, after considering doing so, by current plan type

Figure 4-2

SOURCE: RTI analysis of the Medicare Health Insurance Options Survey.

4.1.3 Beneficiary Reported Experience with Plan

Rating of Overall Health Insurance. Survey respondents were asked to rate their overall health care insurance coverage on a scale from 0 to 10, where 0 is the worst health care insurance coverage, and 10 is the best. PPO Demonstration enrollees on average offered lower scores (6.9) than did HMO or FFS enrollees (both 7.4). We created a three-category grouping of the rating to better understand differences in responses across plans (Table 4-2).

Table 4-2
Rating of satisfaction with insurance, by plan type
	Percentage of respondents
	PPO	HMO*	FFS*
Sample size, n = 18,859. NOTES: 10 indicates the best health insurance coverage possible, 0 indicates the worst coverage possible. Differences statistically significant at p ≤ .01 are noted with an asterisk (*). SOURCE: RTI analysis of the Medicare Health Insurance Options Survey.
Ratings of 0 to 5	30.7	23.5	23.3
Ratings of 6 to 8	43.7	43.0	39.3
Ratings of 9 to 10	25.6	33.5	37.4

The proportion of beneficiaries rating their overall insurance as a "9" or "10" was highest among those in FFS, followed by those in HMOs, and lowest by those in PPO Demonstration plans. Conversely, the proportion of beneficiaries in the lowest rating category was highest among PPO enrollees. We stratified mean insurance ratings by income, health status, out-of-pocket expenses, worry about out-of-pocket expenses, any out-of-network use, and PPO-network provider access problems to investigate the lower ratings of PPO enrollees. We found that PPO enrollees had lower satisfaction ratings in every category within each of these six characteristics except in two instances: (1) for the less than $10,000 income category and (2) for those rating their general health as poor. Low-income enrollees, possibly eligible for Medicaid, and those in poor health may have appreciated the greater freedom of provider choice in PPOs than in HMOs.

Out-of-pocket Expenses. To further investigate reasons for the lower plan ratings of PPOs relative to other beneficiaries, we compared these groups on some key plan features, beginning with out-of-pocket expenses. More PPO Demonstration enrollees reported higher out-of-pocket costs, and a smaller proportion reported no out-of-pocket costs. Almost half (45.4 percent) of PPO enrollees paid more in out-of-pocket costs with their current plan than with their prior insurance arrangement. Only about 23.5 percent paid the same, and 24.9 percent paid less. The same was not true of HMO enrollees, of whom only 27.9 percent paid more in out-of-pocket costs. Exclusive of monthly premium, a much larger proportion of PPO Demonstration enrollees (90.5 percent) had some out-of-pocket expenses in the prior 6-month period compared with HMO enrollees (82.4 percent) and FFS beneficiaries (77.8 percent). Experience with out-of-pocket expenditures of greater than $1,000 was somewhat similar among PPO and FFS beneficiaries (12.1 percent and 13.1 percent) and slightly lower for HMO enrollees (9.0 percent). As expected, enrollment in PPOs appears to come with higher out-of-pocket costs than in HMOs or FFS. (Many FFS enrollees have Medigap or additional retiree coverage to pay for expenses not covered by traditional Medicare.)

We then stratified by income level responses to the question probing the level of beneficiary concern about affordability of out-of-pocket costs. More than half of low income enrollees (less than $10,000 in annual income) in all plan types were "very worried" about affording out-of-pocket costs, as shown in Figure 4-3. However, for FFS and HMO enrollees, concern about affordability fell steadily with higher income. In contrast, anxiety about costs was as high or higher for low- to middle-income PPO Demonstration enrollees ($10,000 to $30,000 income) as for lower-income PPO enrollees. Even among upper-middle-income PPO enrollees ($30,000 to $50,000 income), nearly half (47 percent) were very worried about out-of-pocket expenses. Only among high-income beneficiaries ($50,000+) did concern abate, and even in that category, more than one-quarter of PPO enrollees were very worried about costs. Although PPO enrollees frequently cited a desire to control costs as a reason for enrolling, many appear to have experienced significant out-of-pocket costs. This may account, in part, for the somewhat lower overall plan ratings cited by PPO enrollees.

Below is a bar graph, click here for the text describing this graph.

Figure 4-3
Percentage of beneficiaries very worried about out-of-pocket expenses by income group, by plan type

Figure 4-3

Sample size, n = 19,716.
Differences statistically significant at p ≤ .01 are noted with an asterisk (*).
SOURCE: RTI analysis of the Medicare Health Insurance Options Survey.

Access to Care. Next we compared survey responses about access to care. We asked respondents four questions to determine whether they had had access problems in four areas: (1) having a personal doctor or nurse; (2) needing to see a specialist in the last six months; (3) needing care, tests, or treatment in the last six months; and (4) needing approval from their health plan for any care, tests, or treatment in the last six months. Our survey found that a lower proportion of PPO than HMO or FFS enrollees had experienced a "big" problem for each of the four measures designed to detect an access problem, with PPO access statistically better than HMO access for two of the measures. Respondents in all groups most often cited problems getting to see a specialist in the last six months, with 16 percent of PPO enrollees, 17 percent of HMO enrollees, and 16 percent of FFS beneficiaries reporting a "big" problem in this area. Overall, access to care in PPOs appears to be as good as, if not slightly better than, in other plan types.

Out-of-Network Utilization. We found it surprising that only a slightly higher proportion of PPO than HMO enrollees reported out-of-network use (12 percent versus 10 percent), given that out-of-network coverage is the most important difference between the two plan types. When asked whether they had paid more money to see providers out of their insurer's network, a much larger proportion of PPO enrollees (62 percent) had paid more money than HMO enrollees (38 percent). Because we would expect that all plans with a network would require higher cost sharing for non-network use, these low percentages indicate that many beneficiaries, especially HMO enrollees, did not understand the concept of "preferred providers," at least not this terminology. This is consistent with the finding from our "knowledge" question on preferred providers that many beneficiaries did not understand that PPO enrollees have to pay more to see providers who are not on a plan's preferred list. It is also consistent with the approximately one-fifth of beneficiaries across plan types reporting that they "did not know" whether they had paid more for non-network services. Our findings suggest, however, that in PPOs, where the network/non-network distinction is central, a higher proportion of enrollees understand the higher cost sharing for non-network providers than in HMOs.

We also included a question for only PPO respondents in the survey to assess whether providers had refused a respondent service, given that most PPO plans were offered by insurers who also had closed HMO panels of physicians, and that some providers not under contract with the insurer's HMO might refuse service. Approximately 13 percent of PPO enrollees indicated that they had been refused service by a non-network provider. In short, problems with access to non-network providers had been experienced by a small fraction of PPO enrollees, about the same proportion that had accessed any non-network services.

4.1.4 Summary of Enrollee/Non-Enrollee Survey Analysis

The evidence from this survey suggests that the most frequent reason beneficiaries switched to PPOs was to reduce their overall costs of medical care. The analysis found that although access is no worse in PPOs than in HMOs and FFS, PPO enrollees were less satisfied with their overall health insurance than HMO enrollees and FFS beneficiaries. This lower insurance rating may be attributed in part to higher out-of-pocket expenses reported by PPO enrollees. The fact that PPO enrollees report being more concerned about out-of-pocket expenses than HMO or FFS enrollees is also consistent with this hypothesis.

4.2 Demonstration Disenrollment Rates

In addition to analyzing Demonstration enrollee and non-enrollee reasons for plans choice, we also compared voluntary disenrollment trends among Demonstration and Medicare HMO plans. The findings are based on secondary data analyses of Demonstration PPOs, combined with data from the Medicare CAHPS® Disenrollment Reasons Survey, an annual survey sponsored by CMS to determine reasons beneficiaries leave their HMO plans. The findings described in this report are based on data collected from the sample of disenrollees in the Demonstration PPOs that were included in the 2004 Medicare CAHPS Disenrollment Reasons Survey and from a subset of the disenrollees from Medicare HMOs included in the same CMS Regions and/or Medicare health plan market areas. The target population for the annual Reasons Survey consists of Medicare beneficiaries who voluntarily leave their MA organizations and continuing cost contracts during the calendar year.

4.2.1 Disenrollment Rates

The overall disenrollment rate for PPO Demonstration plans in 2004 was more than 3 percentage points higher than the overall rate for HMO plans in the same market areas. The disenrollment rates for PPO Demonstration plans ranged from 3.5 percent to 46.3 percent with an average overall rate of 10.5 percent. The disenrollment rates for the 80 HMO plans in the same market areas ranged from 1.6 percent to 31.0 percent with an average overall rate of 7.4 percent.

