ASSESSING TUBERCULOSIS DIAGNOSTIC DELAY IN INDIA USING DISTRIBUTIONAL AND TAIL-BASED APPROACHES: A CROSS-SECTIONAL STUDY
Journal of Public Health Research and Community Health Development · 2026
DOI: 10.20473/jphrecode.v10i1.85960 ↗Abstract
Background: The delay in the diagnosis of tuberculosis is a key challenge in managing the disease, particularly in highly burdened areas. The mean and median do not provide insight into the extent to which patients' waiting periods vary, nor into their length. Purpose: This study examined how diagnostic delay is distributed among TB patients and identified which factors drive the longest delays. Methods: We conducted a cross-sectional study using patient-level TB data. Diagnostic delay was defined as the gap between symptom onset and TB diagnosis. We fit the data to three distributions Gamma, Weibull, and Lognormal and flagged delays in the top 10% as “extreme.” Logistic regression was then used to examine how extreme delays related to demographics and healthcare access. Results: Delays ranged from 1 day to 359 days. The median was 32.5 days, while the mean delay was 49.2 days, indicating right-skewed data. More than 10% of patients took more than 100 days to be diagnosed. Of the three models, the one that fits best was the Gamma model. Those people living in rural areas and those seeking treatment from private institutions had high representations in the extremely delayed diagnosis group. Conclusion: Diagnostic delays for TB vary widely, they are skewed and inconsistent in such a way that a means cannot capture all the information. Examining the entire distribution, particularly its tail, provides insight into how severe the delays are and where they are most severe.
EUVIMED