Considerations of psychosocial illness phase in cancer survival

Rarely are biomedical, clinical and psychosocial data considered simultaneo usly as influences on cancer patient outcomes. This study utilized medical record and interview data from 152 adult cancer patients with various tumor types in a model of survival estimation. Predictors included disease stage of the neoplasm (TNM stage), clinical functioning of the patient (Karnofsk y performance status), and psychosocial demands of the disease course (psyc hosocial illness phase). Psychosocial illness phase captures developmental time phases of illness (i.e. 'crisis', 'early chronic', 'late chronic' and 'terminal'), essentially locating patients along the disease course relativ e to treatment and treatment response. The analysis utilized the Kaplan-Mei er (product-limit) method to estimate stratum specific survival functions. Model comparisons employed the differences in the likelihood ratio chi-squa res between nested models, and Cox proportional hazard models assisted in e xplaining the effects of the predictors on survival times. Results indicate that psychosocial illness phase makes an independent contribution to survi val time estimation (p < 0.05) when all three dimensions are considered sim ultaneously. Copyright (C) 2001 John Wiley Ltd. Sons, Ltd.