From quality measurement to improvement action: A stakeholder-integrated mixed-methods implementation analysis of adult outpatient care in Kenya

Quality measurement has limited improvement value when patient, provider and manager evidence is collected without an accountable route to action. This study examined how a Stakeholder-Integrated Quality Measurement-to-Improvement Framework (QMIF) linked role-specific outpatient quality evidence to short-cycle improvement in Kenyan Level IV facilities. A convergent mixed-methods implementation analysis was embedded in a 26-facility non-randomised, non-equivalent pretest-posttest quasi-experiment in Kisumu and Nyeri Counties (13 interventions; 13 comparisons). Baseline was collected on 10-15 April 2026, structured QMIF exposure ran from 15 April to 2 July 2026, and endline was collected on 3-5 July 2026. The longitudinal panel comprised 339 participants (260 patients, 54 providers and 25 managers) and the qualitative strand comprised 84 interviews. Theoretical interpretation was informed by Knowledge-to-Action (KTA), the Consolidated Framework for Implementation Research (CFIR) and RE-AIM. Mixed-effects ANCOVA with a facility random intercept provided baseline-adjusted inferential context for the primary perceived-quality outcome; difference-in-differences (DiD) and Hedges g were supportive. Adjusted effects favoured QMIF for patients (+21.4 percentage-of-maximum-possible [POMP] points; 95% CI 17.8-25.0), providers (+20.4; 16.3-24.6) and managers (+16.1; 11.9-20.4); all p<.001. Corresponding Hedges g values were 1.64, 1.87 and 2.13. Knowledge-attitude-practice (KAP) and reported quality-improvement (QI) action were retained as descriptive implementation indicators: intervention-arm endline KAP was 84.3%, 84.8% and 83.7% for patients, providers and managers, while QI-action POMP was 63.0%, 63.8% and 61.7%, respectively. Framework Method analysis and mixed-methods joint displays suggested a recurrent pathway of measurement, feedback, prioritisation, PDSA action and remeasurement. Sixteen of 18 stakeholder-domain combinations met the study's pragmatic criterion, while provider reliability (3.98/6) and manager appropriateness (3.96/6) remained boundary cases; medicines, diagnostics, staffing, equipment and procurement remained structural constraints. The combined evidence supports a plausible implementation pathway, but formal mediation, moderation and fidelity-outcome analyses were not performed. The study concludes that stakeholder-integrated measurement can support short-cycle outpatient QI while unresolved structural constraints require escalation; it recommends replication with stronger fidelity measurement, objective service indicators, external psychometric validation and longer follow-up.