Effects and costs of predictive risk stratification in primary care: randomised stepped wedge trial

Background The ageing UK population needs safe approaches to reduce emergency hospital admissions. Predictive risk stratification modelling (PRiSM) estimates risk that individuals will suffer emergency admission to hospital within 12 months and selects patients for preventative community care to avoid admissions. Aim To evaluate the introduction of (PRiSM) into primary care. Method Funded by NIHR, we used randomised stepped wedge design to estimate (cost) effectiveness of introducing PRiSM software into 32 participating practices in urban South Wales, supported by practice-based training, clinical support through two local ‘GP champions’, and technical support through telephone help-desk. Outcome measures included: emergency hospital admissions (primary), other hospital activity, and GP activity, all estimated from routine data; patient-reported SF-12 health-related quality of life scores; and NHS costs. Results Across 230,000 participants, PRiSM implementation increased: emergency hospital admission rates by 1.1% (95% confidence interval [CI] = 1.0% to 1.3%); Emergency Department attendance rates by 3.0% (95%CI = 2.8% to 3.2%); outpatient visit rates by 5.5% (95%CI = 5.1% to 5.8%); GP activity by 1.1% (95% CI = 0.7% to 1.4%); and NHS costs per patient by £76 (95%CI = £46 to £106). Questionnaires completed by 1400 randomly sampled participants showed that: PRISM improved SF-12 physical scores by 1.5 points (95%CI = 0.8 to 2.2); but not SF-12 mental scores (95%CI = −1.5 points to +0.3). The direct cost of introducing PRiSM was £0.11/patient/year. Conclusion The introduction of PRiSM increased emergency hospital admissions and other NHS activity without clear evidence of benefit.