The Duke--UNC Health Profile (DUHP) was developed as a brief 63-item instrument designed to measure adult health status in the primary care setting along four dimensions: symptom status, physical function, emotional function and social function. Reliability and validity were tested on a group of 395 ambulatory patients in a family medicine center. Temporal stability Spearman correlations ranged from 0.52 to 0.82 for the four dimensions. Cronbach's alpha for internal consistency was 0.85 for emotional function. Guttman's reproducibility coefficients were 0.98 for physical function and 0.93 for social function, and the scalability coefficients were 0.89 for physical and 0.71 for social. Observed relationships between DUHP scores and demographic characteristics of the respondents correlated well with those predicted by the investigators (overall Spearman correlation 0.79). Convergent and discriminant validity was supported by strong associations between components of DUHP and those on the Sickness Impact Profile (SIP), the Tennessee Self-Concept Scale (Tennessee), and the Zung Self-Rating Depression Scale (Zung). DUHP with SIP monocomponent-heteromethod Spearman correlations ranged from 0.34 to 0.45, and those for DUHP with Tennessee ranged from 0.68 to 0.81. DUHP with Zung monoitem--heteromethod correlations ranged from 0.54 to 0.57. It is concluded that this evidence supports the reliability and validity of the DUHP as an instrument suitable for studying the impact of primary health care on the health outcomes of patients.