RTI uses cookies to offer you the best experience online. By clicking “accept” on this website, you opt in and you agree to the use of cookies. If you would like to know more about how RTI uses cookies and how to manage them please view our Privacy Policy here. You can “opt out” or change your mind by visiting: http://optout.aboutads.info/. Click “accept” to agree.


New paper analyzes health equity and disparities from IRS tax documentation submitted by U.S. nonprofit hospitals

Analysis identifies affordability and mental health among main themes


RESEARCH TRIANGLE PARK, N.C. — A new paper by experts at RTI International, a nonprofit research institute, was published in the Journal of Medical Internet Research (JMIR), the leading peer-reviewed journal for digital medicine and health and health care in the internet age. The paper, “Text Analysis of Trends in Health Equity and Disparities from IRS Tax Documentation Submitted by U.S. Nonprofit Hospitals between 2010 and 2019,” was authored by RTI experts Emily Hadley, Laura MarcialWes Quattrone and Georgiy Bobashev

Many U.S. hospitals are classified as nonprofits and receive tax-exempt status partially in exchange for providing benefits to the community. The RTI authors used text analysis to examine trends in health equity and disparities based on IRS tax documentation submitted by these hospitals. 

“Hospital community benefits tax documentation has historically been cumbersome for both researchers and the public,” said Emily Hadley, a research data scientist at RTI and one of the paper’s authors. “It was exciting to use data science tools to illuminate national trends in health equity and disparities and highlight opportunities for hospitals to seek better alignment with community needs. Our work demonstrates the potential for text analysis to support greater transparency and accountability and facilitate stakeholder-driven research with large amounts of text data.” 

The IRS collects proof of compliance using the Schedule H form that nonprofit hospitals submit as part of the annual IRS Form 990 (F990H). This includes a free-response text section known for being ambiguous and difficult to audit. For this reason, the researchers used natural language processing (NLP) to evaluate this text section with a focus on health equity and disparities. This research is among the first to use NLP for this text analysis. 

When the team analyzed the text, they found an increased usage of text related to 29 themes around health equity and disparities. They also found that more than 90% of hospital reporting entities used a term in 2018 and 2019 related to affordability, government organizations, mental health and data collection. The themes with the largest relative increase were LGBTQ (1676.6%), social determinants of health (SDOH) (958.4%) and environment (522%).  

The authors also found that terms related to homelessness varied geographically from 2010 to 2018 and terms related to equity, health IT, immigration, LGBTQ, oral health, rural, SDOH and substance use had statistically significant geographic variation in 2018. Terms related to substance use saw the largest raw percentage point increase: only a quarter of hospital reporting entities used any substance use language in 2010 while more than two-thirds of hospital reporting entities used a substance use term in 2019. However, usage in themes like LGBTQ, disability, oral health, and race and ethnicity ranked lower than public interest in these topics and some increased mentions of themes with large increases in usage were to explicitly say that no action was taken by a hospital on those themes.  

Overall, the paper reveals that hospital reporting entities are demonstrating an increasing awareness of health equity and disparities topics in community benefits tax documentation, but these do not necessarily correspond with general population interests or additional action. 

Read the full study

Learn more about RTI’s work around Community Benefits: