This evaluation involved a retrospective review of TB and HIV clinical records for adult TB patients aged 15 years and above who enrolled in all TB Directly Observed Therapy (DOT) facilities in Federal Capital Territory (FCT) and Ogun State between October 1, 2012 and September 30, 2013. Health facility TB registers and additional supplemental TB data sources such as Presumptive TB registers, TB treatment cards, and Local Government Area TB registers were reviewed to determine the total number of TB patients seen in 2013 with an HIV status recorded as positive or unknown. These patients were traced to HIV facilities to ascertain their treatment status and timing of ART initiation relative to TB treatment, using HIV care cards and ART registers. In both states, all PEPFAR-supported ART treatment facilities that were providing ART services as of October 1, 2012 were included. In FCT there were 16 facilities and in Ogun State there were 28.

In-depth interviews were conducted with 333 key health care workers selected randomly across all the study facilities using a semi-structured questionnaire, in order to assess knowledge and practices related to TB/HIV management as well as provider perceptions about the barriers to provision and uptake of ART.

Data were collected using a specifically designed data abstraction form that captured demographics, TB and HIV treatment, timing of ART, and TB treatment outcome information, following a retrospective review of the existing TB and HIV clinical and program recording and reporting systems. Descriptive statistics, including frequencies and percentages were used to summarize the data, and 95% confidence limits for all statistics were computed, taking into account the design effect resulting from within-clinic correlation. Unadjusted and adjusted odds ratios (ORs and AORs) and 95% confidence intervals (CI) were calculated to describe associations between categorical variables and ART uptake. Multiple logistic regression was used to assess for predictors of timely ART uptake. We defined ‘timely ART’ per WHO guidelines as ART started within two weeks of TB treatment initiation for patients with a CD4 cell count <50cells/μL, and within eight weeks for all other cases. For the purposes of the multivariate analysis, we also considered ART initiation to have been timely for patients with a missing CD4 cell count only if it was started within two weeks, presuming that in the absence of a CD4 test the patient should be treated sooner rather than later. SAS 9.3^® (SAS Institute, Cary, NC) was used with procedures designed to account for clustered data, such as PROC SURVEYFREQ and PROC SURVEYLOGTSTTC. The in-depth interviews were transcribed, coded, and thematically analyzed.

The protocol and all supporting data collection tools were reviewed by the United States Centers for Disease Control and Prevention (CDC) Associate Director for Science and the Nigeria Research and Ethics Committee, and was determined as a non-research program evaluation as the study involved the use of routine programmatic data with minimal risk to subjects.

Data were abstracted from a total of 4,810 TB patients from 44 DOTS sites in FCT and 70 in Ogun State. These were made up of 1,249 TB patients with a positive or unknown HIV status (574 and 675, respectively). Of the HIV-positive TB patients, median age was 36 (interquartile range [IQR] 30–43), 308 (54.1%) were female, and median CD4 cell count was 178cells/μL (IQR: 80–298) (Table 1). Most of these patients were between ages 25–44 (69.5%), and seen within secondary health care facilities (55.7%). Three hundred and fifty-one (65.9%) accessed DOT at health facilities while 182 (34.2%) had TB treatment supporters for treatment at home. Two hundred and forty six (44.7%) had pulmonary smear-positive TB, 277 (50.4%) had pulmonary smear-negative TB, and 27 (4.9%) had extra-pulmonary TB. The majority of patients were new TB cases (92.9%) (Table 1).

Among 574 HIV-positive TB patients evaluated, 90 (15.7%, 95% CI: 7.1–24.2) had already started ART before TB treatment, while 256 (44.6%, 95% CI: 35.9–53.3) were not documented to have initiated on ART during the six-month course of TB treatment at all. CD4 cell count was missing for 199 patients (34.7%). Of the 273 patients with a known CD4 cell count greater than or equal to 50cells/μL who were not already on ART when they started TB treatment, 82 (30.0%, 95% CI: 22.4–37.7) started ART within the recommended eight weeks of TB treatment initiation, while 60 (22.0%, 95% CI: 12.4–31.5) started ART during TB treatment but only after eight weeks. Among the 54 patients with a CD4 cell count <50cells/μL who were not already on ART at the time of TB treatment initiation, only eight (14.8%, 95% CI: 6.6–23.0) commenced ART within the recommended two weeks of TB treatment initiation, and 21 others (38.9%) started ART, but after two weeks had passed. The remaining 25 (46.3%) were not documented to have started ART at all while on TB treatment (Table 2). Of these 25, 14 achieved TB cure or treatment completion, three defaulted, three died, four transferred out and one had no TB treatment outcome reported.

In our multivariate analysis, CD4 cell count was found to be the only factor associated with Timely ART provision (using the previously stated definition of “timely ART provision”), with AORs of 2.7 (95% CI: 1.4–5.3) and 2.7 (95% CI: 1.2–6.2) for CD4 cell count categories 50–199cells/μL and 200–499/μL, respectively, when compared with having a CD4 cell count less than 50cells/μL (Table 3).

TB treatment outcome was documented for 558 (97.4%) of patients with a positive HIV status (Table 1). Seventy eight percent of patients with a CD4 cell count ≥50cells/μL had documented cure or treatment completion (a positive outcome), compared with 63.2% of those with CD4 cell count <50cells/μL and 57.2% of those with a missing CD4 cell count. Removing the 41 patients who transferred out, the odds of a positive outcome for patients with a CD4 cell count ≥50cells/μL was 2.6 (95% CI: 1.6–4.3) times that for patients with a missing CD4 cell count, but not significantly higher than for patients with a CD4<50cells/μL. For patients with documented treatment failure or default, median time to this outcome was 67.0 days (IQR: 43.0–141.1) for the eight patients (11.8%) with CD4<50/μL, 93.0 days (IQR: 63.0–154.0) for the 30 patients (44.1%) with CD4 ≥50cells/μL, and 72.0 days (IQR: 27.5–119.0) for the 30 patients (44.1%) with a missing CD4 cell count. Median time to death of those who died during follow up was 41 days (IQR: 34.0–53.0) for the six patients (10.0%) with CD4<50cells/μL, 55.0 days (IQR: 41.0–99.0) for the 22 (36.7%) with CD4≥50cells/μL, and 30.0 days (IQR: 4.0–53.0) for the 32 (53.3%) with a missing CD4 cell count.

Semi-structured interviews were conducted with 333 health care workers to assess their perceptions and current practices of patient referral between TB and HIV programs and facilities (Table 4). Of 312 interviewees that responded to the question on which TB/HIV service delivery model was used in their facility, 125 (40.1%) worked in a facility with an integrated TB/HIV service delivery model. Of 240 respondents, 137 (57.1%) reported having CD4 services available. The preferred mode of referral of patients between HIV and TB services was by escort (40.1%) followed by use of referral slips (38.4%). Most of the staff interviewed (84.4%) reported having TB/HIV guidelines available, however only 59.6% of 285 respondents reported prior training on TB/HIV. Additionally, only 38.8% of 242 respondents reported adhering to recommendations for ART initiation in TB/HIV co-infected patients irrespective of CD4 count, and only 41.9% knew that patients with a CD4 less than 50cells/μL should start ART within two weeks of TB treatment initiation.

Key recommendations from healthcare worker respondents for improving TB/HIV services for patients ranged from capacity building for health care workers (12.4%), proper patient follow up (19.0%), to patient counseling and health education (38.5%).