Evaluating Accuracy, Impact, and Operational Challenges of GeneXpert Use for TB Case Finding Among HIV-infected Persons

Background: In Botswana, as in the rest of sub-Saharan Africa, undiagnosed tuberculosis (TB) or TB diagnosed late in the course of disease is thought to be the most common cause of death among HIV-infected persons receiving antiretroviral therapy (ART) as well as those not yet on ART. Common reasons for failure to diagnose TB early, include: (1) failure to appropriately screen HIV-infected patients for TB; and (2) difficulty in diagnosing TB with traditional diagnostic methods.

Public Health Intervention: The recent development of the Xpert MTB/RIF assay for the GeneXpert platform (Xpert) has revolutionized TB diagnostic capability for clinicians managing HIV-infected patients. Among HIV-infected adults, TB diagnostic sensitivity of Xpert (82.4%) has been proven superior to that of smear microscopy (44.6%). In line with World Health organization (WHO) guidelines, the Botswana Ministry of Health (MOH) and the United States Centers for Disease Control and Prevention (CDC), initiated a project in July 2012 to rapidly rollout 13 Xpert devices in service of 22 HIV care and treatment clinics while simultaneously initiating an evaluation to answer important operational research questions that can inform future national scale-up. To maximize the impact of the Xpert device in improving detection of active TB, the Xpert rollout was preceded by strengthening of TB screening procedures by: (1) ensuring the 22 HIV care and treatment clinics adopted the WHO-recommended four-symptom TB screen for adults; (2) situating trained TB case-finding nurses in all 22 facilities; and (3) training health facility personnel in both smear-microscopy-based and Xpert-based TB diagnostic algorithms. The combination of these strengthened TB screening procedures and rollout of the Xpert device is referred to as the "Xpert package" in this proposal.

Key Evaluation Objectives: The evaluation component of this project has two key objectives: (1) to evaluate whether the new MOH-recommended Xpert-based pulmonary TB diagnostic algorithm (including Xpert testing of sputum samples for all TB suspects and chest x-ray for Xpert-negative TB suspects) is more sensitive than the pre-Xpert smear-microscopy-based algorithm (smear microscopy and chest x-ray for smear-negative TB suspects) in diagnosing culture-positive TB disease; and (2) to evaluate the impact of the whole "Xpert package" on all-cause mortality during the first six months of ART, among adult patients.

Rationale for the First Evaluation Objective: Although it is expected that the Xpert-based TB diagnostic algorithm will be both more sensitive and more specific than the pre-Xpert algorithm, superiority of the Xpert- over the microscopy-based algorithm in Botswana has not yet been evaluated.

Rationale for the Second Evaluation Objective: In Botswana, undiagnosed TB or TB diagnosed late in the course of disease is thought to be the most common cause of death among HIV-infected persons receiving antiretroviral therapy (ART). The Xpert package could conceivably reduce TB-related mortality by ensuring all patients are screened for TB and ensuring TB suspects have access to a sensitive TB test (Xpert). Since the start of our Botswana stepped wedge trial in 2012, two separate but related randomized trials in Africa, the TB-NEAT study and the XTEND trial, have fallen short of demonstrating an association between use of Xpert for patients with symptoms of TB and reduced TB mortality. However, limitations in both these trials mean that the question of whether Xpert can reduce early ART mortality has not been definitively answered.

Design: Because the Xpert package is expected to have a beneficial effect for patients enrolling in the 22 study clinics, a phased rollout of the Xpert device using a stepped wedge design was chosen. Prior to Xpert device rollout, all 22 facilities initiated the intensified TB screening procedures simultaneously. Subsequently, 13 Xpert devices were activated in a step-wise manner over nine months with either one or two Xpert devices activated per month and the order of Xpert device rollout randomly assigned.

This step-wedge design involved enrollment of three cohorts:

1. Prospective cohort A (enrolled pre-Xpert device rollout): All patients who attended one of the 22 HIV clinics for the first time after study start (July 31, 2012), but before Xpert device rollout, were eligible for this cohort.
2. Prospective cohort B (enrolled post-Xpert device rollout): All patients who attend one of the 22 HIV clinics for the first time after Xpert device rollout were eligible for this cohort.
3. Retrospective cohort: All patients who initiated ART at one of the 22 HIV clinics for the first time in the 24 months before study start.

To answer the first primary study question, sensitivity of the pre- Xpert TB diagnostic algorithm in prospective cohort A will be compared with the post- Xpert algorithm sensitivity in prospective cohort B.

To answer the second primary study question, the investigators will compare 6-month ART mortality rates of ART enrollees in the retrospective cohort with 6-month ART mortality rates in prospective cohort B.

Power to Answer First Primary Study Question: Fisher's Exact Test sample size formula for comparing two proportions and Proc Power features in SAS version 9.2. software, were used to estimate study power. Power estimates were adjusted for the expected design effect to account for intra-cluster correlation. For design effect calculations, an intra-class correlation of 0.05 was assumed. With an anticipated sample size of 6,136 prospective enrollees (1,878 pre-Xpert and 4,258 post-Xpert), the study is estimated to have >80% power to detect pre- versus post-Xpert TB diagnostic sensitivities, assuming a culture-positive TB prevalence of 4% among prospective cohort enrollees, if pre-Xpert sensitivity is ≤52.5% and post-Xpert sensitivity is ≥82.5%.

Power to Answer Second Primary Study Question: As of August 2015, 10,131 ART patient medical records had been abstracted for the retrospective cohort. Published literature suggests pre-Xpert 6-month ART mortality rates should be about 15 per 100 person years (PY). Using published cluster randomized trial (CRT) power formulae adapted for stepped wedge trials and anticipated study sample sizes, the study has >80% power to detect a difference in 6-month all-cause ART mortality between cohorts R and B of 50% if cohort R mortality is ≥6/100 PY and of 40% if cohort R mortality is ≥13/100 PY.

Procedures: All 22 health facilities were assessed prior to training to troubleshoot potential logistical barriers to successful study implementation. All personnel involved in study procedures were trained before study start and closely supervised during implementation. Patients enrolled in the prospective cohorts are being followed for six months after HIV clinic enrollment, or through the end of TB treatment, whichever is later. Tracing of ART patients in retrospective and prospective cohorts, who are documented as late for appointments or lost to follow-up (missed last scheduled appointment by >60 days) has been performed according to national guidelines to facilitate estimation of true 6-month ART mortality.

Time line: Prospective cohort enrollment started in July 2012 and was complete by March 2014. Retrospective cohort enrollment was complete by March 2015. Patient follow-up and data entry will be complete in March 2016 at which time analysis to answer the first two primary study questions will be possible.