Patient-centered Mobile Technology Interventions to Improve Maternal Health in Uganda

Aim 3 Overview. The investigator will pilot test the intervention to assess feasibility, acceptability and preliminary efficacy to improve ANC utilization and skilled births. Randomization of 120 adult pregnant women of gestational ages ≤ 20 weeks of amenorrhea to standard of care (information giving) (n=40), mHealth SMS intervention (n=40), and mHealth SMS intervention plus mHealth social support engagement (n=40) will be done. Feasibility and acceptability will be primary outcomes. ANC visit attendance and proportion of births attended by a skilled provider will be collected for power estimates for future studies, designed to test the efficacy of the intervention.

Study Setting: This study will take place at Mbarara Regional Referral Hospital, Mbarara, Uganda. Participants will be recruited through the antenatal Clinic, which receives over 30,000 women attending routine antenatal care annually (both normal and high-risk pregnancies). The hospital conducts over 10,000 deliveries per year. Maternity services, including delivery, are largely provided free of charge through public hospitals and health centers.

Study Procedures: Study participants randomized to either of the intervention arms will receive the CommCare application (eMobilize-Uganda) that will include message content and information developed in aim 2. Participants will get adequate instructions on how to use the application to retrieve information. The times and length for individual login sessions will be recorded and transmitted to the server. The smart phone will serve as a gateway to display and visualize intended message content in form of voice, video or text. Once visualization is complete, login data will then be transmitted from the gateway device to a secure web-based server for storage or further viewing through a web portal following a prompt to confirm completion of viewing session and log out to enable submission of data to the server for review via password access of any device that can access the web. Log-ins during periods of inadequate cellular reception will be stored for later transmission. Study participants will be given solar chargers and reminded to charge the phones as needed during enrolment. SMS will also be sent via CommCare. Application log-ins will be considered as proxies for accessing information to alter existing predisposing factors (such as negative health beliefs) that could enable and improve perceived need to seek care. Engagement of social networks will be to enable service utilization and motivate individuals to seek care through regular scheduled SMS reminders.

Randomization: Prior to study initiation, a random number generator will be used to determine arm assignment for study participants. Research assistants will be informed of the arm assignment at the time of participant enrolment. Study arms will be defined as follows: Arm 1: will only receive the standard of care (information giving), as a "Control arm". Study participants in this arm will receive no SMS reminders nor social supporter notifications. Arm 2, also called "Scheduled SMS arm": Will receive an mHealth intervention targeted to the study participant (such as health information on an eMobilize-Uganda application or messaging and SMS reminders, or a voice call if at high risk). A weekly SMS reminder on the impending ANC appointment and expected date of delivery at their preferred time and day of the week will be sent to study participants. The content of the SMS reminders will be customized and determined by everyone at enrolment. If the participant has no preference, we will suggest "This is your ANC visit reminder, encouraging you to attend". This technology is already integrated and running in Uganda via the Yo! Uganda Gateway. Arm 3: Will receive an mhealth intervention targeted to the participant plus an intervention targeted to engage the social supporter. Study participants will receive health information and SMS reminders same as those of "scheduled SMS arm" above + weekly SMS notifications to the 2 pre-identified social supporters. Notifications will bear upcoming ANC visit and delivery due date for the study participant they are supporting for all the study follow-up period (also called the "social support engagement arm"). Social supporters will be able to personalize the SMS content at enrolment. They will be advised to assist study participants with any problems that may affect ANC attendance or facility delivery, but will not be given specific instructions on what to do.

