case studies digital coaching clinical trial AI coaching weight loss study telehealth fitness
Clinical Trial: Digital Coaching Impact | Kalsofit Case Study
Dr. James Park, Kalsofit Research Team 12 min read
Clinical Trial: Digital Coaching Impact
This randomized controlled trial (RCT) represents the first peer-reviewed comparison of AI-powered digital coaching against traditional in-person personal training for sustained weight management.
Study Methodology
Design
- Type: Single-blind RCT (participants blinded to hypothesis)
- Duration: 12 months
- Follow-up: 6-month post-intervention monitoring
- Registration: ClinicalTrials.gov NCT05849221
Participants
- N: 240 adults (ages 25–55, BMI 27–35)
- Recruitment: Primary care physician referrals
- Exclusions: Eating disorders, uncontrolled diabetes, recent bariatric surgery
- Randomization: Block randomization by age, gender, baseline BMI
Arms
Arm A: AI Digital Coaching (n=80)
- Kalsofit Premium with AI Coach
- Wearable device provided
- Weekly AI-generated plans
- 24/7 chat-based coaching support
- Monthly video check-ins with human coach
- Cost: $29/month
Arm B: Traditional Personal Training (n=80)
- 2x weekly 60-minute sessions with certified trainer
- Custom nutrition plans
- Weekly weigh-ins
- Cost: 480/month(60/session × 8)
Arm C: Self-Directed (Control, n=80)
- Standard calorie-counting app
- Generic exercise recommendations
- No coaching support
- Cost: Free
Primary Outcomes
Weight Loss at 12 Months
| Metric | AI Coaching | Personal Training | Self-Directed |
|---|---|---|---|
| Mean weight lost | -9.2kg | -10.1kg | -3.4kg |
| % achieving ≥5% loss | 78% | 81% | 34% |
| % achieving ≥10% loss | 52% | 58% | 12% |
| Statistical significance | p<0.001 vs Control | p<0.001 vs Control | Baseline |
Between Arms Comparison
- AI vs PT: No statistically significant difference (p=0.34)
- AI vs Control: Highly significant (p<0.001)
- PT vs Control: Highly significant (p<0.001)
Secondary Outcomes
Body Composition
- Fat mass lost: AI -7.8kg vs PT -8.4kg vs Control -2.1kg
- Lean mass preserved: AI +0.3kg vs PT +0.7kg vs Control -0.4kg
- Waist circumference: AI -8.2cm vs PT -9.1cm vs Control -3.1cm
Metabolic Health
| Marker | AI Coaching | Personal Training | Self-Directed |
|---|---|---|---|
| A1C change | -0.4% | -0.5% | -0.1% |
| LDL change | -12 mg/dL | -14 mg/dL | -3 mg/dL |
| HDL change | +6 mg/dL | +7 mg/dL | +1 mg/dL |
| Triglycerides | -28 mg/dL | -32 mg/dL | -8 mg/dL |
Behavioral & Psychological Outcomes
Adherence
- Session completion: AI 87% vs PT 76% vs Control 42%
- Program retention at 12 months: AI 71% vs PT 68% vs Control 31%
- Wearable usage (AI arm): 6.2 days/week average
Psychological Metrics
- Exercise self-efficacy: AI +32% vs PT +28% vs Control +8%
- Nutrition knowledge: AI +41% vs PT +35% vs Control +12%
- Quality of life (SF-36): AI +18 points vs PT +19 points vs Control +4 points
Cost-Effectiveness Analysis
Cost Per Kilogram Lost
- AI Coaching: $38/kg
- Personal Training: $571/kg
- AI delivered equivalent outcomes at 6.6% of the cost
Cost Per Quality-Adjusted Life Year (QALY)
- AI Coaching: $1,200/QALY
- Personal Training: $18,400/QALY
- Both well below the $50,000/QALY intervention threshold
Limitations & Future Research
- Population: Predominantly educated, urban participants. Generalizability requires further study.
- Duration: 12 months may not capture long-term maintenance.
- Technology literacy: AI arm required basic smartphone proficiency.
- Hybrid potential: Future research should examine AI + periodic human touch models.
Clinical Implications
AI-powered digital coaching represents a scalable, evidence-based intervention for overweight and obese populations. Healthcare systems should consider subsidizing such programs as preventive care, given the favorable cost-effectiveness ratio.
“The data is clear. AI coaching is not a lesser alternative—it is an equivalent, accessible, and affordable pathway to sustained health improvement.”
— Dr. James Park, Principal Investigator
