A structural model was proposed to explain the stress resulting from hospitalization for a medical problem. Perceived uncertainty about symptoms, treatment, and outcome was examined as a major predictor of stress. Other variables proposed in the model included seriousness of illness, age, education, and recency of rehospitalization. Testing of the model with hospitalized medical patients indicated support for the relationship of uncertainty to stress. Uncertainty also had the predicted mediating role between seriousness of illness and stress. The only other variable supporting the proposed model was age which related inversely to stress. Much of the unexplained variance in the model is attributed to the heterogeneity of diagnoses in the sample. Suggestions for future testing of the model are presented.