Predictors of Intervention Session Completion in a Randomized Clinical Trial of a Behavioral Cancer Pain Intervention

Abstract

ContextSome patients with cancer are able to complete psychosocial pain management intervention sessions, and others find it difficult to do so. ObjectivesConduct a secondary analysis of a randomized clinical trial (N = 178) that compared delivery formats (in-person vs. videoconference) of a pain coping skills training (PCST) intervention for patients with cancer to examine if intervention session completion predicts postintervention outcomes of pain severity and interference, psychological distress, physical well-being, and pain self-efficacy; and identify predictors (i.e., demographics, medical characteristics, baseline outcome scores) of session completion. MethodsSession completion (i.e., completing all four sessions vs. missing at least one session) was tested as a predictor of postintervention outcomes. Predictors of session completion were then examined. ResultsIn both study conditions combined, PCST session completion predicted improvement from baseline to postintervention in pain severity (β = −0.27; P = 0.03), pain interference (β = −0.25; P = 0.048), and pain self-efficacy (β = 0.23; P = 0.07). Participants in the videoconference condition were significantly more likely than those in the in-person condition to complete all sessions (83% vs. 65%; P = 0.006). Participants with at least some college education (odds ratio [OR] 4.36; P = 0.04), a diagnosis of breast cancer (OR 6.73; P = 0.04), and higher levels of pain self-efficacy (OR 2.32; P = 0.02) were more likely to complete videoconference sessions. Participants who lived closer to the medical center (OR 0.64; P = 0.07), had early stage cancer (OR 3.82; P = 0.07), and fewer medical comorbidities (OR 0.59; P = 0.04) were more likely to complete in-person sessions. ConclusionCompleting PCST sessions is important for improving pain outcomes. Efforts to increase session completion (e.g., videoconference delivery) should be considered.