Predictors of loss to follow-up among men with tuberculosis in Puducherry and Tamil Nadu, India

Thomas J. Zhou, Subitha Lakshminarayanan, Sonali Sarkar, Selby Knudsen, C. Robert Horsburgh, Muthuraj Muthaiah, Carolyn K. Kan, Padmini Salgame, Jerrold J. Ellner, Gautam Roy, Helen E. Jenkins, Natasha S. Hochberg

Research output: Contribution to journalArticlepeer-review

5 Scopus citations


Identifying predictors of loss to follow-up (LTFU; treatment lapse 3 2 months) among people with tuberculosis (TB) may assist programmatic efforts in controlling the spread of TB. Newly diagnosed smear-positive TB patients were enrolled in the Regional Prospective Observational Research for TB study in Puducherry and Tamil Nadu, India. Treatment records were used to identify LTFU of those who were enrolled from May 2014 through December 2017. This nested case–control study evaluated male TB patients. Predictors were assessed using multivariable logistic regression. Of 425 men with TB, 82 (19%) were LTFU. In the adjusted analyses of males, divorced/separated marital status (adjusted odds ratio [aOR] 3.80; 95% CI: 1.39–10.38) and at-risk alcohol use (aOR 1.92; 95% CI: 1.12–3.27) were significant predictors for increased risk of LTFU, and diabetes was a significant predictor for decreased risk of LTFU (aOR 0.52; 95% CI: 0.29–0.92). Of 53 men with recorded date of last treatment visit, 23 (43%) and 43 (81%) had LTFU within the first 2 and first 4 months of treatment, respectively. Addressing at-risk alcohol use and providing more intensive follow-up could lead to improved treatment completion.

Original languageEnglish (US)
Pages (from-to)1050-1056
Number of pages7
JournalAmerican Journal of Tropical Medicine and Hygiene
Issue number3
StatePublished - Sep 2020

All Science Journal Classification (ASJC) codes

  • Parasitology
  • Infectious Diseases
  • Virology


Dive into the research topics of 'Predictors of loss to follow-up among men with tuberculosis in Puducherry and Tamil Nadu, India'. Together they form a unique fingerprint.

Cite this