Are randomized trials really the best way to evaluate Complementary Therapies efficacy? Novel patient recruitment, study design, and data analysis: How to design CAM trials which patients will join

Background: Less than 3% of cancer patients participate in randomized trials, possibly even less with CAM trials because these therapies are so widely available. However, many cancer patients use CAM therapies, despite the lack of reliable survivorship data from randomized trials. Over the past 15 years, very few federally funded randomized CAM cancer trials have been published, and only a few in which cancer survival was the primary endpoint. Alternative trial designs, which allow for patient preference, are needed if we are to realize any promise of reliable data. Methods: We conducted a 10-year retrospective study to investigated the effect of Chinese herbal medicine, vitamins and acupuncture on survival in a consecutive case series of all non-small cell lung cancer patients (n=239) presenting at Pine Street Clinic. We examined the effect of short-term vs. long-term CAM therapy, and CAM compared to conventional alone, using concurrent controls from Kaiser Permanente and California Cancer Registries. We used traditional Cox regression, and causal inference with Propensity Score and Marginal Structural Models. Results: CAM combined with conventional therapy reduced risk of death in stage IIIA by 46% (HR=0.54; 95% CI 0.41, 0.70), stage IIIB by 62% (HR=0.38; 95% CI 0.28, 0.50), and stage IV by 69% (HR=0.31; 95% CI 0.20, 0.48), vs. conventional alone. Survival rates for stage IV patients treated with PAM+V were 82% at 1 year, 68% at 2 years, and 14% at 5 years. Application: These innovative trial designs allow researchers to obtain credible CAM cancer survival data analysis results, using clinical records gathered in the community care setting.