This paper examines the consequences of patient enrollment procedures in a capitation-based family practice in Krakow (Poland), where the local city government used two different methods for preparation of patient lists. In the first, the city gave the individuals living within the practice area the option of withdrawing from being enrolled in the practice; in the second, individuals were given the option of enrolling in the practice. These two enrollment procedures, identified as 'active-negative' and 'active-positive' respectively, provide a natural experiment for investigating the effects of an enrollment methodology on the economics of a physician's practice. An examination of the data indicates that self-selecting enrollees utilize significantly greater quantities of health care compared to others, and university educated individuals and individuals more likely to fall ill are more likely to self-select into a practice. The study suggests that in order to reduce demand-side adverse selection, either the system of active-positive enrollment should be modified, or capitation rates should be risk-adjusted by health status rather than by demographic variables only. The policy implications of this study become even more significant as more and more physicians leave their salaried jobs to start state-financed independent practices.