Our Take
A screening tool that works with data already captured at diagnosis is useful, but the study doesn't show whether identifying high-risk patients actually changes outcomes or survival rates.
Why it matters
TB kills 7.4% of diagnosed patients within a year in India (per this study), with most deaths in the first two months. Early triage could direct limited resources to the sickest patients, but only if clinics act on the prediction.
Do this week
TB programme leaders: pilot the calculator at your largest public facility before June 2027 so you can measure whether flagging high-risk patients leads to faster treatment intensification or reduced mortality.
A simple TB death predictor ships from Indian researchers
Researchers at India's ICMR-National Institute of Epidemiology (NIE) in Chennai, working with the Tamil Nadu TB programme, developed a calculator that predicts death risk at the moment of TB diagnosis. The tool uses five clinical measurements: body mass index, oxygen saturation, respiratory rate, pedal oedema, and ability to stand without support. Combined with age, sex, disease site, prior treatment history, and lab confirmation, the model identifies high-risk patients.
The study analysed 55,971 adult TB patients notified across Tamil Nadu between July 2022 and June 2023 (per the published data). Overall, 7.4% died within one year of diagnosis. Nearly 68% of those deaths occurred within the first two months. The calculator performed almost as well as more complex models that rely on data captured later through India's Ni-kshay TB database.
The researchers recommend that severe-illness indicators be routinely captured at diagnosis and have developed a prospective calculator for use at that point. The work was published in BMJ Open and led by Suseendar Shanmugasundaram, Hemant Deepak Shewade, and Manoj V Murhekar.
Early screening is only useful if it changes care
A tool that identifies risk using data already available at diagnosis has operational appeal. It requires no new infrastructure, no delays, no expensive tests. In a resource-constrained setting, knowing which 7.4% of patients are at highest risk could theoretically direct palliative care, intensive monitoring, or urgent interventions to those most likely to die soon.
But the study does not report whether identifying high-risk patients with this calculator actually improves outcomes. It shows the calculator works as a predictor. It does not show what happens next. Without evidence that clinics acting on the prediction reduce deaths, the tool remains a diagnostic curiosity rather than a clinical intervention. A screening tool that sits unused changes nothing.
India accounts for 27% of the global TB burden, per WHO estimates. Reducing mortality remains a stated public health priority despite improvements in diagnosis and drug access. If this calculator is deployed and used to triage care, it could matter. If it becomes another number in a patient file, it won't.
How to pilot and measure impact
If you run a TB programme, the next step is a pragmatic trial. Deploy the calculator at one or two large public facilities. Randomise half the patients to have their risk score flagged to the clinician at diagnosis; leave the control half unscored. Measure whether the flagged cohort receives different care (more frequent follow-up, additional imaging, earlier drug-resistance testing, palliative services) and whether mortality differs between groups over 6–12 months.
Without that comparison, you cannot separate the value of the tool from the effect of awareness or facility-level improvements that might have happened anyway. Adoption without measurement is bureaucratic cover, not public health.