Screening and testing for tuberculosis among the HIV-infected: outcomes from a large HIV programme in western Kenya

People living with HIV (PLHIV) are at increased risk of tuberculosis (TB). TB is also the leading opportunistic infection contributing to about one-third of deaths in this population. The World Health Organization recommends regular screening for TB in PLHIV. Those identified to have any TB-related symptoms are investigated and treated if diagnosed with TB. We sought to evaluate outcomes of intensified case finding and factors associated with undesirable screening for TB in a large HIV programme in western Kenya.

We conducted a retrospective study using routine programme data from the AMPATH HIV electronic medical records database for PLHIV in care between 2015 and 2016. Screening for TB was assessed by the recorded presence of cough ≥2 weeks, fever, night sweats, unintentional weight loss, chest pain and/or breathlessness. Undesirable screening was defined as being screened in < 90% of patient clinical encounters. Data were analyzed by encounters and per-patient. Factors associated with undesirable screening were analyzed using log-binomial regression and presented as relative risks.

There were 90,454 PLHIV, 65% females, median age 40 years, median follow-up time of 1.5 years. Total encounters were 683,898, of which screening for TB was recorded in 87%. 1424 (1.6%) PLHIV were not screened at all during the study period. 44% (95% CI: 43.6-44.3) of PLHIV were screened in < 90% of their clinical encounters (undesirable screening). TB-related symptoms were reported in 0.7% of screened encounters, while in 96% of PLHIV, no symptoms were reported. Overall, in 8% of symptomatic encounters sputum microscopy and/or chest radiography results were recorded. 92% of PLHIV did not have TB-related laboratory results recorded for all their symptomatic encounters. Factors which increased the risks of undesirable screening included: attendance at paediatric clinics (aRR: 1.27, 95% CI: 1.20-1.34), being on antiretroviral therapy (aRR: 1.16, 95% CI: 1.13-1.18), having more clinical encounters (aRR: 1.04, 95% CI: 1.04-1.04), and higher patient volumes in a clinic.

There were missed opportunities for screening and testing for TB. Screening was reduced by being on ART, having increased patient-encounters, the clinic setup, and by high patient volumes. HIV programmes should focus on quality of TB care in HIV clinics.