Objective: The study investigated whether any of four measures of religiousness predicted longer survival for 145 African-American and 177 White women diagnosed with breast cancer in Connecticut between January 1987 and March 1989.
Method: Multivariate Cox proportional hazards models included a religious predictor and sociodemographic, biomedical, treatment, behavioral, and medical care covariables.
Results: The no denomination group had a hazard ratio (HR) of 4.39 (95% Confidence Interval (CI) = 1.42, 13.64). Other measures of religiousness did not yield statistically significant results but showed a consistent pattern of nonreligiousness being more strongly and negatively related to breast cancer survival in African Americans than in Whites.
Conclusions: Exploratory models confirmed that lack of religiousness was associated in this sample with poor breast cancer survival among African American women.