Comparing Two AI Coaching Interventions to Boost Physical Activity in Cancer Survivors: A Randomized Controlled Trial

Abstract

Background: Physical activity is essential for cancer survivors, improving function, reducing fatigue, and lowering recurrence risk, yet adherence remains poor due to physical and psychosocial barriers. Artificial intelligence (AI)–driven health coaching offers scalable, personalized support. Objective: To compare the effectiveness of two AI coaching interventions—SmartText and Alfa Fitness—in increasing physical activity among cancer survivors. Methods: In a three-arm randomized controlled trial at a tertiary center in Lahore, Pakistan, 60 post-treatment survivors were allocated equally to Control, SmartText, or Alfa Fitness. Interventions were delivered for 4 weeks. The primary outcome was change in average daily steps, measured with calibrated pedometers. Secondary outcomes were health-related quality of life (FACT-G) and exercise motivation (BREQ-3). Analyses used repeated-measures ANOVA with post-hoc comparisons and 95% confidence intervals (CIs). Results: Alfa Fitness significantly increased daily steps (+3,618; 95% CI: 2,490–4,764; p<0.001), while SmartText produced a smaller, nonsignificant gain (+1,619; 95% CI: –328 to 3,566; p=0.051). Control participants showed minimal change (+886; 95% CI: –895 to 2,667). FACT-G improved by +5.2 points (95% CI: 2.1–8.3) with Alfa Fitness and +2.8 points (95% CI: 0.1–5.5) with SmartText, compared with +1.1 in Control. No adverse events occurred, and adherence exceeded 90%. Conclusion: Interactive AI coaching (Alfa Fitness) significantly enhanced physical activity and quality of life among cancer survivors, outperforming text-based messaging. These findings support integrating adaptive AI tools into survivorship care, with larger, longer trials needed to confirm sustainability and cost-effectiveness.