New Delhi: A web-based calculator developed by the Indian Council of Medical Research’s National Institute of Epidemiology (ICMR-NIE) can predict a tuberculosis (TB) patient’s risk of dying at the time of diagnosis, using just a handful of clinical measurements.
Researchers from ICMR-NIE say the death prediction tool helped reduce TB deaths in Tamil Nadu. Despite the results, the calculator has not yet been adopted by the Central TB Division (CTD), the central authority responsible for managing and implementing India’s flagship public health initiative aimed at eliminating the disease, the National Tuberculosis Elimination Programme.
The researchers also cautioned that the calculator has not yet been externally validated outside Tamil Nadu.
The calculator, described in a study published last week in the British medical journal, The BMJ Open, estimates the probability of early TB death within two months of diagnosis, and overall death within a year, using routinely collected clinical information available during the patient’s first assessment.
Researchers say the approach is designed for resource-constrained settings where laboratory investigations may not be immediately available.
Tuberculosis, caused by a bacterium called Mycobacterium tuberculosis, remains one of the world’s deadliest infectious diseases despite being preventable and curable.
India continues to carry the world’s highest TB burden, with more than three lakh TB-related deaths reported every year, according to the Global Tuberculosis Report 2024. Nearly 70 percent of these deaths occur within the first two months of diagnosis, making early identification of severely ill patients critical.
“This initiative has shown that it is possible to bring down TB-related deaths remarkably by following scientifically designed tools and methods. Severe illness can be quickly identified through triaging (preliminary assessment), and patients can be promptly admitted after diagnosis,” Dr Hemant Deepak Shewade, senior medical scientist at ICMR-NIE and author of the study, told ThePrint.
Also Read: A US man lived two days without lungs. Why it can revolutionise transplant practice
Developed from Tamil Nadu’s mortality-reduction programme
The calculator emerged from Tamil Nadu Kasanoi Erappila Thittam (TN-KET), a statewide initiative implemented between 2022 and 2025 that used a differentiated TB care approach, identifying patients at high risk of death and providing them with more intensive monitoring and treatment to reduce TB mortality.
Implemented in routine health system settings by the Tamil Nadu State TB Cell with support from key directorates under the state’s Department of Health and Family Welfare, the initiative covered all 38 districts of the state except Chennai.
Under TN-KET, frontline health workers triaged every newly diagnosed adult TB patient using five simple clinical measurements at diagnosis. These included body mass index (BMI), pedal oedema (fluid accumulation in the feet and ankles), respiratory rate, oxygen saturation, and the patient’s ability to stand without support. Patients identified as severely ill were prioritised for hospital admission and closer monitoring.
The TB Death Prediction Calculator used data from 55,971 adults diagnosed with TB in the state’s public health system between July 2022 and June 2023. Overall mortality was 7.4 percent, with 67.9 percent of deaths occurring within the first two months of treatment.
Patients flagged as severely ill had a mortality rate of 19.8 percent, compared with 5.1 percent among those not classified as severely ill.
The initiative also translated into measurable improvements on the ground. Dharmapuri’s TB death rate fell from 12.5 percent to 7.8 percent, Karur’s from 7.1 percent to 5.3 percent, and Villupuram’s from 6.1 percent to 5.2 percent.
An earlier TN-KET analysis also found that within six months of its rollout in April 2022, overall TB deaths declined by nearly 10 percent while early TB deaths fell by almost 20 percent.
After two years of implementation, in 2024 and 2025, a sustained statewide reduction of at least 30 percent in the TB death rate was observed.
How is it different from Ni-kshay?
India’s national TB programme already has a differentiated care module within Ni-kshay, the government’s digital TB patient management and surveillance platform. However, researchers argue that the existing risk assessment is difficult to use prospectively because it relies on 16-20 variables, many of which require laboratory investigations.
By contrast, the new calculator uses only five easy-to-infer clinical measurements, along with basic demographic and disease information such as age, sex, TB site, previous treatment history and microbiological confirmation, to generate an individual’s probability of death at diagnosis.
The study found that a model based only on five simple clinical measurements used to identify severely ill patients performed almost as well as one that used all the patient information routinely recorded in Ni-kshay at the time of diagnosis.
Adding a few more patient characteristics available at diagnosis further improved its accuracy, while including the remaining information from Ni-kshay did not significantly improve prediction.
“The current risk stratification within the Ni-kshay Differentiated TB Care module depends on data that are often unavailable at diagnosis. Seventy percent of TB deaths occur early. We need an actionable assessment at diagnosis before patients are sent back to peripheral centres for ambulatory treatment,” Shewade told ThePrint.
“Rather than assigning patients to broad risk categories, the calculator provides clinicians with an estimated probability of death, allowing them to prioritise admissions based on both patient risk and available hospital capacity,” he added.
Yet to find a place in the national programme
Despite support from ICMR, presentations to the World Health Organization and offers of technical assistance to other states, the calculator has not been adopted by the Central TB Division (CTD).
Dr Soumya Swaminathan, Principal Advisor to the NTEP, said reducing TB mortality remains a national priority, but the programme has not endorsed this specific tool.
“It is not adopted by all states. CTD has prioritised mortality reduction but not advocated for this particular tool. Lack of beds for infectious patients is a big issue in many states, so sick patients don’t get admitted,” she told ThePrint.
Shewade said the standalone calculator could help other states replicate the Tamil Nadu model without having to implement the entire TN-KET programme.
“We feel that scaling the model across India will be beneficial in bringing down the number of people dying every year due to the disease,” he said.
The TN-KET programme has now ended, and Tamil Nadu has gone back to using the national TB programme’s reporting system. However, many of the methods introduced under TN-KET to identify and manage seriously ill TB patients continue to be used across the state.
Study cautions
Researchers say one of the study’s biggest strengths is that it was built using routine programme data from nearly 56,000 adult TB patients treated across 30 districts in Tamil Nadu, making it broadly representative of patients seeking care in the public health system.
Because most TB deaths occur within the first two months of diagnosis, they argue that a prediction tool based on information available at the initial clinical assessment is more practical for real-world use than models requiring extensive laboratory data.
The authors, however, caution that the calculator should be interpreted with some limitations. About 17 percent of eligible patients were excluded because they were not triaged, and these patients differed in some clinical characteristics from those included in the analysis.
The model is designed only for use at the time of diagnosis and cannot predict changes in a patient’s risk later during treatment.
Most importantly, the researchers said that the calculator has not yet been externally validated outside Tamil Nadu.
“As this is the first statewide programme to capture these specific triage variables at diagnosis, no independent dataset was available for comparison. Future studies should evaluate the performance of these models in different regional contexts to ensure wider generalisability,” the researchers noted.
(Edited by Sugita Katyal)
Also Read: India gets first approved precision treatment for advanced prostate cancer. How Pluvicto works
It will catch up in the rest of the country after realising that cow urine does not work!
Sadly, even academics in premier institutes have joined the appeasement brigade