PMTCT care engagement as a social practice and system : insights from an mHealth intervention and routine tracing in western Kenya

Background: Effective engagement in prevention of mother-to-child transmission (PMTCT) care is crucial to reduce vertical HIV transmission rates and improve health outcomes for women and their infants. Kenya remains one of seven countries contributing to 60% of all new pediatric HIV infections globally, highlighting persistent challenges in PMTCT despite global and local efforts. Although antiretroviral therapy (ART) has become more accessible to women of reproductive age living with HIV, challenges remain in engaging and retaining women in PMTCT care in resource-constrained settings like Kenya. Social stigma, poverty, and deeply rooted gendered power imbalances can hinder consistent care engagement. Little is known about routine tracing of women who miss their scheduled PMTCT clinic visits and how these approaches could integrate with innovative solutions like mHealth to support continuous care.

Aims: To investigate how interactive text-messaging and routine defaulter tracing influence PMTCT care engagement in western Kenya, and to examine how such interventions and policies interface with existing social and health system practices.

Methods: A mixed-methods design was employed, combining a longitudinal cohort study and qualitative research nested within the WelTel-PMTCT randomized control trial conducted across six facilities from July 2016 to February 2019. Study I involved a cohort of 299 women who received weekly interactive text messages over 24 months. Data were collected through questionnaires, logbooks, and data extracts from the intervention platform. Descriptive statistics, multivariable-adjusted logistic regression and negative binomial models were used to explore associations between sociodemographic characteristics and intervention adherence. Study II used semi-structured qualitative interviews and observations with women in care (n=14) and health workers (n=14) to explore their perceptions and experiences of the WelTel-PMTCT intervention. Thematic network analysis and social practice theory were used in the analysis. Study III draws on narrative interviews (n=15) with 10 women with intermittent care engagement and repeated tracing. A compound approach to narrative analysis, combined with an intersectionality lens, was used to uncover factors influencing these women's PMTCT care engagement. Study IV explored defaulter tracing through semi-structured interviews with 19 front-line tracers, supplemented by observations, key informant interviews and document reviews. Thematic network analysis was used, informed by maintenance and repair theory.

Results: Out of the total 31,638 text messages sent as part of the intervention, women responded to 49%, with 48% indicating no issues ("okay" responses) and 1% reporting problems. Younger women (18-24 years) were more likely to respond to <50% of all messages (OR: 2.20; 95% CI: 1.03-4.72) compared to older women. Those with secondary rather than lower education were also more likely to report problems (OR: 1.88, 95% CI: 1.08-3.27). Response rates decreased over time, from 64.5% initially to 38.5% after median 27 months, reflecting possible intervention fatigue (Study I). The interactive text message intervention showed a promising "fit' with existing PMTCT practices by bolstering psychosocial support and fostering stronger relationships between women and health workers. This alignment encouraged a more person-centered approach to care, reducing reliance on informal communication and formalizing phone-based interactions. However, structural challenges, such as network instability and HIV-related stigma, hindered sustained care engagement (Study II). Women with intermittent care engagement encountered intersecting barriers, including stigma, gender imbalances, poverty, and caregiving burdens, with variations observed between married women and young, single women. Social support networks were a motivating factor for care re-engagement and well-being across both groups (Study III). Tracing efforts were variably implemented across the health facilities and relied largely on the resourcefulness of health workers. Without adequate institutional support and formal evaluations, tracing often depended on informal 'repair work' by healthcare workers and was challenged by systemic stigma (Study IV).

Conclusion: PMTCT care engagement in the Western Kenyan context is inherently social and non-linear, shaped by women's broader life circumstances and gendered power imbalances and stigmas. Findings from this thesis illustrate how interactive text messaging and defaulter tracing can enhance PMTCT care engagement. In particular, it highlights the potential of technologies like interactive text messaging to deliver 'warm care' that resonates with women's lived experiences. Effective implementation requires these technologies to work in tandem with existing healthcare systems. However, implementation of defaulter tracing relied on the ingenuity of health workers who were without proper institutional support - underscoring a need for urgent 'repair' via adequate resource allocation to frontline health workers who keep systems going. Overall, this thesis underscores the need for more nuanced, comprehensive approaches to measuring and understanding PMTCT care engagement. Interactive text messaging and defaulter tracing alone cannot resolve the systemic issues of poverty, stigma, and gender imbalances, but they can play a vital role in connecting women to care and foster stronger social bonds. The success of such interventions likely depends on their integration into holistic, person- centred care models that address broader socio-structural barriers. When implemented thoughtfully, mHealth and tracing efforts have the potential to support sustained engagement in PMTCT care and improve health outcomes for women living with HIV and their children.

List of scientific papers

I. Nordberg B, Kaguiri E, Chamorro De Angeles KJ., Gabriel E, van der Kop M, Mwangi W, Lester RT, Were E, Ekström A, Rautiainen S. The use, adherence, and evaluation of interactive text-messaging among women admitted to prevention of mother-to-child transmission of HIV care in Kenya (WelTel PMTCT). BMC pregnancy and childbirth.2024;1 (25). https://doi.org/10.1186/s12884-023-06194-0

II. Chamorro De Angeles K, Oware P, Were E, Ekström AM, Skovdal M, Kågesten A. Care from distance, yet closer together: How an SMS intervention enhanced care engagement for prevention of mother-to- child HIV transmission care in Western Kenya. SSM Qualit. Res. Health. 2022;2. https://doi.org/10.1016/j.ssmqr.2022.100195

III. Chamorro De Angeles KJ, Kågesten A, Skovdal M, Ekström A, Dusabe- Richards J, Kaguiri E, Were E, Faxelid E. Navigating intersecting identities and stigmas - narratives from women who drop in and out of PMTCT care in western Kenya. [Submitted]

IV. Chamorro De Angeles KJ, Storey S, Nordberg B, Kaguiri E, Dusabe-Richards J, Ekström A, Were E, Skovdal M, Kågesten A. Care as repair: work force resourcefulness and adaptation in implementing a PMTCT defaulter tracing policy in western Kenya. [Submitted]