Menu
Menu
Menu
Menu
Interactions between patients and service providers frequently influence uptake of prevention of mother-to-child transmission (PMTCT) HIV services in sub-Saharan Africa, but this process has not been examined in depth. This study explores how patient-provider relations influence PMTCT service use in four government facilities in Kisesa, Tanzania. Qualitative data were collected in 2012 through participatory group activities with community members (3 male, 3 female groups), in-depth interviews with 21 women who delivered recently (16 HIV-positive), 9 health providers, and observations in antenatal clinics. Data were transcribed, translated into English and analysed with NVIVO9 using an adapted theoretical model of patient-centred care. Three themes emerged: decision-making processes, trust, and features of care. There were few examples of shared decision-making, with a power imbalance in favour of providers, although they offered substantial psycho-social support. Unclear communication by providers, and patients not asking questions, resulted in missed services. Omission of pre-HIV test counselling was often noted, influencing women’s ability to opt-out of HIV testing. Trust in providers was limited by confidentiality concerns, and some HIV-positive women were anxious about referrals to other facilities after establishing trust in their original provider. Good care was recounted by some women, but many (HIV-positive and negative) described disrespectful staff including discrimination of HIV-positive patients and scolding, particularly during delivery; exacerbated by lack of materials (gloves, sheets) and associated costs, which frustrated staff. Experienced or anticipated negative staff behaviour influenced adherence to subsequent PMTCT components. Findings revealed a pivotal role for patient-provider relations in PMTCT service use. Disrespectful treatment and lack of informed consent for HIV testing require urgent attention by PMTCT programme managers. Strategies should address staff behaviour, emphasizing ethical standards and communication, and empower patients to seek information about available services. Optimising provider-patient relations can improve uptake of maternal health services more broadly, and ART adherence. © 2014 Gourlay et al.
Kisesa is a relatively poor rural area in north-western Tanzania with a low GDP per capita of <500 US dollars (USD). Most residents earn a living from subsistence farming or small businesses selling local produce. There are four government health facilities in Kisesa: three rural dispensaries in remote villages and one health centre in the trading centre. The district and regional referral hospitals are ≥20 km away. PMTCT services have been operating (intermittently) in all four facilities since 2009, although comprehensive HIV services are only offered in the health centre (since 2008). Regular rounds of demographic and HIV serological surveillance have been conducted in Kisesa, among a population of approximately 30,000, since 1994. A variety of qualitative methods – participatory learning and action (PLA) group activities with mothers and fathers, in-depth interviews (IDI) with HIV-positive and HIV-negative women as well as health providers, and observations of procedures and patient-provider interactions in ANC and MCH clinics – were used in a broader study investigating barriers to PMTCT service use in Kisesa in 2012. PLAs extend more commonly used focus group discussion methods to include further participation from participants, such as role-playing or creating charts or maps, promoting a two-way exchange of knowledge and information between researchers and participants. PLA methods had been useful previously in this setting, to investigate barriers to accessing other HIV services, engaging community members and capturing prevalent community-wide beliefs which complement individual perspectives from IDIs [27]. PLA activities were also designed to generate a locally relevant vignette (short story) about a hypothetical pregnant woman diagnosed with HIV to aid discussion during the IDIs given the sensitive topic area, and limit reporting bias [28]. The study was designed to maximise learning from different methodologies and perspectives, including those of HIV-positive women, to explore direct experiences with PMTCT services, and HIV-negative women, for more general experiences of MCH services, and the wider community and health service providers. PLA activities were conducted with 3 male and 3 female groups from different residence areas (remote (RE), roadside (RD), trading centre (TC)). Participants were randomly selected from a sampling frame, constructed using demographic and HIV sero-surveillance datasets, of 3102 community members aged 15–60 who had ≥1 child, for experiences or views on MCH services. A few HIV-positive women were purposively included in each female PLA group using an approach based on the ‘seeded’ focus group [29]. Fieldworkers were unaware of the HIV status of any individuals. Overall, 30% of 105 individuals visited were not found, and 2% refused to participate. Each group included 8–12 participants. Demographics of PLA participants are summarised in table 2. Male participants were slightly older than female participants, with an average age of 42 years compared to 36 years respectively. IDIs were conducted with 21 women who had been pregnant or given birth since 2009. Ten HIV-positive women were recruited purposively by nurses from each clinic, while eleven women (5 HIV-negative, 6 HIV-positive) were recruited from the community via PLAs, ensuring a mix of women with and without experience of ANC or PMTCT services. Women were discreetly invited for interview by fieldworkers after the PLAs, using a coded list. Fieldworkers were unaware of participants' HIV status, unless participants voluntarily disclosed their status during interviews. Of 8 HIV-positive women invited, 6 accepted. All 5 HIV-negative women accepted. Recruitment