‘They are inconveniencing us’ - exploring how gaps in patient education and patient centred approaches interfere with TB treatment adherence: perspectives from patients and clinicians in the Free State Province, South Africa

This study was conducted at three public health facilities in the Mangaung District (Free State Province, South Africa) that served high density, marginalised communities with rife unemployment and poverty [29]. HIV prevalence and TB mortality were estimated at 11 and 9% respectively in the in the district [29]. The incidence of drug sensitive TB among those commenced on TB treatment in the Free State Province (2017) was 429.8 per 100,000, compared with the national average of 378.6 per 100,000 [30]. Tuberculosis was the leading cause of death, irrespective of age or sex, in the Free State Province [31] and HIV-positive TB mortality accounted for 73% of deaths among patients with TB in 2018 [1]. Further, 94% of patients newly diagnosed with TB were initiated on TB treatment and 90% of those patients who were TB/HIV co-infected were able to access anti-retroviral therapy [29]. Patients documented as LTFU regarding their TB treatment and clinicians that managed patients with TB at these facilities were approached to participate in the study.

This exploratory qualitative study was conducted between January and May 2018 with purposive sampling used to select patients. Three research assistants (RAs) that conducted the in-depth interviews (IDIs) were Good Clinical Practice (GCP) accredited, trained in qualitative data collection, had previous interviewing experience and fluent in the local languages (English, Sesotho, and Setswana). During training, interviewers were assessed and given feedback following direct observations on their ability to collect data using the interview guides. The investigators regularly reviewed the interviews and provided frequent feedback to the research assistants. Debriefing sessions were arranged to discuss experiences of conducting interviews, interview content, how to improve probing skills and address challenges.

The interviewers prior contact with the patients was limited to arranging interviews. The interviewers had engaged with the clinicians previously during quantitative TB work conducted at the facilities. Participants were provided with information on the qualitative component of the study and the interviewer’s role during the interview. The semi-structured interview questions allowed for the successive coding process required for deductive/inductive thematic analysis approach.

Patients and clinicians that were willing to provide written voluntary informed consent and agreed for the session to be digitally recorded were interviewed. Separate interview guides were developed for the patients (Annexure 1) and clinicians (Annexure 2) that explored access to health care, perceived quality of care, TB knowledge and how treatment support impacted TB treatment completion. The interviews were audio-recorded and were between 30 to 90 min. Basic demographic information (e.g. age, level of education, distance to clinic, and source of income, history of co-morbidities) was collected from patients (Annexure 3: Table 1). Information on position of employment and years of experience was collected from clinicians (Annexure 3: Table 2). The interviewers made field notes after the interviews. It was the author’s assessment that saturation was reached by revising the probes during data collection.

Patients listed on electronic TB records from study facilities; aged ≥18 years and documented as LTFU (as per the WHO treatment outcome definition [32]), were invited to participate in the study. The sample varied by age and gender. Patient contact information was retrieved from the clinical records. Initial attempts were made to contact patients telephonically. Failing which, liaison with the local tracing team was undertaken to trace patients using listed addresses. Adherence to standard practise for LTFU tracing was ensured with only the clinical staff of the facility, the contact tracing team and members of the research team who were tasked with locating LTFU patients having access to the patient identity and contact information. Reports submitted contained anonymised data. From a total of 268 patients identified with the TB outcome of LTFU on the ETR.net and EDR.web (electronic TB registers), we were only able to retrieve 71 case files. Sixty-one (61) of these case files contained invalid contact details and we were thus only able to contact 10 patients for interviews. One patient was unable to be interviewed as he had relocated to another province and two had died. The remaining seven patients identified for interviews were approached either face-to-face or telephonically. Five patients chose to be interviewed at their homes and two chose the facility. (Annexure 4). Patients who commuted to the facility for the interview were reimbursed R150 ($10 US) in lieu of travelling expenses. Inadequate collection of patient contact information and treatment records impacts on tracing and management of patients. Sinai et al. demonstrated that it is commonplace for patients to move between facilities once they have commenced their TB treatment [33]. Often these movements occur without prior notice to the provider and rendering patient tracking difficult to trace [3, 34, 35]. Sinai et al. additionally reflected on the data quality issues within the patients’ records, TB screening registers and case registers and with routine laboratory data including incompleteness, errors and discrepancies [33].

