Recruitment

TB-survivors aged ≥18-years who had successfully completed treatment for PTB disease, but reported chronic or recurrent respiratory symptoms (cough, chest pain, breathlessness, sputum production), were recruited for IDIs. We recruited both individuals who were and were not engaged with health seeking, and recruitment was stratified by HIV status, which is recorded in the majority of TB patients in Malawi.

Health facility sampling was used to identify those who were actively engaged in health seeking. TB-Officers (TBOs)–a cadre of environmental health officers who deliver TB services in Malawi–from the five largest health facilities in Blantyre (Queen Elizabeth Central Hospital (QECH), and Ndirande, Limbe, Bangwe and Zingwangwa Health Centres) were asked to identify and refer sequential patients attending health centres for recurrent respiratory symptoms, whether or not the patient went on to re-initiate TB treatment. Community sampling was used to identify those who were not actively engaged in health seeking, using two approaches designed to capture TB survivors at a range of times from treatment completion. Firstly, a random sample of PTB-survivors who had completed treatment in Bangwe district within the past 2–5 years was taken from the district TB register. These individuals were contacted by phone and approached to participate if they had ever experienced respiratory symptoms after TB treatment completion. Secondly, PTB-survivors reporting residual respiratory symptoms at TB treatment completion were identified from a morbidity surveillance study which was being run concurrently in urban Blantyre, by our team. These individuals were sequentially contacted by phone 6–12 months after treatment completion and approached to participate.

A convenience sample of TBOs from 18 health centres across Blantyre was recruited for participation in two FGDs. Participants were identified at the monthly TB surveillance meetings held for TBOs in Blantyre, and FGDs were held alongside these meetings. Previous experience of qualitative work suggests that TBOs are comfortable sharing their experiences around patient care in group discussions [20].

Facility-based health care workers (HCWs) were purposively recruited for IDIs. The initial sample included TBOs who were engaged in the FGDs, and HCWs who were known by the study team to be involved in TB and respiratory care in Blantyre. A snowballing approach was then used to identify others with interest and experience in managing symptomatic TB-survivors, with purposive sampling to include providers with diverse levels of training and experience.

Data collection

Interview topic guides were developed using existing literature on TB and chronic respiratory disease, and were rooted in previous stakeholder engagement work with TB patient groups, health care providers and policy makers from East Africa [21]. Topics explored with TB-survivors included: experiences of receiving and completing TB treatment; perceptions of health and wellbeing after treatment; the meanings ascribed and actions resulting from residual or recurrent respiratory symptoms; drivers and experiences of health care seeking for these symptoms; perceptions of risk around recurrent TB; and experiences of re-treatment. Topics explored with HCWs included: current practice at TB treatment completion; beliefs about the meaning and drivers of residual or recurrent symptoms; approaches to the investigation of those returning to care, including the use of TB and respiratory diagnostics; approaches to decision making around care, including antibiotic use and TB retreatment. Both TB-survivors and HCWs were asked about perceived challenges in care for symptomatic TB-survivors, and for their suggestions of how to address these.

Written informed consent was obtained from all participants, with support from a guardian or impartial witness for those unable to read. FGDs and IDIs were conducted in Chichewa, the local language spoken in Blantyre, Malawi, by fluent research assistants (WMP and EM). Interviews and focus groups were held at sites where confidentiality could be maintained, and which were comfortable to or chosen by the participants. This included locations outside, in participant’s homes, or at local health facilities. Discussions were audio-recorded and transcribed and translated into English. The research team kept notes of interview findings and contexts, with memos used to support reflexivity about the perspectives and position of the research team. Debriefings between the research team members took place regularly to share insights and observations, discuss emerging themes and relationships between them, and iteratively review the topic guide. Final sample size was determined by saturation, with initial plans for IDIs with up to 20 HCWs and 40 TB-survivors.

