The cascade of care in diagnosis and treatment of latent tuberculosis infection: a systematic review and meta-analysis

Background WHO estimates that a third of the world's population has latent tuberculosis infection and that less than 5% of those infected are diagnosed and treated to prevent tuberculosis. We aimed to systematically review studies that report the steps from initial tuberculosis screening through to treatment for latent tuberculosis infection, which we call the latent tuberculosis cascade of care. We specifically aimed to assess the number of people lost at each stage of the cascade. Methods We did a systematic review and meta-analysis of study-level observational data. We searched MEDLINE (via OVID), Embase, and Health Star for observational studies, published between 1946 and April 12, 2015, that reported primary data for diagnosis and treatment of latent tuberculosis infection. We did meta-analyses using random and fixed effects analyses to identify percentages of patients with latent tuberculosis infection completing each step in the cascade. We also estimated pooled proportions in subgroups stratified by different characteristics of interest to assess risk factors for losses. Results We identified 58 studies, describing 70 distinct cohorts and 748 572 people. Steps in the cascade associated with greater losses included completion of testing (71·9% [95% CI 71·8–72·0] of people intended for screening), completion of medical evaluation (43·7% [42·5–44·9]), recommendation for treatment (35·0% [33·8–36·4]), and completion of treatment if started (18·8% [16·3–19·7]). Steps with fewer losses included receiving test results, referral for evaluation if test positive, and accepting to start therapy if recommended. Factors associated with fewer losses were immune-compromising medical indications, being part of contact investigations, and use of rifamycin-based regimens. Interpretation We identify major losses at several steps in the cascade of care for latent tuberculosis infection. Improvements in management of latent tuberculosis will need programmatic approaches to address the losses at each step in the cascade. Funders Canadian Institutes of Health Research.