4.2.2 Reasons Cited for Disenrollment

PPO Demonstration disenrollees were significantly more likely to cite information problems than were HMO disenrollees, and the most frequently cited reason by both groups was "After joining the plan, it wasn't what you expected." HMO disenrollees were more likely to cite Doctor Access reasons than PPO Demonstration disenrollees. This is most likely because PPO Demonstration disenrollees had the freedom to go outside the plan's provider panel for physician care, while HMO disenrollees did not. HMO disenrollees were more likely to cite problems getting care than PPO Demonstration enrollees. HMOs are a more restrictive product than PPOs, typically requiring referrals and pre-authorizations that may not be required in PPOs. Finally, we found that PPO Demonstration disenrollees were significantly more likely than HMO disenrollees to cite problems with the affordability, cost, and premiums of the plan. PPOs are a higher cost product, requiring higher upfront premiums than HMOs, and higher enrollee cost sharing if out-of-network services are accessed.

4.2.3 Destination After Disenrollment

Our analysis found that a larger proportion of HMO disenrollees (61 percent) than PPO Demonstration disenrollees (47 percent) say that they returned to another managed care plan after disenrollment, instead of returning to fee-for-service (FFS) Medicare.

SECTION 5
COST IMPACT AND BIASED SELECTION

Our evaluation of the PPO Demonstration included a cost impact and biased selection analysis. The purpose was to answer two related questions. First, what was the impact of the PPO Demonstration on Medicare expenditures for beneficiaries who enrolled in the PPO Demonstration? Second, were beneficiaries who enrolled in the PPO Demonstration different from beneficiaries who did not enroll with regard to expected medical care expenditures; that is, was there biased selection in PPO enrollment?

5.1 Background and Methods

5.1.1 Cost Impact

We conducted a cost impact analysis for the first year of the PPO Demonstration (2003). The cost impact analysis was an estimate of the impact of the PPO Demonstration, including any biased selection that it experienced, on Medicare program expenditures. The estimated impact on program payments is the difference between actual payments for PPO enrollees (including risk-sharing payments¹³) and estimates of what would have been paid for PPO enrollees had they not enrolled in a PPO. If PPO enrollees had not enrolled in a PPO, they would have been enrolled in either FFS or another MA plan. If the local market shares of MA and FFS were reasonably stable, the simplest and most reasonable assumption is that PPO enrollees would have remained in FFS if they were enrolled in FFS at the end of the prior year (2002), and similarly for MA.

We lacked prior enrollment information for PPO enrollees newly enrolled in the Medicare program in 2003. Without any good means of predicting whether these beneficiaries would have enrolled in FFS or MA in the absence of the PPO Demonstration, we randomly assigned them to either FFS or MA in the same proportion as continuing enrollees in PPO plans.

Thus, estimating program expenditure impacts required three component estimates:

Actual 2003 Medicare payments for 2003 PPO enrollees, including risk-sharing payments;
Estimated 2003 MA payments for 2003 PPO enrollees who were:

Enrolled in an MA plan in December 2002, or

New to the Medicare program in 2003 and randomly assigned to MA;

Estimated 2003 FFS payments for 2003 PPO enrollees who were:

Enrolled in FFS in December 2002, or

New to the Medicare program in 2003 and randomly assigned to FFS.

Because of timing and project resource constraints we did not include the second and third years of the Demonstration (2004–2005) in our cost impact analysis. Thus we only present an incomplete cost impact analysis of the Demonstration.

5.1.2 Biased Selection

Our biased selection analysis addressed how PPO enrollees differed from non-enrollees with regard to expected expenditures. Analyzing biased selection is important for several reasons. First, the analysis indicates whether PPOs appealed to a broad cross-section of Medicare beneficiaries, both healthy and sick and of different demographic characteristics. Many studies have found that traditional Medicare HMOs experience favorable selection (Mello et al., 2003 and PPRC 1996 provide reviews of the literature), which was attributed to reluctance of sicker beneficiaries with established providers to accept HMO restrictions on provider choice. With the greater access to a wider range of providers that PPOs provide, they may be more appealing to beneficiaries in poorer health undergoing more frequent medical treatment. Indeed, studies of commercial PPOs have found that they experience selection intermediate between traditional FFS indemnity plans and closed panel HMOs (Wei, Ellis, and Ash, 2001). PPO Demonstration plans also may have been less averse to enrolling sicker beneficiaries because the risk-sharing provisions in their Demonstration reimbursement contracts limited their potential financial downside from such beneficiaries.

Second, biased selection has implications for Medicare program payments. If healthier beneficiaries choose PPOs rather than FFS, and this choice is not fully accounted for by the MA capitation payment formula, Medicare program payments could rise. In a widely cited study, Brown et al. (1993) found that Medicare HMOs increased program payments by 5.7 percent because of favorable selection. Riley et al. (1996) made a similar estimate. The implementation of health-based risk adjustment should improve the accuracy of Medicare capitation payments. But in the first year of the PPO Demonstration, 2003, a demographic model still comprised 90 percent of MA payments, with the other 10 percent adjusted by inpatient diagnoses. Even in the third and final year of the PPO Demonstration, 2005, a demographic model still comprised 50 percent of MA payments, with the other 50 percent adjusted by all-encounter diagnoses. Hence, considerable opportunity for profiting from risk selection existed during the PPO Demonstration.

Measures of biased selection may be categorized by timing relative to PPO enrollment, and type. Previous studies have measured expenditures and health status indicators prior to enrollment, during enrollment, and post-enrollment. The use of pre- and post-enrollment measures was driven partly by limited availability of data during enrollment. But these measures also have conceptual advantages and disadvantages. Prior use differences between PPO enrollees and non-enrollees may overstate selection bias if there is "regression to the mean" in use and expenditures once enrollment occurs. Prior use differences for new enrollees also may not be representative of selection among the larger numbers of "continuing enrollees."¹⁴ Indicators measured during the period of enrollment may be confounded by the different utilization patterns, benefit design, cost sharing, and quality of care of PPOs versus FFS or other MA plans (Tchernis, Normand, Pakes, et al. 2006; Robinson and Gardner, 1995). Indicators measured for PPO disenrollees may not be representative of all PPO enrollees (Cox and Hogan, 1997). Types of indicators that were used to measure biased selection include expenditures (Pauly and Zheng, 2003); utilization (Hill and Brown, 1990); mortality (Cox and Hogan, 1997); diagnoses; functional status (Lichtenstein et al., 1991), self-rated health status (Riley et al., 1996); and risk scores (Greenwald et al., 2000; Feldman, Dowd, and Wrobel, 2003; Pope et al., 2006; Kautter et al., 2007).

In this study, we analyze biased selection indicators—demographics, predicted expenditures, and risk scores—measured during the period of PPO enrollment in 2003, which was the first year of the Demonstration. We then provide an update to our biased selection analysis for 2005, which was the last year of the Demonstration.¹⁵ These characteristics for PPO enrollees were compared to the same indicators measured for FFS and MA enrollees residing in the service areas of the PPO plans. The selection indicators we studied—unlike utilization for example—were not confounded by measurement during the period of PPO enrollment (PPO enrollment might affect utilization differently than FFS or HMO enrollment). A beneficiary's demographics and diagnostic profile should be largely unaffected by PPO versus FFS or HMO enrollment.

5.1.3 Presentation of Results

In this section of the report, because of the potential sensitivity of the results, in contract-specific analyses we refer to each of the 33 individual PPO Demonstration contracts analyzed by an arbitrary "contract number" running from 1 to 33 that we generated to distinguish the contracts, but maintain their anonymity. In many of our analyses, we generated tables with and without enrollees of the Contract 15 PPO Demonstration contract. Contract 15 enrolled over half of the beneficiaries ever enrolled in the Demonstration in 2003, most of whom transferred directly from the sponsoring organization's earlier Medicare HMO product. Because of Contract 15's large size and enrollment rollover, combining it with the other Demonstration contracts may provide somewhat different results than examining the other contracts separately.

5.2 Cost Impact Results

5.2.1 Overall Results

Table 5-1 presents cost impact results for all PPO Demonstration plans. Payments from CMS to PPO Demonstration plans in 2003 totaled $482,470,996, including capitation payments for 89,334 PPO enrollees and risk-sharing payments to 19 participating PPO plans. We estimated that PPO enrollees would have cost CMS $441,620,391 in the absence of the Demonstration. Thus, the PPO Demonstration resulted in an estimated $40,850,605 in higher payments by Medicare in 2003. These payments were made up of $6,779,686 in net risk-sharing payments and $34,070,919 in extra capitation payments. The extra amount represented higher payments of $457 per PPO enrollee in 2003, an increase of 9.3 percent over the costs CMS would have incurred for these beneficiaries in the absence of the Demonstration.

