Identification of Predictors (Risk Factors) of Multidrug-Resistant Tuberculosis (MDR-TB) in the Amravati region using Logistic Regression Analysis

Abstract

Multidrug-resistant tuberculosis (MDR-TB) is a global public health threat and burden on the health system. Facilities for its diagnosis and treatment are limited in many high-burden countries, including India. This study aims to determine predictors (risk factors) associated with MDR-TB among the new and old TB patients from the Amravati region to provide future guidance. A cross-sectional study was conducted on all new and old TB patients registered under DOTS in the Amravati region from May 2012 to December 2019. Complete information on 402 patients, which included treatment details and other relevant clinical data, was used in this study. Binary logistic regression analysis is employed to determine the predictors (risk factors) for MDR-TB. The significant predictors for MDR-TB found are ‘Number of times TB treatment taken in past’, ‘Positive follow-up of at least 4 months treatment or Relapse/defaulter from previous DOTS', and MDR-contact, HIV positive, irregular TB patients, CBNAT-detected TB patients, whereas the factor ‘number of past outcomes-type’ did not predict MDR-TB. The study shows that the MDR contact and CBNAT test results are the strongest predictors of MDR-TB. These factors among newly diagnosed TB patients should be taken as an alarm for the screening of MDR-TB.  Mantel-Haenszel approach is also discussed in the study.