Patient and provider perspectives on how trust influences maternal vaccine acceptance among pregnant women in Kenya

Background: Pregnant women and newborns are at high risk for infectious diseases. Altered immunity status during pregnancy and challenges fully vaccinating newborns contribute to this medical reality. Maternal immunization is a strategy to protect pregnant women and their newborns. This study aimed to find out how patient-provider relationships affect maternal vaccine uptake, particularly in the context of a lower middle-income country where limited research in this area exists. Methods: We conducted semi-structured, in-depth narrative interviews of both providers and pregnant women from four sites in Kenya: Siaya, Nairobi, Mombasa, and Marsabit. Interviews were conducted in either English or one of the local regional languages. Results: We found that patient trust in health care providers (HCPs) is integral to vaccine acceptance among pregnant women in Kenya. The HCP-patient relationship is a fiduciary one, whereby the patients’ trusts is primarily rooted in the provider’s social position as a person who is highly educated in matters of health. Furthermore, patient health education and provider attitudes are crucial for reinstating and fostering that trust, especially in cases where trust was impeded by rumors, community myths and misperceptions, and religious and cultural factors. Conclusion: Patient trust in providers is a strong facilitator contributing to vaccine acceptance among pregnant women in Kenya. To maintain and increase immunization trust, providers have a critical role in cultivating a positive environment that allows for favorable interactions and patient health education. This includes educating providers on maternal immunizations and enhancing knowledge of effective risk communication tactics in clinical encounters. This research was part of a larger in-depth, mixed-methods, multi-tiered, national study on maternal vaccine acceptance in Kenya. A research question within this larger study was the role of healthcare providers on maternal vaccine acceptance. This paper focuses on the findings garnered from pregnant women and healthcare providers’ (HCPs) interviews; it specifically addresses how patient trust within the provider-patient relationship affects maternal vaccine acceptance among pregnant women. Several factors made Kenya a fitting study-site for this project, the first being the recent re-emergence of anti-vaccine rhetoric against the tetanus vaccine. Secondly, Kenya has a large birth cohort (over 1.5 million births in 2012) and has a long standing history / partnership with the CDC that’s resulted in a strong, surveillance system. Lastly, there is considerable regional, cultural, religious, and tribal diversity in Kenya, which enabled our team to gather a variety of perspectives from those who work in the maternal/child health field. Thus, four field sites representing both urban and rural settings were selected, allowing for the examination of cultural, economic, and geographical differences within Kenya (Table 1). Clinics were selected based on their geographic location within the study areas and previous partnerships with the Kenya Medical Research Institute (KEMRI) – Centers for Disease Control (CDC) collaboration. Site Descriptions A major tourist area, Mombasa is located in the south east of Kenya and represents a semi urban setting. Accessibility to healthcare facilities differs on proximity to Mombasa town; Health facilities included Coast General Provincial Hospital (a KEMRI/ CDC influenza site). It is the second largest public hospital in Kenya with a bed capacity of 672. There are about 80 healthcare staff members working in child and maternal health. In 2015, the maternal clinic saw 1882 new patients and 4226 returning patients. This hospital is part of the influenza surveillance platforms. As the capital city of Kenya, Nairobi represents the urban setting. Accessibility to different types of health facilities (private/ public) is higher in Nairobi than anywhere else in the country. Health facilities included Mbagathi District Hospital which is located in Nairobi right next to KEMRI and CDC Kenya. This hospital has a bed capacity of 320. There are about 53 healthcare workers handling maternal and children issues. The average number of pregnant women seen per month at the hospital is 500. Located in Northern Kenya, it is a hard to reach and sparsely populated area. It is the most unique of all four locations as it is primarily composed of a nomadic community. Accessibility to any health care facilities is poor. Health facilities included Marsabit District Hospital which is located in the north Eastern part of Kenya. The hospital has a capacity of 86 beds. This location allowed the team to access a different population seen at the other hospitals. Located in Western Kenya and close to Kisumu, Siaya represents the rural setting. Accessibility to healthcare facilities differs depending on proximity to Kisumu which is another major city in Kenya. Health facilities included Siaya County Referral Hospital which serves a large number of rural and low social economic patients. The bed capacity is 200. There are 26 health care workers within the maternal clinic that care for and see about 300 to 400 new and returning pregnant women. It is located about 72 Km from The Centre for Global Health Research at KEMRI Kisumu Field Station Both pregnant women and healthcare providers were selected through convenience sampling methods at clinical facilities (Table 2). Pregnant women