Skip to main content

Resources

Research & Publications

We do not commission our own evaluations and we do not hold sign-off on the findings. What follows includes the paper about the programmes we closed.

Peer-reviewed

Published papers

Each entry states the finding in plain language, because a title alone tells a non-specialist reader nothing about what was learned.

  • Randomised trial Open access 2026

    Community health worker home visits and under-five mortality: a cluster-randomised trial across six districts

    Okonjo A, Mwangi J, Adeyemi G, et al. · Journal of Global Primary Care

    What it found

    Door-to-door screening by trained, paid community health workers reduced under-five deaths by 31% over four years against matched control districts.

  • Economic evaluation Open access 2025

    Cost per outcome in integrated versus vertical primary health programmes

    Whitcombe H, Menon R, Bergström S. · Programme Economics Quarterly

    What it found

    Integrated delivery cost 18% more to establish and 34% less per outcome after year four, driven almost entirely by reduced repeat presentation.

  • Cohort study Open access 2025

    Retention of clinical staff following handover to national health systems

    Bekele T, Diallo F, Kimani J. · International Journal of Health Workforce

    What it found

    Staff retention two years post-handover was 79% where posts were nationally salaried from the outset, and 41% where they were not.

  • Controlled study Open access 2024

    Psychological first aid delivered by non-specialists in acute displacement

    Delacroix H, Tanaka Y. · Humanitarian Health Review

    What it found

    Non-specialist delivery matched specialist outcomes at three months on two of three measures, and was available roughly nine weeks sooner.

  • Field evaluation 2024

    Solar-direct-drive cold chain performance in off-grid clinics: a 36-month field evaluation

    Mwangi J, Raghunathan P. · Journal of Field Immunisation

    What it found

    Temperature excursions fell by 62% against generator-dependent refrigeration, with maintenance cost lower after month fourteen.

  • Retrospective analysis Open access 2023

    Why three programmes were closed: a retrospective analysis of programme failure

    Bergström S, Okonjo A. · Global Programme Practice Quarterly

    What it found

    All three closures shared one feature: the programme was designed around available funding rather than an assessed local need.

Open data

The underlying data

Published so our conclusions can be checked rather than taken on trust. Patient data is anonymised at source and never leaves the country it was collected in; what is published here is aggregated beyond re-identification.

  • Community health worker trial, anonymised

    Household-level visit and outcome data across six districts, 2021–2025.

    1.4M rows CC BY 4.0 Download CSV · 1 KB
  • Clinic activity, monthly

    Attendance, referral and stock-out counts for every clinic we operate.

    412 sites CC BY 4.0 Download CSV · 5 KB
  • Cost-per-outcome model

    The spreadsheet behind the economic evaluation, with assumptions exposed.

    Workbook CC BY 4.0 Download CSV · 1 KB

We do not mark our own homework

Evaluations are carried out by independent academic partners who choose their own methods and publish their own conclusions. We see the findings when the journal does. Where a study has shown our work to be ineffective, that study is on this page.