Welcome to the TB Death Prediction Calculator

TB death prediction calculator for use at TB diagnosis developed by ICMR-NIE Chennai, is based on predictive models developed using triage variables from the Severe TB Web Application (TB SeWA) of Tamil Nadu Kasanoi Erappila Thittam (#TNKET) and baseline characteristics from NTEP's Ni-kshay. Currently the models in the calculator are based on 55000 adult TB patient's data from Tamil Nadu (Jun 2022- Jul 2023 cohort). The triage variables' data that contributed to building the calculator are as a result of statewide implementation of #TNKET. As more data becomes available (Jul 2023- June 2024 cohort and later), the prediction equation will be updated in the backend from time to time to generate even more robust predictions.

The calculator provides predicted probabilities of early and overall TB-related death at diagnosis for an adult (>=15y) with TB (not known to be drug-resistant at diagnosis) based on variables available at the time of diagnosis. Focus is on predicting death at TB diagnosis itself as majority TB deaths are early and appropriate action regarding inpatient care for severely ill people with TB needs to be taken at diagnosis. After sending the patient to the nearest primary health care/ health subcenter facility for treatment, many people with TB and their family may refuse referral for admission.

Four models are available in the calculator. Model 1 uses all baseline characteristics (predictor variables) routinely captured in Ni-kshay. Model 2 is based solely on five triage variables captured in TN-KET: body mass index (BMI), pedal oedema, respiratory rate, oxygen saturation, and ability to stand without support. The predictive performance of model 2 is similar to model 1. Model 3 includes the five triage variables from TN-KET along with baseline characteristics routinely captured in Ni-kshay that are easily available at diagnosis (age, gender, previous history of TB treatment, site of TB, microbiological confirmation). Model 4 includes all the Ni-kshay baseline characteristics and all triage variables from TN-KET (model 1 plus 2).

Appropriate model can be used at TB diagnosis based on context. If there is commitment from states in India to capture the five triage variables like Tamil Nadu is doing, model 3 can be used at TB diagnosis even in resource constrained settings.