We report a meta-analysis seeking socio-demographic and medical moderators of the effectiveness of psycho-oncological interventions. The dataset comprised 146 published and unpublished prospective controlled trials where outcomes were anxiety, depression, and general distress. Analyses took into account the confounding potential of two study design features discovered in preliminary analyses, and produced a coherent pattern of moderation suggestive of the critical influence of elevated baseline distress. Being older, having lower income, being male, or having a cancer other than (early stage) breast cancer produced higher statistically significant or trend effects, and being single produced nonsignificant effects double in magnitude to effects produced by married patients. Elevated baseline distress itself produced a significantly higher effect. Implications for research and practice are discussed.