Table 5-1
Cost impact of the PPO Demonstration, 2003
	Enrollees¹	Actual Demonstration payments²	Estimated payments without Demonstration³	Demonstration Cost Impact (Actual–Estimated)
	Enrollees¹	Actual Demonstration payments²	Estimated payments without Demonstration³	Total	Per PPO enrollee	Percent estimated payments without Demonstration⁴
NOTES: ¹ Includes Medicare beneficiaries enrolled in a PPO Demonstration plan for at least 1 month in 2003. ² Demonstration payments include all capitation payments to PPO Demonstration plans and total reconciliation payments made as part of the risk-sharing arrangement. ³ Estimated Medicare payments on behalf of PPO enrollees to managed care organizations or FFS providers in the absence of the Demonstration. Risk-sharing payments would not have occurred in the absence of the Demonstration and are thus $0. ⁴ Total Demonstration Cost Impact divided by estimated payments without Demonstration. ⁵ Risk-sharing amount does not include amount for one insurer, representing three Demonstration contracts. This insurer's risk-sharing amount was not finalized at the time this report was prepared. Also, an additional six Demonstration PPO plans deferred the 2003 reconciliation. ⁶ Non-Risk-sharing amount includes actual payments to PPO Demonstration plans or estimated payments to health plans and FFS providers in the absence of the Demonstration. ⁷ Includes only non-risk-sharing payments. SOURCE: RTI International analysis of 2003 Medicare claims, enrollment, reconciliation, and payment files.
TOTAL	89,334	$482,470,996	$441,620,391	$40,850,605	$457	9.3%
Risk Sharing⁵	89,334	6,779,686	0	6,779,686	76	—
Non-Risk Sharing⁶	89,334	475,691,310	441,620,391	34,070,919	381	7.7%
Prior Enrollment Status⁷
Prior Year FFS	17,419	60,046,849	47,919,537	12,127,312	696	25.3%
Prior Year HMO	64,983	400,341,568	378,726,053	21,615,515	333	5.7%
New Medicare Enrollee	6,932	15,302,893	14,974,801	328,092	47	2.2%

Table 5-1 also shows actual (other than risk-sharing) payments and estimated payments for 2003 PPO enrollees by their prior enrollment status.¹⁶ Among 2003 PPO enrollees, 64,983 were enrolled in an HMO at the end of 2002.¹⁷ These beneficiaries were estimated to have incurred an extra $21,615,515 in CMS expenditures during 2003 because of the PPO Demonstration, or an additional $333 per beneficiary. These expenditures were more than half of the total extra non-risk-sharing payments made by CMS, but only represented a 5.7 percent increase above what CMS would have incurred absent the Demonstration. These extra payments arose from the 99 percent of FFS county payment rate paid under the Demonstration, which was higher than the usual Medicare capitation rate in some counties.

PPO enrollees who were in FFS prior to 2003 incurred estimated extra payments per beneficiary (other than risk sharing) of $696 under the Demonstration. These expenditures represented a 25.3 percent increase in payments for these beneficiaries over what Medicare would have paid for them had they remained in FFS Medicare, a much larger increase than for prior HMO enrollees. While large on a per capita basis, this resulted in only $12,127,312 in total extra payments because there were only 17,419 PPO beneficiaries who had been enrolled in original FFS Medicare at the end of 2002. Higher Demonstration payments for prior FFS enrollees arose from favorable selection of healthier beneficiaries into PPOs (see Section 5.3), and MA capitation rates that were higher than average FFS expenditures in some counties.

PPO new Medicare (program) enrollees saw the lowest extra costs at only $47 per beneficiary, or 2.2 percent higher than without the Demonstration. New Medicare enrollees may not have been enrolled in a PPO for very many months in 2003. This, together with their relatively small numbers, limited their total cost impact.¹⁸

Table 5-2 and Table 5-3, which have the same table shell as Table 5-1, depict the cost impact of the Demonstration for non-Contract 15 and Contract 15 PPO enrollees, respectively. Although Contract 15 comprised 58 percent of PPO Demonstration enrollment, it accounted for less than half of the Demonstration's cost impact. Three factors mostly accounted for this. First, and most important, almost all of Contract 15's enrollment (94 percent) was previously enrolled in Contract 15's HMO. The average cost impact of the Demonstration was much lower for prior HMO enrollees than for prior FFS enrollees. Second, the estimated per capita cost impact for prior FFS enrollees was lower for Contract 15 (17.3 percent) than for non-Contract 15 (26.6 percent) PPO enrollees. Third, Contract 15 did not participate in risk sharing in 2003, and thus received no risk-sharing payments. As a percentage of payments absent the Demonstration, we estimated that Medicare paid 5.9 percent more for Contract 15 PPO enrollees, but 17.4 percent more for non-Contract 15 PPO enrollees.

Table 5-2
Cost impact of the PPO Demonstration, 2003, excluding Contract 15
	Enrollees¹	Actual Demonstration payments²	Estimated payments without Demonstration³	Demonstration cost impact (Actual–Estimated)
	Enrollees¹	Actual Demonstration payments²	Estimated payments without Demonstration³	Total	Per PPO Enrollee	Percent estimated payments without Demonstration⁴
NOTES: Beneficiaries enrolled in the Contract 15 plan are excluded from this table. ¹ Includes Medicare beneficiaries enrolled in a PPO Demonstration plan for at least one month in 2003. ² Demonstration payment includes all capitation payments to PPO Demonstration plans and total reconciliation payments made as part of the risk-sharing arrangement. ³ Estimated Medicare payments on behalf of PPO enrollees to managed care organizations or FFS providers in the absence of the Demonstration. Risk-sharing payments would not have occurred in the absence of the Demonstration and are thus $0. ⁴ Total Demonstration Cost Impact divided by estimated payments without Demonstration. ⁵ Risk-sharing amount does not include amount for one insurer, representing three Demonstration contracts. This insurer's risk-sharing amount was not finalized at the time this report was prepared. Also, an additional six Demonstration PPO plans deferred the 2003 reconciliation. ⁶ Non-Risk-sharing amount includes actual payments to PPO Demonstration plans or estimated payments to health plans and FFS providers in the absence of the Demonstration. ⁷ Includes only non-risk-sharing payments. SOURCE: RTI International analysis of 2003 Medicare claims, enrollment, reconciliation, and payment files.
TOTAL	37,913	$ 149,204,532	$ 127,064,610	$ 22,139,922	$ 584	17.4%
Risk sharing⁵	37,913	6,779,686	0	6,779,686	179	—
Non-Risk Sharing⁶	37,913	142,424,846	127,064,610	15,360,236	405	12.1%
Prior Enrollment Status⁷
Prior Year FFS	15,607	52,292,084	41,306,838	10,985,246	704	26.6%
Prior Year HMO	16,479	77,593,709	73,444,357	4,149,352	252	5.6%
New Medicare Enrollee	5,827	12,539,053	12,313,415	225,638	39	1.8%

Table 5-3
Cost impact of the PPO Demonstration, 2003, including only Contract 15
	Enrollees¹	Actual Demonstration payments²	Estimated payments without Demonstration³	Demonstration cost impact (Actual–Estimated)
	Enrollees¹	Actual Demonstration payments²	Estimated payments without Demonstration³	Total	Per PPO enrollee	Percent estimated payments without Demonstration⁴
NOTES: Only beneficiaries enrolled in PPO Demonstration Contract 15 are included in this table. ¹ Includes Medicare beneficiaries enrolled in the Contract 15 Demonstration plan for at least one month in 2003. ² Demonstration payments include all capitation payments to the Contract 15 Demonstration plan. ³ Estimated Medicare payments on behalf of Contract 15 enrollees in the absence of the Demonstration. ⁴ Total Demonstration Cost Impact divided by estimated payments without Demonstration. ⁵ Non-Risk-sharing amount includes actual payments or estimated payments to Contract 15. SOURCE: RTI International analysis of 2003 Medicare claims, enrollment, and payment files.
TOTAL	51,421	$ 333,266,464	$ 314,555,781	$ 18,710,683	$ 364	5.9%
Risk Sharing	51,421	0	0	0	0	—
Non-Risk Sharing⁵	51,421	333,266,464	314,555,781	18,710,683	364	5.9%
Prior Enrollment Status
Prior Year FFS	1,812	7,754,765	6,612,699	1,142,066	630	17.3%
Prior Year HMO	48,504	322,747,859	305,281,696	17,466,163	360	5.7%
New Medicare Enrollee	1,105	2,763,840	2,661,387	102,454	93	3.8%

Extra costs per beneficiary by Demonstration contract ranged from $1,427, to –$1,041 (savings to Medicare from the Demonstration). All but 5 of 33 health plans incurred extra costs for Medicare by enrolling beneficiaries under the PPO Demonstration. The five plans that saved Medicare money enrolled only 1,524 beneficiaries, or 1.7 percent of the Demonstration's total enrollment. Many of the PPO contracts increased Medicare payments in excess of 20 percent. This large increase in Medicare payments was the result of the reconciliation of risk-sharing payments or the large percentage of Medicare beneficiaries from fee-for-service enrolling in these PPO contracts.

The per capita and percentage cost impact by contract was highly correlated with the percentage of a contract's enrollees who were previously enrolled in FFS versus HMO. Contracts that derived a high percentage of their enrollees from FFS tended to have large per capita and percentage cost impacts. Conversely, contracts that drew a large proportion of their enrollment from HMOs had smaller per capita and percentage cost impacts. This result is consistent with the greater cost impact of the Demonstration on prior FFS than on prior HMO enrollees shown in Table 5-1.

