Tuberculosis (TB) killed approximately 1.4 million people in 2015 and is the leading cause of mortality among people living with HIV (PLHIV), globally [1]. In 2015, an estimated 400,000 people died of HIV-associated TB, with the majority of these deaths occurring in sub-Saharan Africa [1]. South Africa has among the highest rates of HIV and TB incidence in the world [1]. The provision of isoniazid preventive therapy (IPT) can decrease the incidence of active TB in HIV-infected individuals by 32% [2]. World Health Organization (WHO) recommendations call for ruling out active TB disease in PLHIV prior to providing IPT [3]. Tuberculin skin test (TST) positivity is used to identify those HIV-infected individuals who would benefit from at least 36 months of IPT [3]. The South African Ministry of Health adopted the WHO guidelines in 2014 [4], though with certain caveats regarding duration based on TST results and anti-retroviral therapy (ART) status.

The use of TST, currently the most common screening test for latent TB, remains suboptimal as a positive result does not differentiate between active disease versus prior infection [5]. Moreover, in certain populations false negative results are common, particularly among PLHIV with CD4 counts < 200 cells/mm3 [6]. Several operational barriers have been identified regarding TST in highly burdened health systems and resource-constrained settings. These include: a lack of human resources to administer the test, staff training, and tuberculin shortages [7]. The TST also requires patients to make at least two visits to a health facility, increasing the possibility for losses to follow up.

Newer interferon-gamma release assays (IGRA) have the advantage of using specific Mycobacterium tuberculosis antigens, which provide higher specificity [8]. One such assay, the QuantiFERON TB Gold In-Tube (QGIT) test, has been approved and licensed for use in multiple settings, generating debate regarding the possibility of a more effective and widely used screening tool [9]. While IGRA tests are more expensive for a health system than TST in the short term, they may be cost effective and create cost savings in the longer term if they result in significantly better detection and treatment of latent TB with subsequent prevention of TB disease, as has been observed in some settings [10].

Several studies are underway assessing the use of QGIT in clinical care, including the TEKO trial, which is a cluster-randomized trial among 14 clinics in North West Province of South Africa and the parent study for the current qualitative sub-study and analysis. The TEKO (meaning “to test” in Sesotho) study is formally titled, Quantiferon Gold Test for Detecting Tuberculosis (TB) Infection in HIV/AIDS Patients in South Africa and is registered with ClinicalTrials.gov (NCT02119130). The larger TEKO trial will compare the cost-effectiveness of the QGIT collected at the same time as the blood drawn for routine CD4 count among HIV-infected adults, compared with TST only, on the proportion of PLHIV initiated on the appropriate IPT regimen per South African national guidelines [4]. There were 3398 eligible PLHIV enrolled in TEKO and followed for 2 years. Participants were newly HIV-infected, over 18 years of age, and eligible for TST/IPT. The trial started enrollment in 2015 and intervention implementation and evaluation of key outcomes is still ongoing. Follow-up data collection from the TEKO trial will be completed in 2018.

Prior to the TEKO trial, a formative study was conducted to explore attitudinal and operational factors surrounding the implementation of QGIT and TST among patients and providers. Here we present findings from this qualitative exploration, focused on the relevance and acceptability of using QGIT to address barriers to latent TB screening.

We conducted formative qualitative research prior to the initiation of the TEKO trial from April to June 2014. The study utilized a cross-sectional design and a total of 56 people participated in one individual, in-depth interview. The sample included 28 PLHIV receiving care in public clinics and 28 of their clinical care providers. As indicated by qualitative research norms, we sought between 20 and 30 participants per population group (e.g. patients and providers) [11, 12]. The 14 clinics included were all general primary health care clinics, which also offered HIV services and were selected to participate in the TEKO trial, which at that point had not yet begun.

Eligibility for the qualitative study included patients who were 18 years of age or more, living with HIV for at least 6 months, and able to converse in English. One female and one male patient participant were interviewed in each participating clinic. These patients were selected by providers based on their knowledge of patients who might be willing and able to narrate their experience with their clinical care. Two clinical care providers that provided TB care were interviewed in each participating clinic. We sought different types of providers involved in providing TB screening and care to patients (e.g. nurses, counselors, and community health workers).

