The problem of planning for the expansion of a rural primary health care delivery system is conceptualized as determining the set of facilities to be added to an existing system so as to maximize the incremental benefit to the community subject to a cost constraint. The proposed approach involves a five step procedure: (1) Facility attributes or characteristics (e.g., travel time, time to get an appointment, type of health care facility, and hours of operation) which are relevant to the patient in his health care seeking behavior are identified; (2) An individuals overall preference for alternate health care facilities is expressed in terms of a weighted linear model of facility attributes. The weights or importances that an individual attaches to the different facility attributes are estimated from his rank order overall preference judgments of a set of fictitious, but realistic facilities. This procedure is repeated for each individual in a representative sample of households from the region of interest; (3) Each consumers preference model is transformed into a benefit function expressing the individuals benefit for a facility in $/year; (4) A method for determining the total incremental benefit to the community from a set of proposed facilities is provided; (5) A heuristic procedure then determines a set of facilities which yields near-optimal total incremental benefit subject to the cost constraint. A practical application of the proposed approach reveals that the consumer preference model has good predictive validity and that it can be reliably estimated.