Study population and sampling.

Participants were from VL-endemic countries in East Africa and South-East Asia, and included national neglected tropical disease officers, national ministry of health functionaries, subnational programme managers, district and referral hospital physicians and nurses, nongovernmental organizations, academic researchers, respective regional and country office WHO personnel, and members of WHO’s South-East Asia Regional Technical Advisory Group for VL. We employed a convenience sampling methodology, by sending invitation emails to relevant listservs (computerized list of names and e-mail addresses). We obtained email addresses of stakeholders from databases available within WHO headquarter office, and contacted WHO regional offices for the African Region, Eastern Mediterranean Region and the Region for South-East Asia. The Ministry of Health of India, and WHO partners like Médecins Sans Frontières (MSF), Drugs for Neglected Diseases Initiative (DNDi) and Leishmaniasis East Africa Platform (LEAP) were also contacted. Based on obtained email addresses (n = 260), the majority of invited participants were from South-East Asia (n = 197, 76%). Those who accepted the invitation received the consent form and the online survey.

Data collection.

We developed the survey tool based on questionnaires previously developed for other WHO guidelines [13–16], and pre-tested it before sending to potential participants. The questionnaire covered participant demographics, valuation of the outcomes of interest, and judgements related to impact on equity, acceptability, and feasibility of the two interventions being compared (S1 Appendix). It also included open-ended questions on implementation considerations, monitoring and evaluation, and research priorities. The latter were similar to those asked during the semi-structured interviews (see below). We asked participants to rate the importance of the outcomes of interest on a 9-point Likert scale. For the remaining measures, we used the response options provided in the EtD. We sent two reminder emails after 7 and 13 days respectively. We used DataForm online tool from LimeSurvey. The landing page started with a consent form; participation was voluntary and participants could decline or exit the survey at any stage. Data were collected anonymously and independently by a data manager.

Data analysis.

We categorized the responses to the valuation of outcomes into critical (values of 7, 8 and 9), important (values of 4, 5 and 6), and of limited importance (1, 2 and 3). We used Stata (version 13) to conduct descriptive analyses. We analysed responses to open-ended questions along with the data from the semi-structured interviews.

Study population and sampling.

We included the same stakeholder groups as in the survey study (referred to as “non-patient stakeholders”), as well as VL patients with HIV-coinfection. For the non-patient stakeholders, we invited those who had responded positively to a question about their willingness to participate in a semi-structured interview. The lead investigator (JK) conducted the interviews with non-patient stakeholders after obtaining informed oral consent. At the treatment centre in Ethiopia where the clinical trial was conducted [7], physicians attempted to contact all the patients who had participated in the trial, while ensuring voluntary participation, confidentiality, and anonymity. A similar process was followed in India [8], except that a random sample of patients was contacted. Wherever possible, attempts were made to contact all VL patients under trial by the treating physicians. Patients were given information about this study and were invited for the interview. Local principal investigators (DP and NM) then contacted patients who volunteered and expressed their willingness, obtained their informed oral consent and conducted the interviews.

Data collection.

We developed an interview guide and discussed it with interviewers to maximize consistency. It covered the same constructs as the questionnaire but sought deeper views. Interviews with non-patient stakeholders were conducted in English, while interviews with patients were conducted in local language, using simplified language. We interviewed patients over the telephone, and non-patient stakeholders using an online meeting platform, and audio recorded all interviews. We stopped recruiting patients when we reached data saturation. We anonymized all transcripts using a unique identifier per interviewee.

Data analysis.

We used the framework thematic analytical approach guided by the EtD framework, which consisted of the following seven stages [17]: (1) transcription of the recordings verbatim (AEH and JK). Since interviews with patients were conducted in local language, the local investigators (DP and NM) translated interviews to English and transcribed them at the same time; (2) familiarization with the transcripts (AEH, CR and JK); (3) data indexing, which consisted of labeling each meaningful datum, initially applied to a few transcripts using the predetermined constructs of EtD framework (AEH, CR and JK). Texts that did not fit the predetermined constructs were also labeled; (4) discussion of labels leading to the creation of categories, forming the basis for the analytical framework (AEH, CR, JK and GHA); (5) application of the analytical framework to all transcripts, adding new labels that did not fit into the existing analytical framework when encountered (AEH, CR and JK); (6) charting and examining allocation of subcategories for each category, and reexamination of the labels that did not fit the existing analytical framework (AEH, CR and JK); (7) comparison and contrasting of categories and subcategories to finalize the results (AEH, CR, JK and GHA). All coders (AEH, CR and JK) first coded the same interview as part of a calibration exercise. The rest of the interviews were coded by one person. Coders met to determine the analytical framework, and compare and contrast findings. At the end of the process, we generated a complete narrative of the findings, supplemented with quotes from individual interviews.

Impact on equity, acceptability, and feasibility of treatment alternatives.

When comparing combination therapy with monotherapy, the majority of participants (n = 105, 60%) responded that providing combination therapy would result in increased health equity (n = 70, 40% “increased”, n = 35, 20% “probably increased”). Most participants (n = 139, 79%) responded that combination therapy is more acceptable (n = 97, 55% “more acceptable”, n = 42, 24% “probably more acceptable”), and that combination therapy is more feasible than monotherapy (n = 100, 57%; n = 58, 33% “more feasible”, n = 42, 24% “probably more feasible”) (Fig 3).

Download:

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

Fig 3. Impact on equity, acceptability, and feasibility of combination therapy as compared with monotherapy.

https://doi.org/10.1371/journal.pntd.0010624.g003

Valuation of the outcomes of interest.

