You will know that despite being HIV positive you are not alone: Qualitative study to inform content of a text messaging intervention to improve prevention of mother-to-child HIV transmission

Background: Prevention of mother-to-child HIV transmission (PMTCT) relies on long-term adherence to antiretroviral therapy (ART). Mobile health approaches, such as text messaging (short message service, SMS), may improve adherence in some clinical contexts, but it is unclear what SMS content is desired to improve PMTCT-ART adherence. Objective: We aimed to explore the SMS content preferences related to engagement in PMTCT care among women, male partners, and health care workers. The message content was used to inform an ongoing randomized trial to enhance the PMTCT-ART adherence. Methods: We conducted 10 focus group discussions with 87 HIV-infected pregnant or postpartum women and semistructured individual interviews with 15 male partners of HIV-infected women and 30 health care workers from HIV and maternal child health clinics in Kenya. All interviews were recorded, translated, and transcribed. We analyzed transcripts using deductive and inductive approaches to characterize women’s, partners’, and health care workers’ perceptions of text message content. Results: All women and male partners, and most health care workers viewed text messages as a useful strategy to improve engagement in PMTCT care. Women desired messages spanning 3 distinct content domains: (1) educational messages on PMTCT and maternal child health, (2) reminder messages regarding clinic visits and adherence, and (3) encouraging messages that provide emotional support. While all groups valued reminder and educational messages, women highlighted emotional support more than the other groups (partners or health care workers). In addition, women felt that encouraging messages would assist with acceptance of their HIV status, support disclosure, improve patient-provider relationship, and provide support for HIV-related challenges. All 3 groups valued not only messages to support PMTCT or HIV care but also messages that addressed general maternal child health topics, stressing that both HIV-and maternal child health–related messages should be part of an SMS system for PMTCT. Conclusions: Women, male partners, and health care workers endorsed SMS text messaging as a strategy to improve PMTCT and maternal child health outcomes. Our results highlight the specific ways in which text messaging can encourage and support HIV-infected women in PMTCT to remain in care, adhere to treatment, and care for themselves and their children. Trial Registration: ClinicalTrials.gov NCT02400671; https://clinicaltrials.gov/ct2/show/NCT02400671 (Archived by WebCite at http://www.webcitation.org/70W7SVIVJ). We conducted a qualitative study to inform the text message content for the Mobile WACh-X study ({“type”:”clinical-trial”,”attrs”:{“text”:”NCT02400671″,”term_id”:”NCT02400671″}}NCT02400671), a triple-arm, placebo-controlled, unblinded RCT designed to assess the impact of unidirectional and bidirectional SMS text messaging on maternal adherence, retention, and clinical outcomes in PMTCT-ART programs in Kenya [20]. Using purposive sampling, we recruited women, male partners, and HCWs from three sites; two in rural Western Kenya and one in periurban Nairobi. Focus group discussions (FGDs) were conducted with HIV-infected pregnant women seeking antenatal care (ANC) services or HIV-infected postpartum women who had an uninfected child aged ≤2 years. Women were purposively recruited during routine visits to ANC clinics, comprehensive HIV care clinics, and maternal child health (MCH) clinics. To provide a range of experiences and perspectives, we selected pregnant and postpartum women based on the following experiences with ART: using ART in the peripartum period only; using ART within and outside of the peripartum period; and no ART experience. Women were eligible to participate if they were aged ≥14 years, were HIV-infected and pregnant or postpartum, had daily access to a mobile phone, and were willing to receive SMS text messages. We conducted semistructured individual interviews with male partners and HCWs. Both HIV-infected and -uninfected male partners were recruited for participation. We recruited HIV-infected men in concordant relationships during their routine HIV clinic visits. In addition, HIV-uninfected men were referred to the study by HIV-infected female partners attending MCH clinics; female partners were given a referral form inviting male partners to the clinic to learn more about the study. Eligible men were aged ≥18 years and had an HIV-infected female partner who was pregnant or had a child aged ≤2 years and was accessing ANC or MCH services. We purposively recruited providers aged ≥18 years from ANC and MCH clinics where they worked. Men were eligible to participate if they were directly involved in caring for HIV-infected pregnant women or HIV-exposed infants. Overall, 87 women participated in 10 FGDs (6-10 women per FGD); 15 men and 30 HCWs participated in semistructured individual interviews. We conducted two rounds of data collection between January and June 2015. In the first round, 6 FGDs were conducted with HIV-infected women, 15 individual interviews were conducted with male partners, and 30 individual interviews were conducted with HCWs. The objectives of the first round of data collection were to explore general opinions about health-related SMS text messages, determine comprehension and acceptability of predeveloped text messages, and elicit ideas for additional messaging themes in order to refine the message content. A second round of 4 FGDs elicited women’s feedback on the refined message content. Both FGDs and interviews were conducted using a semistructured discussion guide including open-ended questions exploring three main topic areas: (1) challenges and resources for attending a clinic and adhering to ART, (2) perspectives on using SMS text messaging to support adherence, and (3) perceptions of specific message content to guide message refinement. We asked participants to provide feedback on messages in four content areas: general support, breastfeeding, family planning, and ART adherence. All messages shared a common format: they opened with a greeting to the recipient from a nurse (“[Name], this is [nurse name] at [clinic name]”), followed by a message addressing one of the content areas (Supplementary Material 1) [20]. Interviews and FGDs were conducted by a trained Kenyan social scientist who was not involved in providing clinical care for participants. Pilot messages were read aloud by the discussion facilitator, and participants were probed for additional message content they would like to receive, beyond what was included in initial pilot messages. Sociodemographic information for all participants was collected via a tablet-based questionnaire using Open Data Kit. Interviews and FGDs were conducted in English, Kiswahili, and Dholuo, depending on participants’ preference. FGDs ranged from 90 to 130 minutes in length, and interviews ranged from 19 to 49 minutes. All interviews and FGDs were audiorecorded, transcribed, and translated into English, if necessary, by the interviewer, who was fluent in all three languages. This study was reviewed and approved by the University of Washington Institutional Review Board and Kenyatta National Hospital and University of Nairobi Ethics and Research Committee. All study participants provided written informed consent. We performed a descriptive content analysis to identify key concepts emerging between and across groups of women, male partners, and HCWs. Dedoose software (version 7.6.6, Sociocultural Research Consultants LLC, Los Angeles, CA) was used for data management and analysis. An initial codebook was deductively and inductively generated by JF, KBS, and KR after reviewing the literature and reading a subset of FGD and interview transcripts. Next, the codebook was refined iteratively by reviewing additional transcripts and revising initial codes. We used the final codebook to perform consensus coding and facilitate discussion until reaching an agreement on the code application. All transcripts were coded independently by one team member (JF, KBS, or KR) and reviewed by another team member. All disagreements in code application were resolved through group discussion with all three coders. The analytic framework focused on challenges living with HIV, current resources or strategies used to engage in HIV care, preferences and perceived utility of the specific SMS text message content, and benefits or challenges to using SMS text messaging to engage in care. Furthermore, we identified themes related to the analytic framework categories and combined them into a conceptual diagram.