An exponential spatial interaction model was used to analyze primary-health-care-seeking patterns of rural consumers in northern California. A macrotheory of spatial interaction and a microtheory of consumer search were combined as a logically consistent rationale for the model. Ordinary least squares proved superior to nonlinear least squares and a limited dependent variable maximum likelihood technique in predicting health-care-seeking patterns. The model was used to address policy questions including the impact of removing local providers on patient flows to remaining providers and the market shares that could be expected by a new provider. © 1987 American Agricultural Economics Association.