Table A in S4 Appendix presents findings about the valuation for the following outcomes: mortality, disease-related complications, clinical cure, relapse, non-serious side effects, serious side effects, and patient satisfaction. The valuation of an outcome was driven by both the experience of the outcome itself (e.g., burden of relapse on the patient) and its consequences (e.g., relapse leading to mortality). For certain outcomes, it was also driven by the coinfection status of the patients (i.e., outcomes being more important in this specific population than in mono-infected or any patients).

Both patient and non-patient stakeholders considered reducing mortality as an important outcome. While non-patient stakeholders highlighted death as a critical outcome for any disease, they also noted the importance of mortality as an outcome in this specific population, since coinfection and immunosuppression place them at increased vulnerability. Patients considered survival as a “miracle” (India, P7), and recounted experiences of death among close relationships. Non-patient stakeholders were highly concerned about disease-related complications, for the possibility of being irreversible, being life-threatening, and having financial impacts. All participants perceived clinical cure as an opportunity for patients to resume their daily activities and be socially and economically productive. Non-patient stakeholders additionally placed a higher value on clinical cure at 6 months compared with clinical cure at the time of treatment completion, due to the increased risk of relapse among these patients.

Non-patient stakeholders described relapse as “the main feature of HIV-VL” (India, NP8) and “the biggest challenge” (India, NP5). This was due to how commonly relapse occurs, the risk of recurrent relapses, the resultant patient suffering and patients’ doubts about treatment effectiveness, the impact on patients’ mental health, and the ensuing risk of mortality. Relapse was also perceived to have negative effects from the health provider and community perspectives. From the community perspective, relapse could contribute to spreading the infection to other individuals, as patients are an “important reservoir for the infection” (India, NP2).

Non-patient stakeholders valued non-serious side effects less than the other outcomes, especially as compared with serious side effects. They considered non-serious side effects to be expected, well tolerated and easily treated, whereas serious side effects were viewed to result in additional burden on patients, and increased risk for permanent damage and mortality. Also, non-patient stakeholders were concerned about the risk of additive, more serious side effects, since this patient population is often receiving treatment for other diseases.

Participants rated patient satisfaction as an important outcome, for any patient and any disease. However, one non-patient stakeholder noted that satisfaction is usually overlooked: “we do not look at it normally” (India, NP7).

Impact on equity.

Most participants perceived no differential effect on equity between treatment alternatives as both are currently available free of charge and require hospitalization. They also noted that combination therapy is equally accessible for different geographical, gender and age groups. Nevertheless, few participants were concerned that pregnant women and women of childbearing age who may not prevent pregnancy were ineligible to receive combination therapy (particularly administration of miltefosine in this category of patients), and that it may negatively impact equity.

In addition, a number of views on general equity issues relevant to patients with HIV–VL coinfection emerged (see Table B in S4 Appendix for additional details). To note that half of the patient participants from India reported initially receiving care at a private health facility for substantial amounts of money (range: US$ 133–2600) before receiving care at a government health facility.

Feasibility.

Longer duration of hospitalization was generally perceived to be less feasible due to the economic burden on the health system, the patients and their caregivers. Overall, there were no concerns for adherence issues with combination therapy, due to the treatment being completed at the hospital. One feasibility issue with combination therapy was the added difficulty of counselling-related contraindication in pregnant women (Table C in S4 Appendix,).

Acceptability.

Participants considered combination therapy to be more acceptable than monotherapy, when associated with a shorter duration of hospitalization. Both patient and non-patient stakeholders highlighted the burden of a longer duration of hospitalization on patients and their families, in terms of lost wages and need for accompaniment. In addition, some patients mentioned that they were less comfortable at the hospital than at home, that longer duration of hospitalization resulted in separation from their families and in inability to undertake daily tasks. Nonetheless, few participants preferred longer duration of hospitalization due to medical supervision available at the hospital and poor living conditions at home. The only two female participants noted that use of contraception was not an issue (see Table D in S4 Appendix for additional details).

Implementation considerations, monitoring and evaluation, and research priorities.

Participants reported on a number of implementation considerations, including but not limited to access to care, discrimination and stigma, patient involvement, compassion by health providers, counselling, and holistic care. Monitoring and evaluation suggestions included the parameters (e.g., adverse events, drug interactions) and logistics (e.g., tools, checklists) of monitoring. Research priorities included disease epidemiology, diagnostic aspects, prevention, disease progression and prognosis, need for additional data, treatment-related issues, immunology, relationship with HIV, and health systems issues. More detailed findings are provided in S5 Appendix for the survey study and Table E in S4 Appendix for the semi-structured interviews.

Integrated findings.

For the valuation of outcomes of interest, the qualitative and quantitative findings were generally concordant, in terms of all treatment outcomes being important to a certain extent. However, the qualitative findings had higher discriminatory ability than the quantitative findings. Indeed, while there were very small differences in the survey ratings of the different outcomes, the qualitative findings showed that some outcomes (such as mortality and relapse) were more valued that the rest of the outcomes (such as non-serious adverse events). One area of expansion was through qualitative findings providing justifications for valuation of outcomes.

For the contextual factors, quantitative and qualitative findings were also concordant, with combination therapy being more acceptable and feasible. Qualitative findings expanded on the reasons behind the ratings resulting from the quantitative study. Of note, 28% of participants indicated that combination therapy was “probably less feasible” or “less feasible”. The qualitative study did not provide further insight into these findings.