were approached and consented at the clinics by study personnel. Healthcare workers were initially approached at the clinic; however, due to a nurses’ strike in 2017 when the data collection was ongoing, the protocol was adapted and most of the interviews were conducted at the homes of the healthcare providers. Interview Protocol Pregnant women: pregnant women waiting for their scheduled antenatal care visits at the clinics were approached by research members and asked if they willing to participate in the study. If they were willing, they were taken to a private room/office designated by the hospital for confidential consenting and interviewing. After allowing time for consent review and answering questions, the study team recorded each interview. • Inclusion criteria: Women aged 15–40 · Women in any trimester; Patient at health facility included in the study; Be willing to converse with others in a focus group format (only for message testing phase); Able to provide informed consent (If participant is illiterate, procedures to ensure full understanding of the research and consent process will be implemented according to international and federal guidelines). • Exclusion criteria: have previously participated in this study; Those who do not or cannot provide consent; Failure to meet other inclusion criteria. • Interview Topics: Interview guide for women included discussions on the following topics: (a) have they had a checkup in the last year; (b) have they received any vaccines that they can recall; (c) if they have received vaccines, they will be asked about their understanding of the vaccines they received [e.g., do they know what the vaccine prevents against, did they get the vaccine just because their parent/doctor told them to]; (d) comfort discussing sensitive topics with their doctor and parents; (e) awareness of maternal vaccines; (f) information from peers about maternal vaccines [friends’ vaccination status, anecdotal side effects, discussions on social media, reasons to get it/not get it]; (g) discussions with parents/guardians about maternal vaccines; and (h) motivating factors to be vaccinated. HCPs: providers working at the antenatal care were contacted ahead of time to arrange interviews for times that would work best for them. During this phase of data collection, HCP were on a nationwide strike. To mitigate the effects of delayed data collection, study team members organized interviews outside of clinic. • Inclusion Criteria: Currently working at the selected study sites; Current physician, nurse, nurse midwife, community health worker; Able to provide informed consent. • Exclusion Criteria: Those who do not or cannot provide consent; Failure to meet other inclusion criteria. • Interview Topics: The semi-structured interview guide for providers included discussions on the following topics: (a) proportion of patients they estimate have received or refused maternal vaccines; (b) times at which they recommend maternal vaccine; (c) practices regarding immunization history verification (e.g. immunization information system); (d) barriers or reasons for refusal cited by parents/patients; (e) perceived ability and methods used to address these barriers/refusals; (f) comfort discussing vaccine recommendations with their patients; (g) existing efforts of reminder/recall for maternal vaccinations; and (h) knowledge of Tdap vaccine effectiveness and safety. Semi- structured, in-depth interview guides were developed using grounded theory [16] by the study’s lead anthropologist with input from the study team, in-country partners, and a scientific advisory committee. To understand the social determinants of maternal immunization acceptance among pregnant women in Kenya, open-ended questions explored socio-cultural practices customs, values and beliefs (Additional files 1 and 2). During the initial pilot-testing phase, the guides were reviewed, revised, and updated as new themes emerged. An expert anthropologist trained in-country research team members in qualitative methods including protocol adherence, screening, consenting, qualitative interview methods, transcription and translation. The inclusion and exclusion criteria are described in Table ​Table2.2. Interviews were conducted in either English, Swahili, Kikuyu, Luo, or Borana depending on site and preference of the interviewee. Interviews lasted approximately 30 to 60 min, depending on the extent of the discussion, and were audio recorded. Interviews were conducted in teams of two, allowing for an interviewer and note taker (Table ​(Table2).2). Both KEMRI IRB and Emory IRB approved the study protocol. CDC IRB reliance on Emory’s IRB was obtained. Interviews were transcribed and translated (when necessary). The analysis was done using N-Vivo 11.0 qualitative data analysis software. Identifiable information such as names, dates, and addresses were removed from both the recordings to maintain participant confidentiality. Prior to coding, the qualitative research team developed two codebooks including one for providers and one for pregnant women’s interviews. Once the codebooks were completed, codes were applied to the transcribed interviews; major thematic content emerged from this process. Intercoder reliability was performed among three coders participating in the coding and analysis. A kappa coefficient of ≥0.80 was considered a minimal cut-point for high intercoder agreement among coded content to maintain rigorous qualitative research standards. After 4 rounds of intercoder testing, the team achieved k ≥ 0.80 agreement on codes used in this analysis.