OBJECTIVE: Most studies on medication compliance originate in the medical care setting. In this study the aim was to analyze self-reported compliance with prescription drug treatment in a population. METHODS: Cross-sectional survey, statistical sample, 20-84 years, in the County of Uppland, Sweden. 5404 (68%) answered the questionnaire. A recall period of two weeks was used for use of prescription drugs. Logistic regression analysis was used for the multivariate analyses. RESULTS: The use of prescription drugs was reported by 2604 persons (48.2%). Of these users 63.5% reported that they had on occasion or often forgotten to take their medicines, 20.8% reported that they had changed dosages on their own and 25.5% answered that they had stopped medication on their own accord. Men stopped taking drugs to a greater extent than women; older persons were more likely to continue drug tretment as prescribed. Individuals with a higher educational level and those with a good medication knowledge reported that they stopped taking prescription drugs to a greater extent than others. The same pattern was seen with respect to the changing of dosages. Further, we found that it was less common that users of drugs for hypertension, users of cardiovascular drugs, users of drugs for angina, and users of antibiotics stopped taking drugs on their own accord while it was more common that users of analgesics, and users of hypnotics, anxiolytics and antidepressants stopped taking drugs as compared to other users. CONCLUSION: This epidemiological study highlights that non-compliance is a public health problem, that there are variations between users of different types of drugs, and that much more research is needed on non-medical and medical factors affecting medication compliance in the population.