Data coding and analysis

Data were imported into NVivo (QSR International) for coding. Broad codes were developed for HCW and TB-survivor data. This was done both deductively and inductively–initial codes were based on the research questions and interview topics, and were adapted over time in response to the data and as more refined sub-codes emerged. Coding was led by the research assistant who had led data collection (WMP) with parallel coding of early transcripts and ongoing review by JM and EM. Discussion with the broader research team was used to reconcile coding differences (JM, WMP, EM, NE). Thematic analysis was employed to identify patterns, themes, and key findings across the HCW and TB-survivor data together. This was an iterative process involving several members of the research team (JM, WMP, EM, NE), and required moving back and forth between the coded material, researcher observations and memos, and emerging analysis, and theoretical approaches to understand the data [22]. We borrowed from theories on the TB care cascade [23], and the invisible work of patients and health care workers in completing diagnostic journeys to support this analysis [24, 25]. We employed a technology-in-practice perspective which recognises that medical practice is a result of diverse elements including bodies, samples, equipment, the physical organisation of clinical spaces, and conversations and relationships between those receiving and providing care [26]. Data from HCWs and TB survivors were compared to support triangulation.

The characteristics and qualifications of the research team, and factors affecting research reliability, validity and transferability are described in the Supplementary materials (S1 and S2 Tables).

Ethical approvals

All participants provided written informed consent. Ethical approval was obtained from the Liverpool School of Tropical Medicine (21–069) and Malawi College of Medicine Research Ethics (P.08/21/3377) Committees.

Patient experience of residual or recurrent symptoms

TB-survivors reported negative experiences of residual or recurrent symptoms after TB treatment completion, dominated by themes of uncertainty, fear, and anxiety.

Participants with residual symptoms at treatment completion expressed significant uncertainty, asking why they did not feel better, and whether the TB medication had worked (Table 1, Quote 1). Some sought reassurance from TB-Officers that they had been cured (Table 1, Quote 2), with others pursuing repeat CXR imaging to understand their health and obtain ‘peace of mind’ (Table 1, Quote 3). The concept of post-TB lung disease–described as ‘sores’ within the lung–was frequently used by TB-Officers to explain residual symptoms to patients, but was framed as a temporary problem which would heal over time rather than a chronic condition. For those that failed to improve there was a strong mismatch between the reassurance received and their own experience of ongoing poor health, and this led to further uncertainty and loss of faith in the care received (Table 1, Quote 4). Participants expressed uncertainty about whether, when and how to seek care for symptoms, reporting a lack of guidance from TB-Officers at treatment completion. (Table 1, Quote 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Quotations.

https://doi.org/10.1371/journal.pgph.0003436.t001

Healthcare seeking for recurrent symptoms was largely within the formal health system, with limited reference to traditional/spiritual treatment (Table 1, Quote 6). This biomedical approach was rooted in the belief that TB is a condition best managed within the health system, and fear that symptoms represented recurrent TB disease (Table 1, Quote 7).

Symptomatic TB survivors expressed anxiety that they would be blamed for their symptoms if they returned to care. They worried that they would be accused of poor medication adherence during the initial TB treatment episode, participation in harmful behaviours such as alcohol use or smoking, or ongoing work in harmful environments which they had been advised to avoid (Table 1, Quote 8). These fears were rooted in the counselling received during the initial TB disease episode where these exposures were framed as the cause of TB disease. In addition to fear of being blamed by others, TB-survivors expressed self-blame, with several questioning how their own behaviours might have led to recurrent illness, or expressing concern about ongoing exposure to dusts and fumes in occupations which they were unable to leave for financial reasons (Table 1, Quote 9). These exposures were framed as risk factors for TB disease only, rather than for broader respiratory disease.

Symptomatic TB-survivors returning to care experienced multiple visits for clinical review or investigations, with a lack of clear diagnosis of the cause of their symptoms. This was particularly the case for those who were sputum Xpert MTB/Rif negative on investigation. Most of our participants had received several courses of antibiotics with little perceived benefit. For several of them a diagnosis of recurrent TB disease was met with relief, with TB retreatment felt to offer renewed hope of recovery (Table 1, Quote 10).

The post-TB care pathway

We identified four time points along the post-TB pathway, which were critical in shaping patient experiences (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The post TB care pathway, with four critical intervention points.

1. Counselling and approaches to discharge at the end of TB treatment; 2. Health seeking pathways for return to care; 3. Initial investigation and care when re-engaging with services; 4. Decision making around TB retreatment.

https://doi.org/10.1371/journal.pgph.0003436.g001

TB counselling and discharge from care

Our data suggest limited counselling about residual symptoms at treatment completion, and a lack of follow-up for TB-survivors with ongoing symptoms or concerns.

