In this study the relationship between four psychologic health constructs (depression, anxiety, patient response bias, and hostility) and the Sickness Impact Profile (SIP) measurement of functional status was evaluated. The SIP, Carroll Depression Rating Scale (CDRS); and the Minnesota Multiphasic Personality Inventory (MMPI) were administered to 332 patients hospitalized for treatment of combined medical and psychiatric problems. Pearson's product-moment correlation was high between CDRS and SIP Total score (r = 0.67) and between CDRS and SIP Psychosocial subscale (r = 0.72); correlation was lower between CDRS and SIP Physical subscale (r = 0.44). Six MMPI scales (depression, anxiety, psychasthenia, lie, K, hostility) correlated with SIP Total score (r = 0.18 to 0.50), with SIP Psychosocial score (r = 0.28 to 0.65) and less well with SIP Physical subscale (r = 0.07 to 0.25). Factor analysis of the SIP categories showed two factors with eigenvalues greater than 1. Promax factor rotation showed all SIP Psychosocial categories and all measured psychologic variables loaded most heavily on factor 1. SIP Physical categories loaded most heavily on factor 2. Stepwise multiple regression analysis showed that psychologic variables account for 49% of the SIP total variance, 62% of SIP Psychosocial subscale variance, but only 19% of SIP Physical subscale variance. The CDRS accounts for the major portion of the explained variance with only minor additional contributions from the MMPI scales. We conclude that 1) the SIP discriminates psychosocial and physical dysfunction even in medical patients with extensive psychiatric comorbidity; 2) the SIP measures at least two dimensions of health, one of which is strongly related to depression; and 3) constructs measured by MMPI scales do not have substantial independent contribution to SIP variance.