Incidence and predictors of lost to follow-up among drug-resistant tuberculosis patients at University of Gondar Comprehensive Specialized Hospital, Northwest Ethiopia: a retrospective follow-up study

BACKGROUND: The emergence of Drug-Resistance Tuberculosis (DR-TB) is an increasing global public health problem. Lost to Follow-up (LTFU) from DR-TB treatment remains a major barrier to tuberculosis epidemic control and better treatment outcome. In Ethiopia, evidences on the incidence and predictors of LTFU are scarce. Thus, this study aimed to determine the incidence and identify the predictors of LTFU among DR-TB patients. METHODS: A retrospective follow-up study was conducted among a total of 332 DR-TB patients at the University of Gondar comprehensive specialized hospital. Data were retrieved from patient records from September 2010 to December 2017 and entered in to Epi-data 4.2.0.0 and analysed using Stata14.1 software. The risk was estimated using the Nelson-Aalen cumulative hazard curve. A log-rank test was used for survival comparisons between categories of independent variables. The Gompertz regression model was fitted, and hazard ratio with a 95% confidence interval (CI) was used to measure the strength of associations. Variables with less than 0.05 p-values in the multivariable model were considered as significantly associated with LTFU. RESULTS: Among a total of 332 patient records reviewed, 206 (62.05%) were male. The median age was 30 years (Inter Quartile Range (IQR): 23–40). Forty-one (12.35%) of the participants had no history of TB treatment, while a quarter of were TB-HIV co-infected. Closely all (92.17%) of the patients had pulmonary tuberculosis. The median follow up time was 20.37 months (IQR: 11.02, 21.80). Thirty-six (10.84%) patients were lost from follow-up with an incidence rate of 6.47 (95% CI: 4.67, 8.97)/1000 Person Months (PM). Homelessness (Adjusted Hazard Ratio (AHR) =2.51, 95%CI: 1.15, 5.45) and treatment enrolment year from 2013 to 2014 (AHR = 3.25, 95% CI: 1.30, 8.13) were significant predictors of LTFU. CONCLUSION: This study indicated that LTFU among DR-TB registered patients was high in the first six months compared to subsequent months. Homelessness and year of treatment enrolment were independent predictors of LTFU, requiring more economic support to patients in order to ensure treatment completion. This result can be generalized to patients who are using DR-TB treatment in similar settings.