WHO’s tuberculosis data resources contain several kinds of values. A country-reported notification series and an estimate of incidence should remain distinguishable in a source table. Treating every value as the same kind of count can change the claim being made.
WHO separates country-provided data from its burden estimates and labels provisional notifications as subject to change. Its incidence metadata describes estimated new and relapse cases with uncertainty bounds. The source type and measure name belong beside each value.
Start with the data dictionary
Record the file, variable name, definition, unit, geography, period and release date. Keep any provisional flag and uncertainty fields. Do not infer a variable’s meaning solely from a short column heading or a nearby chart.
The WHO data page also explains that country data can be corrected after a report is written. Preserve the version used for a printed report separately from a later download when comparing them.
A comparison needs matched meanings
Before putting two values in one chart, inspect their case definitions, periods and geographical coverage. A yearly estimate and a partial-period provisional notification count do not become comparable merely because both concern TB.
If showing an estimate with bounds, keep the bounds and their source definition. Do not manufacture precision by displaying extra decimals or presenting a point estimate as a directly observed census.
Do not give a subtraction more meaning than its inputs
A numerical gap between two series does not by itself identify particular people or establish an exact count of undiagnosed disease. Any interpretation of detection, reporting or programme performance needs the applicable methods and evidence.
No TB dataset was downloaded or country outcome assessed here. The guide offers a provenance worksheet, without diagnosis, treatment advice or a claim about an individual’s infection. The practical objective is a chart in which readers can tell what was reported, what was estimated and which version was used.
Official source documentation read 21 October 2026. The worksheets are editorial reading aids; no population result, individual diagnosis or clinical recommendation is supplied.
Sources: Official indicator or dataset documentation; Official indicator or dataset documentation. Related discussion: health information source quality.
Leave a Reply