October 6, 2026 interva.net Editorial

Seven Countries on How to Teach a Doctor to Write a Cause of Death

A qualitative study in Global Health Action interviewed health officials in Bangladesh, Colombia, Ethiopia, Kenya, Rwanda, Sri Lanka and Tanzania about embedding medical certification training in medical schools and licensing — the physician-written track that runs alongside verbal autopsy.

cause of deathCRVSdata qualityglobal health

Nearly 40% of the world’s deaths are never registered, and in low-income countries only 8% of the deaths that are reported carry a documented cause. WHO’s civil registration fact sheet puts the spread starkly: registered deaths range from 98% of the total in the European Region to 10% in the African Region.

Where deaths are neither registered nor medically described, verbal autopsy is the substitute — an interview with a family, run through a probabilistic model. But in places where a physician does sign the certificate, the quality of the resulting mortality statistics depends on something less discussed: whether that physician was ever taught how to fill the form. A study published in Global Health Action looks at how seven countries tried to make that teaching permanent.

Nine interviews, seven systems

The paper comes from Rania A. Khan, an independent consultant, with colleagues from Vital Strategies and from ministries and medical councils in the countries studied. Its method is deliberately narrow. Seventeen countries participating in the Bloomberg Philanthropies Data for Health Initiative were approached; seven answered affirmatively to at least one of two questions — whether a formal order required medical certification of cause of death (MCCD) in pre-service medical education, or in continuing medical education. Bangladesh, Colombia, Ethiopia, Kenya, Rwanda, Sri Lanka and Tanzania made the cut.

Data collection took place in 2024 through virtual semi-structured interviews, mostly with Data for Health country coordinators, totalling nine participants, with an additional round of country partners consulted to triangulate findings. Analysis was thematic, interpreted through a published health-institutionalization framework.

That design sets the evidence level clearly. This is a qualitative account of how policy processes unfolded, from a small and purposively selected group, drawing on self-reported experience. The authors state that members of the research team had supported certification training in the countries they were studying, and warn that this positionality may have shaped interpretation. The paper reports no measure of certification accuracy, no error rates, and no mortality outcomes.

What made training stick

Across the seven accounts, the study’s five cross-cutting themes — government commitment, stakeholder coordination, international partnerships, training integration and professional-development incentives — rarely operated alone. Tanzania’s health ministry began its certification improvement work in 2015 after concluding that national mortality data did not align with WHO standards. Rwanda’s Ministry of Health took ownership early, embedding the topic in its National CRVS Work Plan for 2017–2023. Kenya’s regulator, the Kenya Medical Practitioners and Dentists Council, led first and was later joined by county-level authorities.

Pre-service integration looked different where medical schools set their own curricula. Bangladesh moved through its medical and dental council, which issued a circular in 2017 adding a certification module to a curriculum that all 107 medical schools follow. Tanzania brought universities and the national technical-education council on board from 2017; Rwanda made certification a core component of medical education by 2020; Sri Lanka developed a single curriculum implemented by every medical faculty. Colombia is the counter-example the authors give: university autonomy let schools design their own programmes, and without a binding requirement, ministry support alone did not produce integration.

The lever that turned out to matter most

If one finding recurs with the force of a practical lesson, it is that voluntary training and mandatory training behave differently. Doctors in several countries first treated the certificate as extra administrative work; professional bodies and hospitals had to be brought in to build acceptance.

The countries that tied certification training to licensure saw compliance move. Rwanda’s accredited e-learning course, required for annual licence renewal, was made to work alongside a centralised governance structure and high internet penetration — conditions the authors note may not travel. Kenya’s council is moving toward mandating the training and plans to attach up to 10 continuing-professional-development credits to it. Ethiopia’s module carries a generous 19 credits but remains optional, and compliance is low, especially in rural areas with few trained staff. Tanzania’s course is likewise not yet a licence requirement; in Sri Lanka, certification training sits inside the professional-development framework but credits are not required for renewal, softening the incentive. In Colombia, where the virtual course is optional for practising doctors, about half of those who start it finish.

Against that, the study sets the infrastructure problem. E-learning was used in all seven countries but behaved differently according to connectivity and platform design; Tanzania’s early platform struggled to link course completion to credit tracking. Sri Lanka leaned on face-to-face teaching and, during the economic crisis, distributed roughly 20,000 printed certification guidebooks to hospitals. Ethiopia’s need to translate materials into multiple local languages adds cost and logistics the authors single out as an equity question.

Cascade training appears throughout. Tanzania launched a national training-of-trainers model in 2016 using clinicians from teaching hospitals; Rwanda trained master trainers who then instructed colleagues inside their own institutions; Sri Lanka trained approximately 125 master trainers in person and expanded to 25 hospitals across nine provinces, each writing its own rollout plan. The authors flag the obvious risk — cascade quality depends on supervision and refresher training, which were uneven.

The half of the problem this paper does not touch

The study’s subject is the certified track, where a doctor is present at death. Most deaths in the settings this digest covers are not on that track at all. There, the WHO verbal autopsy instrument and its coding methods carry the load, and the systems that pair nurse- or officer-led interviews with official registration remain the focus of much of the operational literature — Zambia’s digitised death notification and its parallel certification-coded programme being a recent example covered in this digest.

The paper’s authors are careful about what training can deliver on its own. Training improves certification in their account, but better data does not automatically become better decisions: they point to political will, institutional capacity to use the data, and policy environments in which mortality evidence is not deprioritised. Sustainability is stated as an open question rather than a conclusion — external funding, here from Bloomberg Philanthropies, starts these programmes; domestic budget lines are what keep them. Whether donor-supported institutionalisation lasts is, the authors write, a matter for a follow-up study that has not yet been done.

For readers tracking the measurement side of global health, the value of this paper is process detail: which institutions signed which circular, which regulator dangled which credit, which system turned out to be too decentralised to move. It is not a validation study, and nothing in it establishes that the certified causes these systems produce are correct. Certifying a death accurately and interviewing a family about one are different instruments, used in different places, and each has its own documented blind spots. This digest reports on both; it does not advise ministries on which to adopt, and it is not affiliated with the InterVA project, Umeå University or WHO.

Sources

← Back to the news digest