5.2.2 Risk-sharing Payments

We also analyzed CMS 2003 risk-sharing payments to PPO Demonstration contracts. Five of the Demonstration PPO contracts did not participate in risk sharing during 2003. In addition, the three PPO Demonstration contracts of one insurer did not reconcile their 2003 risk sharing payments in time for our analysis.¹⁹ Our results do not include the plans that had not reported for 2003 at the time (2006) of our analysis.

Risk-sharing transactions included payments from CMS to 14 PPO contracts of $7,808,084 and payments from five PPO contracts to CMS of $1,028,398. On net, CMS paid $6,779,686 to PPO contracts. This represented $295 per PPO Demonstration enrollee enrolled in a Demonstration contract with a risk-sharing arrangement with CMS in 2003 and for which risk-sharing amounts were available for our analysis, and 8.8 percent of total 2003 CMS payments to these Demonstration contracts. The $6.8 million in net risk-sharing payments accounted for 16.6 percent of the $40.9 million extra costs of the Demonstration.

The largest CMS risk-sharing payment to a single plan was $2.0 million. Three contracts received $5.5 million of the total $6.8 million in net risk-sharing payments, or 82 percent. Thus, risk-sharing payments were concentrated in a few contracts. Also, risk-sharing payments to some contracts were quite significant on a per capita and a percentage basis. For example, the largest per enrollee risk-sharing payments was $1,155 and accounted for 27.7 percent of total CMS payments to that contract. The largest plan risk-sharing payments to CMS amounted to $426,674. On a per enrollee basis, the largest risk-sharing payments to CMS were $372.

The substantial large positive net risk-sharing payments from CMS to PPO Demonstration plans are surprising in light of the positive cost impact of the Demonstration (CMS paid more for Demonstration enrollees than it would have in the absence of the Demonstration) and the favorable selection of FFS enrollees experienced by Demonstration plans (discussed in the next section). One factor that may have contributed to positive risk-sharing payments was that risk sharing was not mandatory. Plans expecting a favorable risk selection of beneficiaries may have chosen not to participate in risk sharing. However, only 5 of 33 plans did not participate. A final factor could be the method CMS used to determine the medical loss ratio target that was the basis of risk sharing. CMS and the PPO plans may have negotiated the target assuming a neutral PPO selection of enrollees from HMOs, or that PPO enrollees would "look like" HMO enrollees in the same area. As we show in the next section, Demonstration plans received an adverse selection of HMO enrollees, which could have contributed to the positive net risk-sharing payments. Also, we found that the PPO enrollee average risk score was considerably higher than the average area HMO enrollee risk score for the three PPO contracts that received the bulk of the risk-sharing payments.

5.3 Biased Selection Results

Our biased selection analysis compared PPO Demonstration enrollees to HMO and FFS enrollees within the PPO Demonstration service area during 2003, which was the first year of the Demonstration. We compared enrollees on demographic characteristics, predicted expenditures, and risk scores. We also updated our biased selection analysis for 2005, which was the last year of the Demonstration.

5.3.1 2003 Results

Table 5-4 depicts a demographic comparison of each sample: PPO, HMO, and FFS. There is little difference in the male/female proportions across the samples. PPO Medicare/Medicaid dual-eligible enrollment (2.5 percent) was distinctly lower than HMO (7.8 percent), and much lower than the FFS Medicaid proportion (16.0 percent). PPOs did not appeal to poorer beneficiaries, presumably in large part because of their higher premiums. PPO elderly enrollees were younger than the FFS elderly and slightly younger than the HMO elderly. Almost half of PPO enrollment (49.4 percent) was in the youngest elderly age bracket of 65 to 74 year-olds compared to 38.7 percent for FFS and 47.6 percent for HMO. Disabled (under age 65) enrollment in PPOs was a slightly higher percentage of total enrollment (13.0 percent) than in HMOs (10.6 percent), but lower than FFS disabled enrollment (17.4 percent). In sum, compared to FFS, PPO enrollees were more likely to be non-Medicaid, non-disabled, and younger among the elderly. PPO enrollees were more similar to HMO enrollees than FFS beneficiaries. But they were more likely to be non-Medicaid, disabled, and younger elderly than HMO enrollees.

Table 5-4
Demographic distribution of PPO, HMO, and FFS enrollees in the PPO Demonstration service area, 2003
	TOTAL		PPO		HMO		FFS¹
NOTES: ¹ Based on a sample of FFS enrollees (see Kautter et al., 2007). Weighted to represent actual count of Medicare FFS beneficiaries in the PPO service area. SOURCE: RTI International analysis of CMS claims and enrollment data.
TOTAL	9,871,640	100.0%	89,334	100.0%	1,881,960	100.0%	7,900,346	100.0%
Male	4,122,183	41.8	37,811	42.3	790,248	42.0	3,294,124	41.7
Female	5,749,457	58.2	51,523	57.7	1,091,712	58.0	4,606,222	58.3
Medicaid	1,414,461	14.3	2,234	2.5	146,237	7.8	1,265,992	16.0
Female, 0–64	763,072	7.7	6,632	7.4	102,954	5.5	653,486	8.3
Female, 65–74	2,249,975	22.8	24,955	27.9	504,286	26.8	1,720,734	21.8
Female, 75–84	1,957,618	19.8	15,911	17.8	367,228	19.5	1,574,479	19.9
Female, 85+	778,797	7.9	4,030	4.5	117,244	6.2	657,523	8.3
Male, 0–64	820,528	8.3	5,045	5.6	96,114	5.1	719,369	9.1
Male, 65–74	1,743,288	17.7	19,180	21.5	390,678	20.8	1,333,430	16.9
Male, 75–84	1,241,852	12.6	11,351	12.7	248,103	13.2	982,399	12.4
Male, 85+	316,510	3.2	2,230	2.5	55,353	2.9	258,927	3.3

Table 5-5 compares predicted expenditures and risk scores across PPO, HMO, and FFS populations. PPO beneficiaries (risk score = 0.882) were notably healthier than the average beneficiary in FFS Medicare (risk score = 1.030). PPO beneficiaries were predicted to cost about $1,000 less on average than FFS beneficiaries (if the PPO enrollees were in FFS Medicare). PPO enrollees have about the same average risk score and predicted expenditures as HMO enrollees. Hence, PPOs experienced about the same degree of favorable selection relative to FFS as HMOs did. Risk scores and predicted expenditures for Contract 15 and non-Contract 15 PPO enrollees did not differ significantly. PPO enrollees were substantially more expensive than recent (2003) enrollees in HMOs, who enrolled during the same time the (start-up) PPO Demonstration plans operated. The PPO Demonstration plans drew an enrollee population more similar to the entire HMO population than to recent HMO enrollees.

Table 5-5
Predicted expenditures and risk scores of PPO, HMO, and FFS enrollees in the PPO Demonstration service area, 2003
	Enrollees	Predicted expenditures	Risk score
NOTES: ¹ Recent HMO enrollees are beneficiaries who enrolled in an HMO on or after January 1, 2003. SOURCE: RTI International analysis of CMS claims and enrollment data.
Total Sample	9,871,640	$6,696	1.000
PPO	89,334	5,905	0.882
Non-Contract 15	37,913	6,045	0.903
Contract 15	51,421	5,827	0.870
HMO	1,881,960	5,825	0.870
Recent HMO enrollees¹	161,013	4,596	0.686
FFS	7,900,346	6,899	1.030

Table 5-6 shows predicted expenditures and risk scores for 2003 PPO enrollees only, broken down by their prior year enrollment (HMO, FFS, or not in Medicare at the end of 2002), and Contract 15/non-Contract 15 enrollment. Prior year enrollment for this table was assigned in the same manner as for the cost impact analysis.²⁰ PPO enrollees previously in an HMO had much higher risk scores (0.916) than either prior FFS (0.724) or new Medicare enrollees (0.692). Results for predicted expenditures were similar.

Table 5-6
Predicted expenditures and risk scores of 2003
PPO Demonstration enrollees by prior enrollment and Contract 15/Non-Contract 15 status
	Enrollees	Predicted expenditures	Risk score
SOURCE: RTI International analysis of CMS claims and enrollment data.
All PPO Beneficiaries	89,334	$5,905	0.882
Non-Contract 15	37,913	6,045	0.903
Contract 15	51,421	5,827	0.870
New Medicare Enrollees	6,932	4,631	0.692
Non-Contract 15	5,827	4,691	0.701
Contract 15	1,105	4,359	0.651
Prior Year HMO	64,983	6,132	0.916
Non-Contract 15	16,479	6,907	1.031
Contract 15	48,504	5,871	0.877
Prior Year FFS	17,419	4,851	0.724
Non-Contract 15	15,607	4,853	0.725
Contract 15	1,812	4,837	0.722

Excluding the Contract 15 HMO to PPO rollover, PPO plans drew an even less favorable selection from HMOs (prior HMO risk score = 1.031 for non-Contract 15 plans). Comparing to the risk score of the PPO service area HMO population (0.870 from Table 5-5²¹, non-Contract 15 plans drew an adverse selection from HMO enrollees. PPO enrollees drawn from FFS had a much lower risk score (0.724) than the service area FFS population (1.030 from Table 5-5). PPOs were attracting much healthier than average FFS enrollees.

