A cluster-randomized trial of household vs incentive-based tuberculosis contact investigation in rural South Africa

Implementation reach

Bibliographic Data

ID19593141
AuthorsEteri Machavariani (0000-0001-8138-2300, New York University, corresponding author), Samyra R Cox (0000-0001-9953-1693, Johns Hopkins University, corresponding author), Bareng A S Nonyane (0000-0001-5633-7487, Johns Hopkins University, corresponding author), Tinatin Gvelesiani (0009-0000-4941-3824, Tbilisi State Medical University, corresponding author), Neil Martinson (0000-0002-0184-5705, Perinatal HIV Research Unit, corresponding author), David Dowdy (0000-0003-0481-7475, Johns Hopkins University, corresponding author), David W Dowdy, Colleen F Hanrahan (0000-0002-4544-4591, Johns Hopkins University, corresponding author)
EditorsJoseph Baruch Baluku (0000-0002-5852-9674)
Year2025
Volume5
Issue12
Pagese0005580
Publication date2025-12-05
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenuePLOS Global Public Health (JOURNAL)
Journal identifiersISSN: 2767-3375 • E-ISSN: 2767-3375
PublisherPublic Library of Science (PLoS) (PUBLISHER)
DOI10.1371/journal.pgph.0005580
PMID41348846
OpenAlexW4417055030
LanguageEN
References cited30

The World Health Organization recommends TB contact investigation in high-burden countries. We examined the implementation reach of two contact investigation strategies in South Africa . Kharitode TB, a cluster-randomized crossover trial, compared household- and incentive-based contact investigation in 28 clinics (July 2016-January 2020). Clinics used each strategy for 18 months (separated by a six-month “washout”). Adults recently diagnosed with TB (index participants) were enrolled. In the household-based arm, contact persons were screened and provided sputum samples at home. In the incentive-based arm, index participants distributed referral coupons to their contacts, who received a $3.50 incentive upon presenting for screening at clinics. We used mixed-effects logistic regression with random intercepts for clinics to examine factors associated with index participant enrollment and sputum collection from contact persons. In the household-based arm, 782/1,269 (61.6%) index participants consented, 1,882 contact persons were enrolled and sputum samples were collected from 988/1,882 (52.5%). In the incentive-based arm, 780/1,295 (60.2%) index participants consented, 1,940 contact persons were enrolled and sputum was collected from 1,431/1,940 (73.8%). Index participants living with HIV (adjusted odds ratio, aOR=0.56, 95% CI 0.38-0.83) or unknown HIV status (aOR=0.12, 95% CI 0.07-0.20) were less likely to participate in the study. Contact persons in the incentive-based arm were more likely to provide a sputum sample compared to those in the household-based arm (aOR=2.12, 95% CI 1.80-2.50). Regardless of the study arm, cough (aOR=2.27, 95% CI 1.87-2.77), current smoking (aOR=2.22, 95% CI 1.63-3.02), and living with HIV (aOR=1.89, 95% CI 1.36-3.62) were associated with higher likelihood of sputum collection. There were gaps in implementation reach at the stages of contacting and enrolling index participants, enrolling individuals with HIV, and obtaining sputum, especially among those under 18 years and household contact persons

Directly Observed Therapy · Index case · Logistic regression · Odds · Odds ratio · Referral · Sputum · Tuberculosis · HIV/AIDS Research and Interventions · Pneumocystis jirovecii pneumonia detection and treatment · Tuberculosis Research and Epidemiology