A semi-structured guide was used to facilitate the interviews using open-ended questions Additional file 1. The interviewer freely utilized probes and follow-up questions to create additional depth. Interviews focused on participant’s views and experiences with TB including details of the screening, diagnosis, and treatment process from patient and provider perspectives. All participants were asked their thoughts about the use of the QGIT test to screen blood drawn at the same time as CD4 blood draws for latent TB. Interviews continued until saturation, or redundancy in findings per key themes, was reached, which was determined by readings of individual transcripts as they came in. These readings also helped ensure the quality of the data collection and transcription.

Clinical care providers were approached directly by study staff, while patients were recruited with the assistance of clinic staff. All interviews took place at the clinic where the participant worked or received care in a private room. Written informed consent was obtained from all participants. The interviews took place in English and lasted approximately one hour. Institutional Review Boards of the Johns Hopkins School of Medicine and the University of Witwatersrand in South Africa approved the study.

All interviews were audio recorded and transcribed. Thematic content analysis [13] was conducted coding and synthesizing provider and patient participants’ views and experiences on the key domains outlined above, as well as those that emerged through the course of the interview through the use of case summaries of each interview. Both the first and second author coded the data in parallel and worked together to generate the case summaries. Three salient themes were documented and are presented below on the clinic environment, TB management practices, and perspectives on the use of the QGIT.

This qualitative study provides important insights into the operational challenges associated with currently recommended TB screening approaches among PLHIV in resource-constrained settings, including use of the TST, as has been documented in prior work⁷. Given the significant burden currently placed on low- and middle-income country health systems, in terms of high patient caseloads, limited personnel, stocks and supplies, long wait times and distances to access care, identifying a more streamlined approach to systematically detect latent TB infection among PLHIV is urgently needed to prevent the leading cause of mortality in this population [14].

Findings from this formative study conducted prior to the start of the now ongoing TEKO trial suggest the likely relevance and acceptability of using the QGIT to screen CD4 blood draws for latent TB infection among PLHIV in rural South Africa.. The approach was viewed as a simple, but potentially powerful, strategy to overcome several logistical barriers to implementing screening for latent TB infection. Yet, both patient and clinical care providers were hesitant to suggest that such an approach should replace other methods, including the use of the TST.

Our qualitative results strongly suggest the need for further dialogue and training regarding the difference between latent and active TB infection and the implementation of the most recent national guidelines on TB among PLHIV in South Africa. Many questions regarding how to interpret and operationalize the guidelines were documented in this exploratory study. As the TEKO trial proceeds, qualitative work will continue to characterize the rollout of the national guidelines and how their adoption in and across clinics may influence the relative effectiveness of a QGIT plus TST versus TST only approach to screening for latent TB infection among PLHIV.

Differences in providers versus patients’ perspectives regarding the daily practices and implementation of current national TB screening approaches underscore the need to engage both groups in future qualitative work regarding their potentially unique experiences in, and perspectives on, the meaning and results of the TEKO trial.

Limitations of the current analysis include the cross-sectional nature of the in-depth interviews conducted, which included only English speakers in the baseline sample, potentially limiting the full range of perspectives specifically among patients. Additionally, patients in this qualitative sample were generally already on ART for some time, whereas TEKO trial participants are newly diagnosed PLHIV. Despite such limitations, our findings revealed strong consistency in terms of patient TB-related experiences in participating clinics and views on how to further strengthen TB management in PLHIV through the use of novel screening strategies such as QGIT.

While this qualitative work was not intended to generate generalizeable results, the common operational barriers to TB screening documented in the literature suggest that QGIT acceptability may be transferable to other low- and middle-income settings.

Findings indicate that using the QGIT to screen routine CD4 blood draws among PLHIV is an acceptable strategy to increase latent TB detection given the challenges associated with ensuring systematic TB screening in overburdened public clinics.