In short, Demonstration PPOs drew an adverse selection from HMOs (excluding the Contract 15 rollover), but a highly favorable selection from FFS. Consequently, although the service area FFS population was significantly sicker than the HMO population on average, PPO enrollees drawn from HMO were much sicker than PPO enrollees drawn from FFS (or PPO new Medicare enrollees). We can hypothesize that sicker HMO enrollees using more medical services may have been disproportionately attracted by the greater freedom of provider choice in PPOs versus HMOs, whereas only healthier FFS beneficiaries using fewer services were willing to accept the greater constraints on provider choice in PPOs versus FFS.

5.3.2 2005 Results

We provide an update of our PPO biased selection analysis for 2005, the third and final year of the Demonstration. Table 5-7 depicts a demographic comparison of each sample: PPO, HMO, and FFS. The results are similar to those obtained for 2003 (Table 5-4). Again, there was very little difference in male/female proportions across the samples, with HMO having a slightly higher percentage of males (58.3 percent for HMO compared with 56.8 percent for PPO and 56.5 percent for FFS). PPO had a substantially lower percentage of older beneficiaries age 85 and above (7.1 percent) than did FFS (11.0 percent), with HMO somewhat in the middle (9.2 percent). A similar pattern occurred for the percentage of enrollees dually eligible for Medicare and Medicaid, with 3.4 percent of PPO enrollees dually eligible, compared with 15.7 percent for FFS enrollees (and 9.2 percent for HMO enrollees). In addition, PPO enrollees had a one-year mortality rate of 3.3 percent, compared with 4.6 percent for FFS (and 3.9 percent for HMO). Table 5-7 provides additional information not shown for 2003. For example, the percentage of PPO enrollees residing in urban areas (95.4 percent) is substantially higher than for FFS enrollees (87.5 percent), and similar to HMO enrollees (95.8 percent). However, PPO had a higher percentage of enrollees in medium and small urban areas (27.3 percent) than did HMO (16.4 percent).

Table 5-8 has the same table shell as Table 5-7, except that it compares PPO enrollees in Contract 15 versus PPO enrollees in Non-Contract 15 plans. For some characteristics, Contract 15 enrollees are more similar to HMO enrollees than to Non-Contract 15 PPO enrollees, which is not surprising given that most Contract 15 enrollees transferred from a discontinued HMO to the same sponsor's Demonstration plan in 2003. Contract 15 has a similar distribution as HMO plans for age, sex, urbanicity, and mortality. However, for other characteristics, Contract 15 enrollees are more similar to Non-Contract 15 PPO enrollees than to HMO enrollees. These characteristics include dual eligibility and race.

Table 5-7
Demographic distribution of PPO, HMO, and FFS enrollees in the PPO Demonstration service area, 2005
	Total		PPO		HMO		FFS
	Count	%	Count	%	Count	%	Count	%
SOURCE: RTI International analysis of CMS claims and enrollment data.
Total	13,963,248	100	122,293	100	2,264,822	100	11,576,133	100
Age
Under 65	2,433,899	17.4	13,014	10.6	219,030	9.7	2,201,855	19.0
65–74	5,819,478	41.7	62,826	51.4	1,064,854	47.0	4,691,798	40.5
75–84	4,222,267	30.2	37,737	30.9	772,634	34.1	3,411,896	29.5
85 and older	1,487,604	10.7	8,716	7.1	208,304	9.2	1,270,584	11.0
Sex
Male	7,928,043	56.8	69,427	56.8	1,319,275	58.3	6,539,341	56.5
Female	6,035,205	43.2	52,866	43.2	945,547	41.8	5,036,792	43.5
Medicaid Status
Medicaid	2,033,927	14.6	4,153	3.4	208,794	9.2	1,820,980	15.7
Non-Medicaid	11,929,321	85.4	118,140	96.6	2,056,028	90.8	9,755,153	84.3
Race
White	11,964,661	85.7	111,896	91.5	1,912,999	84.5	9,939,766	85.9
Black	1,447,756	10.4	7,260	5.9	267,329	11.8	1,173,167	10.1
Other	550,831	3.9	3,137	2.6	84,494	3.7	463,200	4.0
Urbanicity
Urban	12,412,460	88.9	116,613	95.4	2,169,014	95.8	10,126,833	87.5
Large	8,942,411	64.0	83,182	68.0	1,796,773	79.3	7,062,456	61.0
Medium	2,703,332	19.4	21,151	17.3	307,860	13.6	2,374,321	20.5
Small	766,717	5.5	12,280	10.0	64,381	2.8	690,056	6.0
Rural	1,550,788	11.1	5,680	4.6	95,808	4.2	1,449,300	12.5
Urban-adjacent	1,240,784	8.9	5,627	4.6	93,614	4.1	1,141,543	9.9
Non-adjacent	310,004	2.2	53	0.0	2,194	0.1	307,757	2.7
New Medicare Enrollee Status
New Medicare Enrollee	460,564	3.3	3,405	2.8	44,940	2.0	412,219	3.6
Continuing Medicare Enrollee	13,502,684	96.7	118,888	97.2	2,219,882	98.0	11,163,914	96.4
Mortality
Died Following Year	620,523	4.4	4,078	3.3	89,336	3.9	527,109	4.6
Survived Following Year	13,342,725	95.6	118,215	96.7	2,175,486	96.1	11,049,024	95.5

Table 5-8
Demographic distribution of PPO enrollees by Contract 15 vs. Non-Contract 15, 2005
	PPO		Contract 15		Non-Contract 15
	Count	%	Count	%	Count	%
SOURCE: RTI International analysis of CMS claims and enrollment data.
Total	122,293	100	49,578	100	72,715	100
Age
Under 65	13,014	10.6	3,831	7.7	9,183	12.6
65–74	62,826	51.4	23,730	47.9	39,096	53.8
75–84	37,737	30.9	17,855	36.0	19,882	27.3
85 and older	8,716	7.1	4,162	8.4	4,554	6.3
Sex
Male	69,427	56.8	28,840	58.2	40,587	55.8
Female	52,866	43.2	20,738	41.8	32,128	44.2
Medicaid Status
Medicaid	4,153	3.4	1,853	3.7	2,300	3.2
Non-Medicaid	118,140	96.6	47,725	96.3	70,415	96.8
Race
White	111,896	91.5	44,934	90.6	66,962	92.1
Black	7,260	5.9	3,233	6.5	4,027	5.5
Other	3,137	2.6	1,411	2.9	1,726	2.4
Urbanicity
Urban	116,613	95.4	49,572	100.0	67,041	92.2
Large	83,182	68.0	41,273	83.3	41,909	57.6
Medium	21,151	17.3	5,329	10.8	15,822	21.8
Small	12,280	10.0	2,970	6.0	9,310	12.8
Rural	5,680	4.6	6	0.0	5,674	7.8
Urban-adjacent	5,627	4.6	6	0.0	5,621	7.7
Non-adjacent	53	0.0	0	0.0	53	0.1
New Medicare Enrollee Status
New Medicare Enrollee	3,405	2.8	806	1.6	2,599	3.6
Continuing Medicare Enrollee	118,888	97.2	48,772	98.4	70,116	96.4
Mortality
Died Following Year	4,078	3.3	1,893	3.8	2,185	3.0
Survived Following Year	118,215	96.7	47,685	96.2	70,530	97.0

Table 5-9 compares risk scores across PPO, HMO, and FFS populations. Similar to the results for 2003 (Table 5-5), PPO beneficiaries (risk score = 0.921) were notably healthier than the average beneficiary in FFS Medicare (risk score = 1.014). Further, PPOs had a similar average risk score as HMOs (0.921 versus 0.934), implying PPOs experienced about the same degree of favorable selection relative to FFS as HMOs did. Table 5-9 also provides counts and average risk scores for each sample by prior year Medicare plan. The percentage of PPO enrollees coming from FFS the prior year was 10.4 percent, compared with only 6.5 percent for HMO. PPO enrollees previously in a PPO had the highest average risk score (0.947). PPO enrollees previously in an HMO had much higher risk scores (0.914) than either prior FFS (0.825) or new Medicare (program) enrollees (0.530). It is interesting that for beneficiaries enrolled in FFS in the prior year, PPO and HMO experienced a substantial favorable selection relative to FFS, with an average risk score of 0.825 for PPO and 0.882 for HMO, compared with an average risk score of 1.030 for FFS. On the other hand, beneficiaries disenrolling from an HMO into FFS were substantially sicker than average (risk score = 1.141).

Table 5-10 has the same table shell as Table 5-9, except that it compares PPO enrollees in Contract 15 versus PPO enrollees in Non-Contract 15 plans. Excluding Contract 15, the average risk score was 0.878, indicating an even more favorable selection for the Non-Contract 15 PPO plans. This was mainly due to a low average risk score for Non-Contract 15 plan enrollees that were previously enrolled in a PPO plan (risk score = 0.897) or an HMO plan (risk score = 0.901). As for PPO enrollees previously enrolled in FFS, Non-Contract 15 plans actually had a less favorable selection than did Contract 15 (risk score = 0.836 for the Non-Contract 15 plans versus risk score = 0.783 for Contract 15).