Counselling during TB treatment in Malawi is delivered by TB-Officers, who described limited training in counselling, and on-the job experience determining the messaging used (Table 1, Quote 11). The focus of counselling was largely around a set of behaviours and exposures (e.g. poor TB treatment adherence, occupational dusts and fumes, smoking, alcohol) believed to cause TB disease, and which patients were told to avoid in order to remain well. These factors were framed as being within the control of patients, with recurrent TB disease and symptoms occurring as a result of failing to follow advice (Table 1, Quote 12). Neither the risk of re-infection after treatment completion, nor relapse occurring due to inadequate treatment were mentioned by HCWs as a potential cause of disease recurrence. Whilst TB-Officers described clearly advising patients on when and how to seek care in the event of deterioration after TB treatment completion (Table 1, Quote 13), the majority of TB-survivors did not recall receiving this advice (Table 1, Quote 5).

TB-Officers indicated that they often knew which TB-survivors were at high risk of relapse due to poor adherence during treatment, or broader social determinants of health (e.g. malnutrition, poverty, alcohol excess). However, the TB-survivors interviewed had all been discharged at treatment completion without linkage to follow-up or ongoing care to address these challenges. HCWs confirmed that these patients were usually discharged, with a sense of surrender about what would follow (Table 1, Quote 14).

Health seeking pathways

We identified uncertainty amongst HCWs about health seeking pathways for symptomatic TB-survivors, with no agreed pathway for when and how they should return to care. The health seeking approaches described by TB-survivors were similarly heterogenous, often with initial use of over-the-counter medications in the community prior to health facility review. TB-survivors were more likely to return to TB services directly if they had established a good relationship with the TB-Officers during the initial treatment episode. However, many returned to general medical outpatient departments at health facilities, where both TB-survivors and HCWs described the limited amount of time available for assessment due to patient volume and staff shortages (Table 1, Quote 15).

TB-focused investigations and care

Initial investigations for those who did return to care were focused on excluding recurrent TB disease and drug resistance, with little attention to broader cardio-respiratory disease.

Most HCW described initial investigation with sputum Xpert MTB/Rif and CXR, with samples sent to the reference laboratory for TB culture. TB-Officers emphasized the challenge of differentiating between PTLD and recurrent TB disease on CXR (Table 1, Quote 16), but discussed CXR findings in relation to making a TB diagnosis only. There was no suggestion that CXR was used to identify other types of cardio-respiratory disease, or of linkage of those with likely PTLD to ongoing care. HCWs noted that respiratory and cardiac diagnostics (e.g. spirometry, echocardiography) are only available to clinicians in the tertiary referral hospital. TB-survivors were occasionally referred to other medical teams for investigation of non-TB causes of cough, but described anxiety due to conflicting diagnoses and lack of communication between providers (Table 1, Quote 17).

The use of antibiotics was widespread, with a range of drugs and durations used. In the absence of clear guidelines, several clinicians described using a trial of antibiotics as part of the TB diagnostic pathway, with clinical treatment response used to inform decision making around TB retreatment (Table 1, Quote 18). Others described using antibiotics to treat bacterial infection whilst awaiting results of TB investigations.

Varied decision making around TB retreatment

Most symptomatic TB-survivors were tested with sputum Xpert MTB/Rif, regardless of time from initial treatment completion. TB-Officers and clinicians trusted positive Xpert MTB/Rif results in order to initiate retreatment (Table 1, Quote 19). However decision making for those who were sputum Xpert MTB/Rif negative was perceived as more challenging, and the remit of clinicians and medical officers only.

Clinicians noted the difficulty of differentiating between recurrent TB disease and PTLD amongst those who are sputum Xpert MTB/Rif negative. Strategies used to manage this uncertainty included ad-hoc case discussions with colleagues, or referral to the tertiary centre for further investigation and senior clinical review, but these steps were variably used by individual clinicians (Table 1, Quote 20). Access to experienced colleagues was limited in smaller healthcare facilities (Table 1, Quote 21), and the timing of tertiary centre referral varied widely with several patients reporting to have been retreated several times before referral. There was a sense amongst senior clinicians in Blantyre that early referral for senior TB clinical review and broader investigation would improve clinical management (Table 1, Quote 22). However, HCWs noted the lack of feedback between health facilities and the tertiary centre, and concern that patients would be lost to follow up as key barriers to referral (Table 1, Quote 23). Although the potential harm of unnecessary TB retreatment was implicit in the approach described for decision making around retreatment, the risks and benefits considered were not explicitly described by HCWs, and we did not identify any specific decision-making tools in use.