A new health care index predicts short term mortality for TB and HIV co-infected people

BACKGROUND: Using 2004–2007 TB:HIV Study data from Europe and Latin America, we previously generated a health care index (HCI) for TB and HIV co-infected people. With improvements in diagnostic and management practices, we have now updated the HCI with new data. METHODS: We evaluated nine aspects of health care in Cox proportional hazards models on time from TB diagnosis to death. Kaplan-Meier methods were used to estimate the probability of death by HCI quartile. RESULTS: Of 1396 eligible individuals (72% male, 59% from Eastern Europe), 269 died within 12 months. Use of rifamycin/isoniazid/pyrazinamide-based treatment (HR 0.67, 95% CI 0.50–0.89), TB drug susceptibility testing (DST) and number of active TB drugs (DST þ,3 drugs (HR 1.09, 95% CI 0.80–1.48), DST þ ≥3 drugs (HR 0.49, 95% CI 0.35–0.70) vs. no DST), recent HIV-RNA measurement (HR 0.64, 95% CI 0.50–0.82) and combination antiretroviral therapy use (HR 0.72, 95% CI 0.53–0.97) were associated with mortality. These factors contributed respectively 5, –1, 8, 5 and 4 to the HCI. Lower HCI was associated with an increased probability of death; 30% (95% CI 26–35) vs. 9% (95% CI 6–13) in the lowest vs. the highest quartile. CONCLUSION: We found five potentially modifiable health care components that were associated with mortality among TB-HIV positive individuals. Validation of our HCI in other TB cohorts could enhance our findings.