Why in the News?

Tamil Nadu has become the first Indian state to integrate a TB death prediction model into its State TB Elimination Programme, enabling early identification and hospitalisation of high-risk patients to reduce tuberculosis-related mortality.

What’s in Today’s Article?

• TB Death Prediction Model (Introduction, Working Mechanism, Validation, Significance, Impact, etc.)

Introduction

• In a landmark step towards eliminating tuberculosis (TB), Tamil Nadu has become the first state in India to deploy a predictive model that estimates the likelihood of death in TB patients.
• Integrated with the state’s existing digital triage platform TB SeWA, this model is designed to enable faster hospital admissions for severely ill patients, ultimately reducing TB-related mortality.
• This innovation is a collaborative outcome of the Indian Council of Medical Research’s National Institute of Epidemiology (ICMR-NIE) and the Tamil Nadu State Health Department under the larger framework of Tamil Nadu Kasanoi Erappila Thittam (TN-KET).

The Predictive Model and How It Works

• The newly launched predictive model uses five clinical indicators at the time of TB diagnosis:

• Body Mass Index (BMI)
• Presence of pedal oedema (swelling of feet)
• Respiratory rate
• Oxygen saturation levels
• Ability to stand without support

• Healthcare workers input these variables into the TB SeWA application. Based on this input, the model calculates the probability of death ranging from 10% to 50% for severely ill patients.
• For those not flagged as severely ill, the predicted mortality risk remains between 1% and 4%.
• This sharp differentiation in risk estimation helps frontline healthcare staff prioritise admissions and initiate early treatment, which is especially crucial given that over 70% of TB deaths occur within the first two months of treatment.

Significance of the Integration

• Prior to this model, Tamil Nadu's TB SeWA system helped identify severely ill patients using the five indicators, enabling timely inpatient care.
• The integration of a quantified probability of death now offers an objective assessment of risk, improving decision-making at the primary health level.
• The team at ICMR-NIE noted that while the average time from diagnosis to hospital admission is one day in Tamil Nadu, about 25% of severely ill patients face delays of 3–6 days. The new model is expected to reduce such delays.

Development and Validation of the Model

• The model was developed using data from nearly 56,000 TB patients diagnosed across Tamil Nadu between July 2022 and June 2023.
• It was observed that 10–15% of adults diagnosed with TB were classified as severely ill at the time of diagnosis.
• The model's validation has demonstrated that the five triage variables used in TN-KET are just as predictive of mortality risk as the comprehensive baseline variables in the national Ni-kshay TB portal.
• However, Ni-kshay variables typically take up to three weeks to populate, too late to act upon for high-risk patients. In contrast, the TN-KET system captures triage data within a day, ensuring faster action.

Broader Public Health Impact

• All 2,800 public health facilities in Tamil Nadu, from Primary Health Centres to Medical Colleges, currently use the TB SeWA application. The model supports:

• Real-time triaging
• Objective risk stratification
• Timely hospital referrals

• The success of TN-KET and its associated tools has already contributed to reduced loss in the TB care cascade across two-thirds of Tamil Nadu’s districts.
• This innovation serves as a replicable model for other Indian states, where early TB deaths remain a significant challenge despite free diagnosis and treatment.

Global and National Context

• According to the World Health Organisation, TB remains one of the top causes of death globally.
• India bears the highest burden of TB in the world, with two deaths every three minutes.
• A recent study titled “Time to Death and Associated Factors among Tuberculosis Patients in Dangila Woreda, Ethiopia” identifies old age, low body weight, and TB/HIV co-infection as significant predictors of early mortality.
• Tamil Nadu’s model, by addressing similar risk factors early, aligns with global recommendations for reducing TB deaths.