Empirical evidence of delays in diagnosis and treatment of pulmonary tuberculosis: systematic review and meta-regression analysis

BACKGROUND: Delays in diagnosis and treatment of pulmonary tuberculosis are a major set-back to global tuberculosis control. There is currently no global evidence on the average delays thus, the most important contributor to total delay is unknown. We aimed to estimate average delay measures and to investigate sources for heterogeneity among studies assessing delay measures. METHODS: Systematic review of studies reporting mean (± standard deviation) or median (interquartile range, IQR) of patient, doctor, diagnostic, treatment, health system and/or total delays in journal articles indexed in PubMed. We pooled mean delays using random-effects inverse variance meta-analysis, investigated for variations in pooled estimates in subgroup analyses and explored for sources of heterogeneity using pre-specified explanatory variables. RESULTS: The systematic review included 198 studies (831,724 patients) from 78 countries. The median number of patients per study was 243 (IQR; 160–458) patients. Overall, the pooled mean total delay was 87.6 (95% CI: 81.4–93.9) days. The most important and largest contributor to total delay was patient delay with a pooled mean delay of 81 (95% CI: 70–92) days followed by doctor’s delay and treatment delay with pooled mean delays of 29.5 (95% CI: 25.9–33.0) and 7.9 (95% CI: 6.9–8.9) days respectively. There was considerable heterogeneity in all pooled analyses (I(2) > 95%). In the meta-regression models of mean delays, studies excluding extra-pulmonary tuberculosis patients reported increased mean doctor’s delay by 45 days on average, non-use of chest x-ray and conducting studies in high income countries decreased mean treatment delay by 20 and 22 days on average, respectively. CONCLUSION: Strategies to address patients’ delay could have important implications for the success of the global tuberculosis control programmes.