Table 5-9
Risk Scores of PPO, HMO, and FFS enrollees in the PPO Demonstration service area, 2005
	Total			PPO			HMO			FFS
	Count	%	Risk Score	Count	%	Risk Score	Count	%	Risk Score	Count	%	Risk Score
SOURCE: RTI International analysis of CMS claims and enrollment data.
Total	13,963,248	100	1.000	122,293	100	0.921	2,264,822	100	0.937	11,576,133	100	1.014
Prior Year Medicare Plan
PPO	106,634	0.8	0.946	96,687	79.1	0.947	6,104	0.3	0.873	3,843	0.0	1.045
HMO	2,121,243	15.2	0.954	9,477	7.7	0.914	2,067,015	91.3	0.950	44,751	0.4	1.141
FFS	11,274,807	80.7	1.028	12,724	10.4	0.825	146,763	6.5	0.882	11,115,320	96.0	1.030
Not Medicare Eligible	460,564	3.3	0.580	3,405	2.8	0.530	44,940	2.0	0.543	412,219	3.6	0.584

Table 5-10
Risk Scores of PPO enrollees by Contract 15 vs. Non-Contract 15, 2005
	PPO			Contract 15			Non-Contract 15
	Count	%	Risk Score	Count	%	Risk Score	Count	%	Risk Score
SOURCE: RTI International analysis of CMS claims and enrollment data.
Total	122,293	100	0.921	49,578	100	0.983	72,715	100	0.878
Prior Year Medicare Plan
PPO	96,687	79.1	0.947	45,536	91.8	1.003	51,151	70.3	0.897
HMO	9,477	7.7	0.914	639	1.3	0.979	8,838	12.2	0.901
FFS	12,724	10.4	0.825	2,597	5.2	0.783	10,127	13.9	0.836
Not Medicare Eligible	3,405	2.8	0.530	806	1.6	0.532	2,599	3.6	0.530

5.4 Summary/Conclusions

We estimated that the Medicare program paid more for the 89,334 beneficiaries enrolled in PPO Demonstration plans in 2003 than it would have paid in the absence of the Demonstration. The total estimated cost impact was approximately $41 million. This amounted to $457 per PPO enrollee, and 9.3 percent of estimated expenditures without the Demonstration. Extra expenditures under the Demonstration were the result of the design of the Demonstration, the characteristics of the beneficiaries who chose to enroll in the Demonstration, and the design of the Medicare program. Four factors accounted for the higher expenditures under the Demonstration:

The 99 percent of FFS per capita expenditures payment rate paid to Demonstration plans where this payment exceeded the Medicare Advantage rates in the applicable counties in 2003. This factor accounted for about $21.6 million of the cost impact.²²
Demonstration plans were offered risk sharing with Medicare, which was not available in the regular Medicare program. Net Medicare risk sharing resulted in about $6.8 million in additional payments to plans.
Demonstration plans enrolled a favorable health status selection of beneficiaries previously enrolled in the original FFS program. Capitation payments under the Demonstration were greater than estimated FFS expenditures. This factor, together with the next, accounted for about $12.1 million of the cost impact.²³
The usual Medicare capitation payment rate was higher than average FFS per capita expenditures in some counties. This factor—which operated for all Medicare capitated plans, not just Demonstration plans—increased Medicare expenditures whenever a FFS beneficiary enrolled in a capitated plan, even with a neutral health status risk selection.

Although expenditures were higher under the PPO Demonstration, the Demonstration may have had offsetting benefits—such as expanding the range of plan choices available to beneficiaries and retaining some plans in the Medicare program—that justified higher expenditures in the eyes of policy makers. The first two factors that led to higher expenditures under the PPO Demonstration were unique to the Demonstration, and are not a feature of Local PPOs under the regular Medicare Advantage program in 2006 and beyond. Comprehensive diagnosis-based risk adjustment for Medicare Advantage fully implemented in 2006 should lessen or eliminate the impact of the third factor. Thus, higher Medicare expenditures under the Demonstration do not imply that Local PPOs are raising Medicare payments beginning in 2006 more than other Medicare Advantage plan types.

We estimated that the predicted medical expenditures (costliness or health status) of PPO Demonstration enrollees in 2003 was about the same as that of HMO enrollees in the PPO plans' service areas. Both PPO and HMO enrollees were predicted to be substantially less costly than service area FFS beneficiaries, on average. Hence, PPOs experienced about the same degree of favorable selection relative to FFS as HMOs did. We also updated our biased selection analysis for 2005, the last year of the Demonstration. We obtained qualitatively similar results to those for 2003. In particular, PPO beneficiaries in 2005 were notably healthier than the average beneficiary in FFS Medicare, and had about the same average risk score as HMO enrollees.

SECTION 6
CONCLUSIONS

6.1 Lessons from the Demonstration about PPOs in Medicare

Evaluation of the PPO Demonstration identified a number of lessons learned about the early implementation of PPOs on a wider scale than previously available. These key lessons are the following:

