Web tool recommends personalized TB treatment and flags side effects

Tuberculosis remains a leading cause of death in underdeveloped and developing countries, and managing drug-resistant TB is particularly difficult. Limited access to detailed drug susceptibility testing, lack of knowledge about second-line anti-TB drugs, poor adherence to drug dosages, and the presence of other health conditions all complicate treatment. The number of TB deaths has been rising due to factors such as high levels of resistance mutations in TB strains, delays in delivering treatment, and the absence of personalized medicine approaches. To address these challenges, Sintu Kumar Samanta and colleagues developed a web-based application designed to aid personalized treatment decisions for TB patients. The system is intended to work with the information that clinicians and patients can provide: geographical location, age, sex, medical and travel history, an AMR report, and associated medical conditions (AMCs). Based on those inputs, the application aims to output a list of drugs that are safe for the individual patient and to highlight possible side effects from drug combinations, helping care teams consider both TB treatment and coexisting conditions when making decisions.

The team built the Personalized Tuberculosis Treatment Recommendation System (PTTRS) version 1 as a web-accessible decision aid. PTTRS accepts patient-level inputs — geographical location, age, sex, medical and travel history, AMR report, and associated medical conditions (AMCs) — and then generates an output listing drugs that are safe for that specific patient profile. The application also reports the possible side-effects of various drug combinations used to treat TB, and it provides side-effect information for combinations involving any AMC the patient may have. In addition to PTTRS, the developers offer a polypharmacy side effect predictor that is not limited to tuberculosis and can be used for other diseases. The PTTRS application is available at https://pttrs-bblserver.streamlit.app/, and the polypharmacy side effect predictor can be accessed at https://psep-bblserver.streamlit.app/. Together, these tools are aimed at helping clinicians and patients anticipate harmful interactions and tailor treatment choices to individual medical and antimicrobial resistance profiles.

If adopted in clinical settings, PTTRS could help bridge gaps that currently hinder care in resource-limited areas. By combining AMR reports and patient medical histories, the system may support more timely and personalized treatment decisions, helping to address problems such as limited access to detailed susceptibility testing and gaps in knowledge about second-line anti-TB drugs. The feature that flags side effects of drug combinations is particularly relevant in the context of comorbidities and polypharmacy, where interactions can worsen outcomes. While the abstract does not report clinical trial results or validation metrics, the described application and the polypharmacy side effect predictor suggest a practical route for clinicians to review safer drug options and anticipate adverse effects before prescribing. Making these tools web-accessible also aims to increase availability in underdeveloped and developing nations where the burden of drug-resistant TB is highest, potentially supporting faster, more individualized care decisions that respond to local AMR patterns and patient needs.