Evaluating the impact of targeted health mobile messages on modern family planning continuation among a peri-urban population in Nairobi: a randomized control trial

Participants randomized to the SMS treatment arm will receive a supplemental package of SMS messages. The SMS message content includes: information and counseling on danger signs; counseling messages on common side effects, including information on typical duration; messaging to dispel misinformation spread through local myths and misconceptions around contraception; reminders to seek routine follow-up care according to method-specific protocols; and information about alternative modern contraceptive methods in the case of dissatisfaction. Each message with conclude with an encouragement for the participant to send a reply SMS for free to receive a phone call from a nurse. Message content has been developed to reinforce clinical protocols on facility-based side effects care and counseling provided to follow-up family planning clients at the partner hospital sites. Specific message streams have been developed for each of the following types of modern contraceptive methods: intrauterine contraceptive device (IUCD), hormonal implants, oral contraceptive pills (both combined and progestin-only), progestin-only injectable, and male condoms. Message order for each stream is based on literature on side effect timing and prevalence, as well as local partner's clinical experience.

The control arm will receive the standard family planning counseling and information brochures, including information about follow-up visits for any routine check-ups or for appropriate method continuation.

2015-04-08

2016-12-31

Contraceptive continuation and fertility outcomes:
• Family planning clients reporting continuous use of modern contraceptive method at 6 months
• Family planning clients reporting modern contraceptive switching (defined as a participant discontinuing use of one modern method, but obtaining another modern method at the time of discontinuation of the original method)
• Duration of continuous use of modern contraception (including contraceptive switching)
• Family planning clients reporting switching to long-acting reversible contraception
• Family planning clients reporting unintended pregnancy at 6 months

Care-seeking behaviors:
• Sought family planning-related follow-up care
• Reported experienced method-related side effects
• Sought care for side effects management/advice

Contraceptive knowledge, attitudes, and beliefs
• Knowledge of contraceptive danger signs (number able to name)
• Knowledge of typically harmless side effects (number able to name)
• Knowledge of common myths and misconceptions about contraceptives

Self-reported satisfaction and service quality
• Receipt of high-quality family planning counseling
• Family planning users reporting satisfaction with family planning services

The "Knowledge of common myths and misconceptions about contraceptives" construct will be measured using the following items:
o Can using any method of family planning cause permanent infertility? (Which methods can cause permanent infertility?)
o Can using any method of family planning cause cancer?

The "Family planning users reporting satisfaction with family planning services" construct will be measured using the following items:
o Reports being asked if she had any questions about family planning
o Self-reported satisfaction with services using 5-point Likert scale
o Reports feeling encouraged to discuss any questions or problems with a nurse
o Reports understanding the advantages and disadvantages of the method she chose
o Family planning clients who believed healthcare workers addressed her concerns about contraception
o Family planning clients who believed that all questions about contraception were answers to her satisfaction

The "Receipt of high-quality family planning counseling" construct will be defined as participants reporting that they received counseling on the following topics:
o How to use the method?
o Any information about side effects or problems you might have with the method?
o What to do if you have any side effects or problems?
o When to return for a follow-up visit?
o Other options of family planning methods you could use?

The study will be a randomized controlled trial with two study arms: clients who will receive the standard of care for family planning service provision offered at the partner organization's private hospitals, and clients who will additionally be invited to participate in a mobile-based platform for enhanced family planning counseling and support over a duration of 7-weeks. Follow up surveys will be administered at 3 months, 6 months and 1 year. The first follow up survey will be cross randomized at 3 months with study arms: clients who will receive the first follow up survey at 3 months postpartum and clients who receive their first follow up surveys at 6 months postpartum.

Women between 18 and 40 years old are eligible for inclusion if they have sought care at a Jacaranda Health facility and obtained a new method of family planning, defined as a method that the client was not using directly prior to her consultation. Women must provide a phone number by which they can receive SMS message and phone calls.

Eligible participants will include clients obtaining a new method of modern contraceptive methods (defined as IUCD, implant, oral contraceptive pills, hormonal injectables, or male condoms) for immediate uptake at one of the local partner's maternity facilities located in Kenya’s Kiambu county. A “new” method in this case is defined as a method that the client was not already currently using at the time of her consultation at Jacaranda. Clients receiving follow-up or continuation care (i.e. a client returning for a routine hormonal injectable or additional supply of oral contraceptive packets) will not be eligible for enrollment. Additionally, clients who wish to obtain a method but were not able to based on contraindications or medical eligibility criteria reasons will not be eligible if they are not able to obtain the method at the time of the Jacaranda consultation. Furthermore, although this has not been previously observed at Jacaranda facilities, clients obtaining methods with the intent to delay use (for example, pregnant women aiming to obtain a method for use after delivery) will not be eligible. Eligibility will not be limited to postpartum women (defined as women who have delivered in the past 1 year), with nulliparous, primiparous, and multiparous women eligible for participation if other eligibility criteria are.

Randomization will be conducted using reproducible code (using STATA software). Randomization is not stratified, but is rather simple random randomization using patient ID numbers assigned by the partner organization sequentially by time of first visit to the facility.

The unit of randomization is the individual.

No

0 clusters. All clients will be recruited at maternity hospitals operated by the local partner organization.

The relevant sample for our SMS intervention will be 600 women.

We aim to enroll equal groups of 300 women in the control and intervention arms.

We use two primary outcomes of interest in order to calculate power. Our first outcome is the probability of discontinuation by 6 months, defined as the percentage of women who are not using the original method (method adopted at enrollment) for a continuous period of 6 months. Using Jacaranda’s administrative data, we assume that the method mix at original uptake will reflect Jacaranda family planning clients since October 2013. We thus assume that 10% of our sample will obtain pills (combined or progestin-only), 1.4% male condoms, 28.8% injectables, 41.8% implants, and 18% IUCDs. Assuming an exponential decay in contraceptive continuation and using discontinuation probabilities from the Demographic and Health Surveys, we assume an average probability of discontinuation by 6 months in a sample using this method mix to be 11.9% (KDHS 2008/2009; Ali, Cleland, and Shah 2012; FHI360). We would then have 80% power to detect a decrease in discontinuation at 6 months of 6 percentage points. This change would represent a decrease in the overall probability of discontinuation at 6 months to 5.5%, which would represent a significant reduction in discontinuation. We believe this reduction is plausible, given 50-80% of discontinuation is due to the same types of method-related problems (such as dissatisfaction, inconvenience, side effects or health concerns) that our intervention aims to address (Ali, Cleland, and Shah 2012). Our second outcome is receipt of follow-up care related to contraceptive use by 6 months. Currently, 19% of Jacaranda’s family planning clients return for a second family planning consultation. With 600 women in our study, assuming equal groups of 300 women, we will have 80% power to detect an increase in the probability of returning for care with 6 months of use of 9% points. This change would represent a significant improvement in contraceptive follow-up care, and is believed to be feasible given similar effects sizes on appointment follow-up caused by mHealth reminder interventions in Kenya (Odeny et al. 2012).

Pre-Analysis Plan - 1 April 2015