Demonstration PPO options were available in a wide range of primarily urban counties. Most but not all Demonstration plans were modeled after traditional PPO options offering access to non-network providers for increased cost sharing. Demonstration plans were offered in a variety of counties (179) nationwide. The Demonstration contracts were found in 21 states and in 9 of the 10 CMS regions. However, the Demonstration plans were generally concentrated in the Mid-Atlantic, Midwest, and Southeast states (29 of 35 contracts). Also, open-access Demonstration plans were available in only 6 percent of counties in 2003, rising to 10 percent by 2005. Almost 80 percent (160) of PPO Demonstration plan service area open enrollment counties were located in MSAs (rather than more rural counties), suggesting that PPOs may not be a widespread solution to increasing managed care offerings in rural areas. Organizations we spoke with during the site visits echoed this view, noting that fulfilling CMS provider network requirements in rural areas was very difficult and costly. Also, it was difficult for managed care organizations to achieve the necessary scale for plan viability of 5,000 to 10,000 enrollees in rural areas. As a result, many organizations viewed rural areas as less attractive potential markets.
The PPO Demonstration did not generally attract new managed care organizations to Medicare, although the Demonstration did encourage existing ones to remain and/or expand their options. Although the PPO Demonstration did not attract many new managed care organizations to Medicare, the features of the Demonstration—primarily the favorable payment rates and availability of risk sharing—did keep some organizations in Medicare when they might otherwise have left the program altogether. The best example of this was Horizon in New Jersey, which substituted its existing HMO product with the new PPO Demonstration product because of the higher payment rates available under the Demonstration, and because Horizon was finding it increasingly difficult to offer its former product under the regular health plans program. In addition, some organizations either reentered Medicare markets from which they had previously withdrawn (for example, Aetna in Maryland) or expanded into areas where they did not offer a plan (for example, Health Net had contracts in Arizona but expanded Medicare business into its Oregon organization). In these cases, expanded choices were offered to Medicare beneficiaries as a result of the Demonstration, although not necessarily by new organizations.
Monthly premiums for PPO Demonstration options tended to be higher on average than other MA plans but lower than popular Medigap options. Demonstration PPO premiums were generally higher than competing HMO/CCP options but lower than the most popular Medigap plan (Pope et al., 2005). In the Demonstration, the average monthly plan premium was $90.38 in 2003, declining to $75.63 in 2004 and $60.69 in 2005. These reductions in monthly premiums by Demonstration plans were consistent with an overall trend in premium reductions for MA plans following the MMA-mandated increase in MA capitation payment rates (which took effect in 2004). Most Demonstration plans' premiums were between $50 and $75 in 2003 and 2004, and between $25 and $50 in 2005. PPOs may represent a mid-range product, offering higher premiums and more provider access than HMOs, but lower premiums and less provider access than Medigap. In the view of most organizations participating in the Demonstration, this premium price difference between PPO and other products was important. The organizations we spoke with during the site visits observed that many beneficiaries were willing to pay high Medigap premiums to maintain freedom of provider choice and access to provider services without referral. That said, many organizations also noted that beneficiaries were sometimes willing to consider other options if they could save money, particularly those beneficiaries who were feeling increased financial pressures.
PPO benefit packages focused on out-of-network coverage rather than increased benefits. The key benefit element offered by the Demonstration plans was ability to obtain coverage for out-of-network provider services. In 2003 and 2004, all Demonstration plans covered a core set of services out-of-network, including acute hospitalization, outpatient hospital services, and primary care and specialist physician services. Other standard Medicare benefits—such as skilled nursing facility stays, home health visits, and durable medical equipment—were covered out-of-network by most, but not all, Demonstration plans. In soliciting for Demonstration participants, CMS stressed availability of prescription drug coverage. Most Demonstration plans offered some drug coverage, though PPO drug benefits tended to be less generous on average than those of competing CCPs. For example, only 42 percent of PPO drug benefit plans covered brand drugs, compared with 53 percent of CCP benefit plans.
Enrollment in PPOs started slowly but increased steadily during the Demonstration period. Characteristics of enrollees in the Demonstration were generally similar to those in other MA plans. There was some expectation that PPOs would be more attractive to FFS beneficiaries than other plans (mostly HMOs) because of the PPO model's greater freedom of provider choice. But PPOs drew about the same proportion of their enrollees from FFS as from other MA plans (42 percent versus 40 percent). Also compared with CCPs, Demonstration plans drew a somewhat lower proportion (15 percent versus 23 percent) of their enrollees from recent Medicare program enrollees (beneficiaries new to the Medicare program during the Demonstration period). This finding is inconsistent with the hypothesis that PPOs should have been especially attractive to Medicare "age ins"—those joining the program when they become eligible at age 65. The demographic characteristics of PPO and CCP enrollees were similar, except that PPOs enrolled fewer black persons and other minorities and fewer Medicaid recipients.
Awareness and knowledge of PPOs among beneficiaries was low, but this did not seem to directly impact enrollment. Lack of knowledge about PPOs, however, may have influenced disenrollment among Demonstration enrollees. We found that beneficiary awareness and knowledge of PPO plans was generally low even among those who chose to enroll in a Demonstration PPO plan. Only approximately one-third of PPO enrollees (31.5 percent) and even a lower proportion of HMO enrollees (20.2 percent) and FFS beneficiaries (27.4 percent) recognized either the PPO or POS terms. For all groups, the most frequently cited source of information about PPO or POS plans was the Medicare program or other government agency, followed by advertisements and meetings with a health plan or insurance agent. Although these awareness rates seemed low—particularly among beneficiaries who had in fact enrolled in a PPO—awareness of the PPO label appeared unnecessary for enrollment, and lack of awareness of the term PPO did not necessarily imply a lack of awareness of the features of a specific PPO plan. We found in our case studies of the Demonstration PPOs that many Medicare PPOs did not market themselves as PPOs but rather by a more general plan name. In enrollment decisions, beneficiaries apparently focused on specific plan features that were of interest to them, as opposed to choosing a generic insurance plan type. Therefore, beneficiaries may not have had widespread knowledge of the term PPO and the generic characteristics of PPOs, but they may still have had some basic understanding of their specific plan features. However, PPO Demonstration disenrollees were significantly more likely to cite information problems than were HMO disenrollees.
Enrollees in Demonstration PPOs reported that they had made this choice to save money. However, PPO enrollees were less likely to be satisfied with their choice, possibly because their costs in the PPO were higher than they had expected. The findings from our enrollee/non-enrollee survey suggested that the most frequent reason beneficiaries switched to PPOs was to reduce their overall costs of medical care. The analysis found that although access was no worse in PPOs compared with HMOs and FFS, PPO enrollees were less satisfied with their overall health insurance than were HMO enrollees and FFS beneficiaries. This lower insurance rating may be attributed in part to higher out-of-pocket expenses reported by PPO enrollees. That PPO enrollees reported being more concerned about out-of-pocket expenses than HMO or FFS enrollees was consistent with this hypothesis. PPO Demonstration disenrollees were significantly more likely than HMO ones to cite problems with the affordability, cost, and premiums of their plan.
The PPO Demonstration resulted in increased costs for Medicare. The evaluation analyses estimated that the Medicare program paid more for the 89,334 beneficiaries enrolled in PPO Demonstration plans in 2003 than it would have paid in its absence. The total estimated cost impact was approximately $41 million, which amounted to $457 per PPO enrollee. These higher costs resulted from the higher capitation rate offered under the Demonstration, the availability of risk sharing between CMS and plans, favorable selection into the Demonstration plans, and the higher payments paid by Medicare for enrollees in managed care relative to FFS.
The Demonstration did allow for experimentation with risk sharing in Medicare Advantage. Under the Demonstration, most risk sharing resulted in payments from CMS to plans. For the 2003 contract year, risk-sharing transactions included payments from CMS to 14 PPO contracts of $7.8 million and payments from 5 PPO contracts to CMS of $1.0 million. On net, CMS paid $6.8 million to PPO contracts. This represented $295 per enrollee in a Demonstration contract with a risk-sharing arrangement with CMS in 2003 (for which risk-sharing amounts were available for our analysis), and 8.8 percent of total 2003 CMS payments to these Demonstration contracts. The $6.8 million in net risk-sharing payments accounted for 16.6 percent of the $40.9 million in extra costs of the Demonstration.

6.2 Relevance of the Demonstration to the Evolution of PPOs in Medicare

The PPO Demonstration succeeded in one of its major goals: to increase offerings of PPO plans to Medicare beneficiaries. A number of national, regional and local managed care organizations offered PPO plans under the Demonstration in a wide variety of geographic areas. PPO plans were offered primarily where other Medicare managed care plans were already offered, but this was probably inevitable given the tight time frame of the Demonstration and the difficulty of developing provider networks in areas not already served by managed care. The higher payment rates and other features of the Demonstration encouraged some plans to remain in the Medicare market or to re-enter where they had exited. This was an important accomplishment given the large number of plan withdrawals from the Medicare program around 2003. The risk sharing available under the Demonstration made some plans willing to offer the PPO product, with its potentially risky (to plans) out-of-network utilization feature.

The PPO Demonstration was both a continuation of the trend promoted by policy makers since the Balanced Budget Act of 1997 toward greater variety of private plan types and offerings in Medicare, and the harbinger of the profusion of MA plans offered beginning in 2005. In particular, the PPO Demonstration portended the Local and Regional PPOs established as of 2006 by the Medicare Modernization Act of 2003. In fact, most of the Demonstration plans transitioned to Local PPOs in 2006, suggesting that the Demonstration had a role in the large increase in Local PPOs between 2005 and 2006. The experience gained by CMS and plans in developing and implementing risk sharing was likely valuable under full implementation of the Part D prescription drug plan (in which risk sharing is an important payment feature).

PPOs—in either their Local or Regional varieties—continue to be an important part of the MA program in 2008. They still have not attracted large enrollment, growing much more slowly, for example, than private FFS plans. But given PPOs' continued dominance of the commercial employer-sponsored insurance market, there may be reason to predict that PPOs will play an important role in Medicare's future.

REFERENCES

Brown, R.S., Clement, D.G., Hill, J.W., et al.: Do health maintenance organizations work for Medicare? Health Care Financing Review 15(1):7-23, Fall 1993.

Centers for Medicare & Medicaid Services (CMS), 2006: Medicare Preferred Provider Organization (PPO) Demonstration Fact Sheet. Available at http://www.cms.hhs.gov/DemoProjectsEvalRpts/downloads/PPO_Fact_Sheet.pdf, accessed November 30, 2007.

Cox, D.F., and Hogan, C.: Biased selection and Medicare HMOs: Analysis of the 1989–1994 experience. Medical Care Research and Review, 54(3):259-274, September 1997.

Feldman, R., Dowd, B., and Wrobel, M.: Risk selection and benefits in the Medicare+Choice program. Health Care Financing Review 25(1):23-36, Fall 2003.

Greenwald, L.M., Levy, J.M., and Ingber, M.J.: Favorable selection in the Medicare+Choice Program: New evidence. Health Care Financing Review 21(3):127-134, Spring 2000.

Greenwald, L.M., Pope, G.C., Anderson, W., West, N., Mobley, L, and Kautter, J. Medicare Preferred Provider Organization (PPO) Case Study and Implementation Report." Final Report Prepared for the Centers for Medicare & Medicaid Services, Contract No 500-00-0024, T.O. #5, March 2004.

Hill, J.W., and Brown, R.S.: Biased Selection in the TEFRA HMO/CMP Program. Final Report prepared for Health Care Financing Administration, September 1990.

Kautter, J., Pope, G.C., and Olmsted, E.: Medicare Preferred Provider Organization Demonstration: Cost Impact and Biased Selection in Year One (2003), Final Report Prepared for the Centers for Medicare & Medicaid Services, Contract No 500-00-0024, T.O. #5, April 2007.

Lichtenstein, R., Thomas, J.W., Adams-Watson, J., Lepkowski, J.M., and Simone, B.: Selection bias in TEFRA At-Risk HMOs. Medical Care 29(4):318-331, April 1991.

Mello, M.M., Stearns, S.C., Norton, E.C., and Ricketts III, T.C.: Understanding Biased Selection in Medicare HMOs. Health Services Research 38 (3):961, June 2003.

Pauly, M.V., and Zheng, Y.: Adverse Selection and the Challenges to Stand-Alone Prescription Drug Insurance. NBER Working Paper 9919, 2003.

Physician Payment Review Commission (PPRC): Risk Selection and Risk Adjustment in Medicare. Annual Report to Congress/Chapter 15, 1996..

Pope, G.C., Greenwald, L., Healy, D., Kautter, J., Olmsted, E., West, N.: Impact of Increased Financial Incentives to Medicare Advantage Plans. Waltham, MA: RTI International, 2006.

Pope, G.C., Greenwald, L., Kautter, J., Olmsted, E., and Mobley, L.R.: Medicare preferred provider organization demonstration: Plan offerings and beneficiary enrollment. Health Care Financing Review 27(3):95-109, 2006.

