Naltrexone was approved by the US Food and Drug Administration (FDA) in December 1994. Although it is one of only two medications for alcohol rehabilitation approved in the US, naltrexone is not frequently prescribed. This paper examines the factors limiting physicians' use of naltrexone. Data were collected through a survey of US physician members of two addiction medicine professional associations. Logistic regression analysis was conducted with the dependent variable being any prescription of naltrexone. Independent variables consisted of physician practice characteristics, physician ratings of naltrexone's attributes (e.g. efficacy), physicians' patient characteristics, and physicians' exposure to information about naltrexone. On average, addiction medicine physicians prescribed naltrexone to 13% of their alcoholism patients. The two main self-reported reasons why physicians did not prescribe the medication to more patients were that patients refused to take the medication or comply with prescribing regimes (23%), and that patients could not afford the medication (21%). Multivariate results indicated that physician perceptions of naltrexone's effectiveness and safety were significantly associated with prescribing. Physicians who had more exposure to information about the product (e.g. by reading more journal articles about naltrexone) were more likely to prescribe it. In summary, multiple factors are contributing to the relatively low naltexone prescription rate.