Data collection: will be performed in the local language, Runyankole. Quantitative questionnaire data will be collected from both study participants and their social supporters at enrolment on the following topics: socio-demographic characteristics: maternal age, gravidity, parity, gestational age, prenatal and antepartum high-risk morbidities, depression, health, food insecurity, alcohol use, social support, reproductive health history, perceptions on pregnancy and delivery. Reports of social support received by study participants will not specify the source as it could occur from outside the dyad studied here. Another questionnaire with closed and open-ended questions that explores specific role of social supporter in study participant's life and pregnancy during study period, communication or contact with study participant and what things they talked about or did together, the type of voluntary and requested help or support given to or received by the study participant during pregnancy towards improving her experience and attendance of scheduled ANC visits, challenges and experiences to social support will be administered at exit. At exit (within 2-4 weeks postpartum), a survey will be administered to assess the understanding, likes, dislikes and challenges, perceived usefulness, ease of use, actual use, barriers of the mHealth-based, social support intervention model aimed at improving utilization of maternity services. Other delivery outcomes (having a still birth and or maternal death), women responding to an SMS notification, number of log-ins to CommCare application, time between scheduled and actual ANC review, birth preparedness, mode of delivery, initiation of breastfeeding, 1 and 5-minute Apgar scores (if known), neonatal weight, need for neonatal or maternal resuscitation, use of assisted ventilation by mother or neonate, hospital stay and any other reported maternal or neonatal complications will be documented. ANC attendance and delivery details will be collected by reviewing the routinely provided ANC cards and postnatal discharge forms provided to confirm self-reports. Women will be followed for at least 6 months and will receive the standard routine care provided at the maternity centers. Data will be collected using an online database that will be developed in CommCare to improve data completeness, management and quality control monitoring. Data entry verification will include algorithms that automatically check completed forms for missing, out-of-range, or inconsistent values before a form can be saved on the website. Additionally, exit in-depth interviews with 30 study participants (15 from each intervention arm) and 10 social supporters will also be conducted to explore the patterns of social support and mechanisms of effect of the intervention. This selection will be based on the study participant's prenatal and perinatal outcomes, type of social support, and study participant characteristics. These interviews with be carried out within the 2-4 weeks postpartum. Interview guides will be developed with guidance from observed quantitative results, as well as two theoretical frameworks; the Unified Theory of Acceptance and Use of Technology (UTAUT) and the HUM. These interviews will last for an hour. To maximize data quality, we will ask interviewees to describe actual experiences and events wherever possible. Examples of interview topics will be: (1) Experiences: descriptions of particular experiences with the telephone and SMS type (the most recent reminder or notification, or telephone received and what happened as a result, the most or least helpful/useful reminder, notification or telephone received and what happened as a result, (2) acceptance: advantages and disadvantages of the mHealth intervention, ease of use, its ability to request or get support as needed, guidance on what to do, SMS type; problems experienced with the intervention; ideas for improvement, intention to use in future (3) consequences: changes or lack of changes resulting from use of mHealth intervention and SMS type; (4) comparisons and attitude: differences, similarities and attitude across the SMS types and other interventions.

Quantitative data analysis. I will use summary statistics to compare the health-related and socio-demographic data for both study participants and social supporters between arms. We will verify the technical function of the intervention with the following statistics: Number of successful application log-ins, number of SMS received by the participant over number of SMS anticipated per protocol, number and type of technical problems encountered, number of SMS notifications responded to by the social supporters, number of women responding to an SMS reminder or notification, time between scheduled and actual ANC review. I will assess acceptability by describing the length of application log-ins, engagement of the application, number of likes or dislikes for the intervention and reported ease of use. Although not powered to detect significant differences, I will preliminarily compare maternal health outcomes between the three study arms to consider power requirements for future larger intervention studies. These outcomes will include attendance of at least 4 ANC visits, delivery in presence of skilled personnel, still birth, and maternal death. To do so, I will fit univariable and multivariable Poisson regression models. The multivariable model will comprise of those baseline variables whose p-value will be <0.1 from the univariable model. A study arm will be included in all regression models to control for any potential differences in ANC or delivery outcomes. The study participant's social support and the social supporter's characteristics (type of relationship, income status, food insecurity, alcohol use, gender, previous obstetric history, delivery status, place of delivery, morbidity/health status, alcohol use, gender and involvement in any health or community group will be assessed as potential predictors of ANC and delivery outcomes based on potential impact on the relationship and instrumental support for the study participant. Study participant's social support will be divided into instrumental and emotional support. Interactions between instrumental support and food insecurity using the recommended Household Food Insecurity Access Scale (HFIAS) and income status will also be assessed because of the low resource nature of this setting which may impede the ability to provide physical support despite the intension to do so. Data analysis will be conducted in STATA version 13.

Qualitative data analysis: In-depth interviews will be digitally recorded and transcribed. The goal of the interviews will be to document the participants' experiences of the mHealth-based social support intervention, mechanisms of effect and or contextual explanations observed within the quantitative results, acceptance and patterns of social support. Qualitative analysis will be inductive and categories will be derived from the different social supporter/ study participant interviews and open-ended questions. These responses will be transcribed into English and coded using Atlas.ti. Categories will then be developed and presented with illustrative quotes from study participant and social supporter data to explain the role, type, challenges, acceptance and experience of the intervention during the study period.