Global health estimates lack a single statistical playbook
Global estimates of health indicators sit at the center of international target-setting, yet the statistical machinery behind them remains uneven. For policymakers, these figures often function as the numerical backbone of progress reports, burden comparisons, and accountability frameworks. For methodologists, the same estimates raise a more basic problem: there is no single standard framework for how they should be built.
This BMJ Open protocol lays out a scoping review designed to map the conceptual and statistical approaches used to produce those estimates. The investigators are not trying to pool outcomes or make a clinical efficacy claim. Their task is more structural and, in practice, just as consequential: identify how global health estimates are constructed, where methods converge, and where they diverge enough to threaten comparability across indicators and regions.
One short sentence captures the problem plainly: global estimates are only as credible as the assumptions buried inside them.
Why a checklist now matters for global health metrics
The protocol builds on an already published checklist intended to guide researchers through key conceptual and statistical considerations for global estimates. That checklist is meant to work alongside the Guidelines for Accurate and Transparent Health Estimates Reporting, a framework aimed at making health estimates easier to interpret and audit.
The need for this work is easy to understand from a measurement standpoint. Global health indicators are rarely observed directly in a single clean data stream. They are often assembled from mixed data sources, modeled across countries and time periods, and adjusted to account for missingness, sparse surveillance, or differences in case definitions. Each of those steps can change the final estimate. In effect, the uncertainty does not begin at the end of the analysis; it starts at the point where the input data are selected and weighted.
A scoping review is a sensible design for this question because the field appears to contain many methods rather than one dominant approach. The goal is breadth, not narrow effect estimation. By cataloging methodological choices across published studies, the review aims to show whether the checklist corresponds to what investigators actually do and whether the field has any consistent statistical habits worth formalizing.
How the review will identify and organize methods
The protocol follows the JBI framework for scoping review protocols. Searches will run across EMBASE, Medline, PubMed, Scopus, Web of Science, and the WHO Global Health Estimate Methods webpage. Eligibility is limited to English-language studies, with no date restriction, which should allow the review to capture both older methodological papers and more recent estimate-building strategies.
Two independent reviewers will screen the literature and extract data. That dual-review structure is important in a scoping exercise because methodological descriptions are often incomplete, and the same analytic step can be described in different language across papers. Independent assessment reduces the chance that a single reviewer’s interpretation of a model, adjustment step, or extrapolation procedure determines whether a paper is included or how it is classified.
The planned extraction framework will follow previously published work that identified key elements and checklist items for methodological and statistical approaches used in the development of global health indicators. In practical terms, this should allow the review to map how studies handle core decisions such as data sources, model selection, uncertainty quantification, and other design choices that shape the final estimate. Once those features are extracted, the findings will be summarized through narrative synthesis, descriptive statistics, and visual plots where appropriate, using the PRISMA-ScR checklist.
Statistical heterogeneity is the real target of the review
The most important issue in this protocol is not whether one estimation method is universally best. It is whether the field has developed a coherent language for comparing methods that are often used for similar policy purposes but may rest on very different assumptions. A global prevalence estimate, a mortality estimate, and an indicator tracking coverage toward a target may all require different modeling logic. Even within the same indicator family, researchers may treat missing data, regional extrapolation, and temporal trends in incompatible ways.
That heterogeneity has direct policy consequences. When estimates are used to assess progress toward global goals, the apparent trend can reflect methodological change as much as real-world change. A revised model may improve statistical plausibility while making year-to-year comparisons harder to interpret. A dataset with wider geographic coverage may seem more precise but still rely on assumptions that are not obvious to the end user. In that setting, transparency is not a cosmetic issue; it is the only way decision-makers can judge whether the estimate is fit for purpose.
The scoping review also has a secondary value: it may reveal where the checklist is strongest and where it still leaves open interpretive gaps. If the literature repeatedly omits certain decision points, the checklist may need refinement. If most studies address the same methodological domains but label them differently, the field may need standard terminology more than new statistics.
Publication plans and the limits of the protocol design
No ethical approval is required because the review will analyze published data only. The dissemination plan is conventional but appropriate: peer-reviewed publication and conference presentation for audiences working in statistics and health research.
As a protocol, this paper does not yet produce findings from the literature itself. It establishes the boundaries of the review and the analytic approach, not the conclusions. That matters because protocols can sometimes be mistaken for completed reviews. Here, the value lies in predefining the question with enough specificity to support a credible map of the evidence once the search is complete.
The registration on the Open Science Framework adds another layer of transparency, since it documents the project plan in advance. For a field built on indirect estimates and model-based inference, that sort of registration is not administrative window dressing. It is part of the same credibility chain the review is trying to assess in the literature.
What remains unresolved is not whether global health estimates are useful; they clearly are. The harder question is whether their statistical foundations are sufficiently standardized to support confident cross-study and cross-country comparisons. This scoping review is designed to answer how the field actually works, and whether the tools now in circulation are mature enough to carry the weight placed on them by global health reporting.