Pope, G.C., Olmsted, E., Kautter, J., Mobley, L.R., and Greenwald, L.: The Medicare Preferred Provider Organization Demonstration: Plan Offerings and Enrollment. Waltham, MA: RTI International, 2005.

Riley, G., Tudor, C., Chiang, Y., and Ingber, M.: Health status of Medicare enrollees in HMOs and fee-for-service in 1994. Health Care Financing Review 17(4):65-76, Summer 1996.

Robinson, J.C., and Gardner, L.B.: Adverse selection among multiple competing health maintenance organizations. Medical Care 33(12):1161-1175, 1995.

Tchernis, R., Normand, S-L.T., Pakes, J., Gaccione, P., and Newhouse, J.P.: Selection and plan switching behavior. Inquiry 43(1):10-22, 2006.

U.S. Government Accountability Office: Medicare Demonstration PPOs: Financial and Other Advantages for Plans, Few Advantages for Beneficiaries. Washington, DC: U.S. Government Printing Office, GAO-04-060, 2004.

Wei, Y., Ellis, R.P., and Ash, A.: Risk selection in the Massachusetts State Employee Health Insurance Program. Health Care Management Science 4:281-287, 2001.

COMPUTER OUTPUT

Table 2-1	-ah_request1.do
Table 2-2	- rq1_AH.xls
Table 2-5	H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request6.do H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request6.log H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request5.do H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request5.log
Table 2-6	H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request6.do H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request6.log H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request5.do H:\project\07964\005 PPO\005\pgm\ykaganova\programs\ppo_2003\ah_request5.log
Table 2-7	– ah_request3.log
Table 2-11	– rq3v1_tab10_2003-2007.lst
Table 2-12	-y04a03c3_rr_V2.lst
Table 5-7	Pgm: tabl1_demoPPO_BR_2005
Table 5-8	Pgm: tabl3_demoPPO_BR_2005
Table 5-9	Pgm: tabl2_demoPPO_BR_2005 and tabl6_demoPPO_BR_2005
Table 5-10	Pgm: tabl8_demoPPO_BR_2005

Endnotes

¹ Beginning in April 2004, the Medicare Advantage base county payment rate was increased to at least 100 percent of the FFS average expenditure, eliminating the higher PPO Demonstration payment rate.

² An "affiliated" plan is a plan offered in the same market area by the same parent company that sponsored a demonstration PPO; for example, a United Healthcare Medicare HMO plan offered in the same service area as the United Demonstration PPO is an "affiliated plan".

³ As of this writing, the 2003 risk-sharing reconciliations have not been finalized for 2003 products offered by 7 organizations. The majority of these plans' 2003 reconciliations were deferred because of. low enrollment in contract year 2003. The net CMS risk-sharing payment to plans does not represent all applicable reconciliations.

⁴ An additional $0.3 million of the demonstration cost impact was the result of higher demonstration non-risk-sharing payments to plans enrolling new Medicare program enrollees. We did not distinguish in our analysis the impacts of the various factors affecting payments for new Medicare program enrollees.

⁵ Regional PPOs were required to have an out-of-pocket maximum on beneficiary cost sharing and a unified Part A and B deductible. Also, unlike other MA plans, Regional PPO bids affected the Regional PPO benchmark payment rate.

⁶ In general, relative costs by health plan and health status do not appear very sensitive to the age range chosen. However, to the extent that Medigap premiums are age-rated, Medigap costs may be higher for older beneficiaries and lower for younger beneficiaries.

⁷ Enrollment in the Medicare Enrollment Database is recorded only by contract, not by plan, so it is not possible to exclude enrollment in employer-only plans.

⁸ By 2005, 3 of these 4 contracts offered employer-specific plans. Unfortunately, enrollment by plan is not available prior to 2006, so we are unable to ascertain how much of these contracts' enrollment growth occurred in employer-specific versus open-access plans.

⁹ If beneficiaries who transferred from Horizon's HMO to its PPO Demonstration contract were included as enrollees from an affiliated HMO, the proportion of demonstration enrollees drawn from an affiliated health plan would be much greater.

¹⁰ We did not interview Anthem, whose two Demonstration contracts began after our site visits were concluded.

¹¹ While Demonstration organizations perceived CMS should have done more to publicize and educate beneficiaries about PPOs, CMS did not want to be in a position of emphasizing one M+C health plan option over others available to beneficiaries. Rather, from CMS's perspective, it was the participating plans' responsibility to market and educate appropriately

¹² The HMO comparison sample included HMO and HMO-POS plan types because it was impossible to exclude the latter. Non-demonstration PPO, provider-sponsored organizations, private FFS plans, and demonstration and cost contracts were excluded from the HMO comparison sample.

¹³ CMS payouts resulting from risk-sharing arrangements are likely to decrease in contract years 2004 and 2005 as a result of more accurate identification of target medical loss ratios which were based, in part, on actual experience with the demonstration products, and changes to benefits over time.

¹⁴ This is less of an issue for the PPO Demonstration because all the PPO plans were startups, at least in 2003.

¹⁵ As a check on the robustness of our results, we repeated the biased selection analyses for age groups. The results of the biased selection analysis by age groups were qualitatively the same as for the entire sample.

¹⁶ Risk-sharing payments were made on a plan aggregate basis and cannot be allocated below the plan level.

¹⁷ As stated earlier, in this report we use the term "HMO" loosely to include all enrollees in Medicare health plans, most of whom were in fact enrolled in HMOs. Also, note that the Contract 15 demonstration PPO plan contained 48,504 enrollees from its prior HMO plan who were rolled into its demonstration PPO plan in 2003.

¹⁸ In addition, we were able to use only a demographic expenditure prediction model for many new Medicare enrollees assigned to FFS in the absence of the demonstration, limiting our ability to capture any favorable selection in PPO enrollment of new Medicare enrollees.

¹⁹ An additional six demonstration PPO plans deferred the 2003 reconciliation.

²⁰ See Kautter et al. (2007).

²¹ The service area HMO risk score includes HMO enrollees in the Contract 15 service area. But Kautter et al. (2007) shows that they comprised only 2.5 percent of PPO combined service area HMO enrollees and had an average risk score of 0.952. Thus, the overall service area HMO risk score of 0.870 is a roughly accurate representation of the average risk score of HMO enrollees in the non-Contract 15 combined PPO service areas.

²² The impacted service areas were not extensive. However, we would like to note that Contract 15's service area was one, and much of the impact in terms of added payments was the result of Contract 15's large enrollment base that year.

²³ An additional $0.3 million of the demonstration cost impact was the result of higher demonstration non-risk-sharing payments to plans enrolling new Medicare program enrollees. We did not distinguish in our analysis the impacts of the various factors affecting payments for new Medicare program enrollees.

March 2008

Evaluation of the Medicare Preferred Provider Organization (PPO) Demonstration

Final Report

ACKNOWLEDGMENTS

CONTENTS

List of Figures

List of Tables

ES.1 Background

ES.2 Key Evaluation Findings

ES.2.1 Plan Participation and Availability

ES.2.2 Plan Premiums, Benefits, and Cost Sharing

ES.2.3 Plan Enrollment, Market Share, and Prior Enrollment Status

ES.2.4 Findings from Case Study Plan Interviews

ES.2.5 Findings from Beneficiary Surveys

ES.2.6 Impact of the Demonstration on Medicare Program Payments

ES.2.7 Demonstration Risk-sharing Payments to Plans

ES.2.8 Health Status and Other Characteristics of Demonstration Enrollees

ES.3 Conclusions

Figure 1-1 Structure of the PPO evaluation

Figure 2-1 Service areas of Medicare PPO Demonstration contracts, April 2004

Figure 2-2 Distribution of PPO and CCP counties by urbanicity, 2004

Figure 2-3 PPO and competing CCP and Medigap Plan F monthly premiums, 2004

Figure 2-4 Predicted out-of-pocket cost by plan type: Beneficiaries aged 70–74

Figure 2-5 Enrollment in the PPO Demonstration, 2003–2005

Figure 4-1 Main reason for choosing current plan, by plan type

Figure 4-2 Reasons for not joining a PPO plan, after considering doing so, by current plan type

Figure 4-3 Percentage of beneficiaries very worried about out-of-pocket expenses by income group, by plan type

Evaluation of the Medicare Preferred
Provider Organization (PPO)
Demonstration

Figure 1-1
Structure of the PPO evaluation

Figure 2-1
Service areas of Medicare PPO Demonstration contracts, April 2004

Figure 2-2
Distribution of PPO and CCP counties by urbanicity, 2004

Figure 2-3
PPO and competing CCP and Medigap Plan F monthly premiums, 2004

Figure 2-4
Predicted out-of-pocket cost by plan type: Beneficiaries aged 70–74

Figure 2-5
Enrollment in the PPO Demonstration, 2003–2005

Figure 4-1
Main reason for choosing current plan, by plan type

Figure 4-2
Reasons for not joining a PPO plan, after considering doing so, by current plan type

Figure 4-3
Percentage of beneficiaries very worried about out-of-pocket expenses by income group, by plan type