Clinical managers at each facility in collaboration with the research team identified clinicians that could be interviewed. Clinicians were currently employed at the selected health facilities; and were part of the nursing staff or doctors that were directly involved in TB care and management. Clinicians were approached face-to-face for participation and interviews were arranged at their convenience. Interviews were conducted in private areas within the facility.

Recorded interviews were transcribed verbatim and translated to English where applicable by research assistants. Transcripts were checked for accuracy by the research assistants. The transcripts were further checked, and identifiers removed by two investigators with medical and doctorate level qualifications. Manual coding was used for data analysis. To validate the findings, independent coding and theme identification was conducted by the same investigators. A codebook was developed that comprised a priori codes derived from the original question guide. Emerging thematic codes were developed during the thematic analysis and added to the codebook as required. Concepts from different interviews were pooled together and integrated into common themes. For reliability, four additional investigators reviewed the codebook before themes were finalised using the deductive/inductive thematic approach. The themes are represented as direct quotes. The themes were summarised by the authors as they were developed during the analysis and the direct quotes by the participants were used to support the summary.

Ethics approval for this study was obtained through the Human Research Ethics Committee (University of the Witwatersrand) [Ethics approval no: 170606] and the Free State Department of Health [Ethics approval no. FS_2017RP10_005]. Written consent was obtained from all participants. The interview scripts were coded, and personal identifying details were not collected.

Most patients reported that they were given TB information by clinicians during their clinical visits. However, it appears that the information received by patients did not translate into improved TB clinical outcomes as evidenced by the inability of this cohort of patients to complete their TB treatment regimen. Often patients demonstrated an understanding of the disease but the social determinants of health at play within the community, such as diminished educational and employment prospects, had far reaching consequences that included TB treatment LTFU, fear of disclosure and stigmatization. Patients felt that the TB messaging received was inadequate. The sub-themes derived from the interviews drew attention to the impact of health literacy on TB treatment outcomes and suggested that the rigid implementation of the current TB treatment guidelines stifled active patient involvement in their ongoing treatment. Commonly reported reasons for LTFU in the literature (including long queues, financial difficulty attending clinic and unreliable transport) were inconsistently reported in this study and were not considered for the analysis.

Many clinicians felt that improving the patient’s TB knowledge would reinforce adherence to treatment and thus focused on sharing information on treatment completion, side effects and infection control. However, the inability of clinicians to establish rapport with patients or to identify social support challenged TB treatment adherence by patients. Clinicians perceived this as patients not following their instructions despite what they considered lengthy TB education. Having said this, the clinicians concurred that their medical management of TB lacked the socio-economic dimension to treat a social disease of this magnitude.

“Ehm from the training side with clinicians, ehm mainly its ehm, they emphasise on … as we also know as clinicians that emphasis is mainly on them, informing patients about taking their treatment on a daily basis and the importance and when to take treatment mainly. But training, its, it doesn’t, you know the training that I have attended doesn’t really emphasise on the socio-economic factors of our patient,” (#8).

The subthemes below highlight how the clinician’s individual interpretation of the national TB guidelines resulted in significantly different key messages being communicated to patients, how the fragmentation of the health system leads to ongoing patient and clinician frustration; and begins to outline the importance of community support in addressing TB.

An important strength of this study is the study being based in a high-density population, with a significant TB disease burden. Further, the opinions of both the patients and clinicians were sought allowing for the exploration of TB treatment adherence from differing perspectives. This study explored the reasons for discontinuing TB treatment by approaching those who had stopped TB treatment. However, the absence of the views of those who had successfully completed treatment, for the purposes of meaningful comparison, may be considered a limitation of this study. We were unable to access all the patients that were LTFU as they had moved away, provided incorrect contact details or the data was missing. Their inclusion may have yielded additional insights. Potentially, some of the responses obtained may have been limited given the sensitive nature of TB and its association with HIV, which also carries a significant burden of stigma and discrimination. The study was limited to a single district in the Free State Province, thus the findings cannot be extrapolated to other settings.