continued until data saturation was reached, based on preliminary analyses. Characteristics of women participating in the IDIs are presented in table 3. Participants were aged between 20 and 47 years, with an average age of 34 years. Roughly half lived in more remote rural villages, the other half residing in the trading centre or roadside villages. Among the subset of women recruited from the community for whom data on education was available, most were educated to primary level (as were the majority of women in this community). Six health workers and three officials were interviewed; recruited purposively to include most of the providers directly involved in delivering PMTCT services in the study area, and to provide perspectives from different facility levels and cadres. Five MCH nurses, a doctor at the CTC, and coordinators at the district and referral hospital were selected and agreed to participate. The health workers interviewed were mostly female (5 out of 6), while two out of three health officials were male. Half of the health workers practised in smaller dispensaries and half worked in the health centre. PLAs were facilitated in Kiswahili by fieldworkers following a written protocol of activities, including group discussions, role-plays about a woman attempting to access PMTCT services (participants created their own characters and stories), brainstorming of barriers along the PMTCT cascade, and ranking of barriers. Trained interviewers conducted IDIs lasting 1–3 hours in Kiswahili. The semi-structured discussion guide included a vignette [28], personal experiences of recent pregnancies, and overall perceptions of services received. Interviewers probed for MCH services accessed, health worker conduct, privacy, counselling offered, and clarity of medical advice. Twelve of sixteen known HIV-positive interviewees voluntarily disclosed their status and discussed PMTCT services. A fieldworker interviewed health workers in Kiswahili, while the principal investigator interviewed health officials in English. Discussions included challenges with delivering PMTCT services, and perceived difficulties for patients. Observations were also conducted in communal areas of Kisesa health centre ANC and MCH clinic by the principal investigator and one fieldworker. Structured observation sheets included prompts for patient-provider interactions, privacy, procedures, patient volume and waiting times. Photographs of physical outputs and detailed notes were taken during PLAs. After each PLA or interview, debriefing sessions were held to discuss emerging themes. PLAs and interviews were digitally audio-recorded, following consent from participants, transcribed verbatim, then translated into English. Analysis was conducted in NVIVO9, guided by a framework approach [30]. Mead and Bower's patient-centred care framework (table 1) was used to organise the initial code-frame, which also incorporated aspects of patient-centred care emphasized by other researchers [22]–[24]. The code-frame was refined after applying it to the first few transcripts; adding, sub-dividing and combining codes. Coding was also done inductively to capture new concepts, and participants' own terminology was documented through in-vivo codes. Codes were then grouped into overarching themes. The resulting code-frame was applied to all transcripts (‘indexing’) by identifying content that described dialogue or other interactions between patients and providers, and the extent to which interactions were patient-centred (e.g. discussion of non-medical matters, question-asking by patients, statements suggesting empathy, included in most coding schemes for analysing verbal behaviour and patient-centred care [19]). The manner in which patients described being treated and questioned by providers (and vice versa), and links to outcomes (e.g. adherence to PMTCT components) were also scrutinized. Perceptual data, for example generated when discussing the vignette, was also analysed to reinforce perspectives from personal experiences. Charts were created using MS Excel to compare data across and within cases. The principal investigator read and coded all translated documents. A sub-sample of transcripts was double-coded (by AW) using the same initial code-frame. Each researcher's revised code-frame was then compared and discussed to ensure all emerging concepts were captured. Ethical approval was granted by the London School of Hygiene and Tropical Medicine, Tanzanian Lake Zone and Medical Research Coordinating Committee ethics review boards. All participants were informed about the study. Before commencing the PLAs, verbal consent was recorded from all participants using a digital audio recorder. It is important to offer verbal consent in this setting due to the expected low level of literacy of some of the participants and the fact that people are not used to signing their name as a form of consent. For some HIV-infected women, the existence of a signed consent might be perceived as an unjustified threat to the subject's confidentiality. All IDI participants were given the option of written or verbal consent, for the same reasons outlined. No participants recruited for interview or PLAs were minors aged <18, therefore consent was obtained from all individuals in person, rather than from next of kin, guardians or caretakers. These procedures for obtaining participant consent were approved by the ethical review boards. Recruitment procedures, using the seeded focus group method and including HIV-negative women, were designed specifically to maximise confidentiality for HIV-positive women. PLA participants were advised that the discussion was confidential, but they were not expected to share personal information. Compensation was provided for travel expenses (5000 Tanzanian shillings, 3 USD). Participant anonymity was maintained by using fictitious or generalised names, and labelling recordings, transcripts and quotations with codes (e.g. ‘health worker#1’).
N/A
