Science · Health research · TB and diabetes
TB and diabetes: reading the screening numbers
Nine molecular-test positives in one Khayelitsha clinic underpin the findings.

Ahead of World Diabetes Day on 14 November, a Cape Town study offers a lesson in reading health research: a test's percentage needs a denominator, a setting and a clear purpose.
The PLOS Global Public Health paper, published on 30 September (opens in a new tab), used 2022 data from adults attending diabetes care at Site B Community Health Centre in Khayelitsha. Researchers enrolled 673 people; 661 had analysable results.
One sputum sample was tested with Xpert Ultra regardless of symptoms or X-ray findings. Nine were positive, including two trace results.
Two of the nine reported any cough. Including fever, night sweats or weight loss identified the same two. Clinician-read TB-suggestive X-rays identified five of the nine.
That makes the cough-screen sensitivity 22.2%: two divided by nine. The denominator is the study's molecular-test-positive group. This percentage cannot tell you your personal chance of having TB.
Why the limits matter
Only nine positives leave wide uncertainty. Participants were already in care and may have had earlier symptom screening. Ultra can detect DNA from dead bacteria, the authors warn. Culture was not the study's reference. The findings cannot establish current clinic failures or describe everyone with diabetes.
Screening starts a process
The WHO's screening explanation, updated on 4 November (opens in a new tab), says screening identifies people with a higher likelihood of TB who need further investigation. A positive screening result alone does not confirm disease.
WHO lists diabetes among risk factors for which screening in healthcare settings may be considered. Its available tools include symptom questions, chest X-rays and molecular rapid tests. The combination depends on the population and the tests available.
Molecular tests appearing on that list matters. The distinction between screening and diagnosis is about a test's purpose within a pathway. It cannot be reduced to symptom questions on one side and laboratory equipment on the other.
South Africa's October 2026 Adult Primary Care guide (opens in a new tab), written for healthcare professionals, places symptoms and risk history alongside molecular testing. Its TB diagnostic pathway distinguishes positive, trace and negative results and considers previous TB treatment and further investigations. It is a clinical decision tool, not a self-diagnosis checklist.
Our comparison of the documents shows why assessment, test interpretation and follow-up belong together. A screening percentage on its own cannot describe that whole process.
The 2026 World Diabetes Day campaign (opens in a new tab) stresses early diagnosis linked to accessible care. For TB screening, the service question is which approach can find disease reliably and provide follow-up at an affordable cost. Larger studies are needed before these findings could support a universal approach.
Sources, evidence and corrections
- Sensitivity and specificity of tuberculosis screening tools in people with diabetes (opens in a new tab) · 2026-09-30
- Tuberculosis: Systematic screening (opens in a new tab) · 2026-11-04
- Adult Primary Care, Clinical Decision Support Guide, South African Edition (opens in a new tab) · 2026-10
- World Diabetes Day 2026 (opens in a new tab) · 2026 campaign, accessed 13 November 2026 SAST
