MAMAS: Mentoring Adolescent Mothers at School

All study activities will take place in Umlazi township in eThekwini district. The township is served by one regional hospital, Prince Mshiyeni Memorial Hospital, which is the referral hospital for 17 primary health care clinics.

Participants will be involved in the study from delivery to 9 months postpartum. The intervention study will be delivered in two cohorts (n=240 per cohort). The first cohort will be enrolled beginning in May 2017. The investigators anticipate that 4 adolescent mothers will be enrolled per day based on the monthly number of adolescent deliveries at Prince Mshiyieni hospital and our own capacity. Therefore, investigators will be finished enrolling the first cohort by July 2017. The first cohort will complete the end line assessment between December 2017-February 2018.

The investigators will begin enrolling the second cohort in September 2017 and will complete enrollment by November 2017. The second cohort will complete the end line assessment between June-August, 2018. Primary outcome analysis will be complete by October, 2018.

MAMAS Intervention The investigators will roll out the program with two cohorts of adolescent mothers (n=240 per cohort); each mentor mother will work with 10 adolescent mothers per cohort. Within each cohort, there will be an intervention arm and control arm (N=240 per arm for the both cohorts). The participants of the program arm will have 15 group sessions with mentor and up to 2 home visits from 3 weeks-6 months postpartum. Investigators will use participatory visual methods (e.g., photovoice, journaling, and drawing) throughout program implementation with adolescent mothers to understand and further facilitate their process of resilience development. As one example, as part of the mentoring process, adolescent mothers may be asked to engage in reflexive journaling. Mentor mothers will ask adolescent mothers who are comfortable doing so to share excerpts from their journals in a particular group session. Through the group session, the adolescent mothers themselves will identify common barriers they are facing and will work together to overcome them. Mentor mothers will keep weekly logs documenting what was covered in the sessions; these will be abstracted as part of ongoing monitoring and evaluation.

All 480 participants will complete a baseline behavioral and biological assessment (HIV and HSV-2 testing) at 2 weeks postpartum and an additional baseline biological assessment (Trichomonas vaginalis, Neisseria gonorrhea and Chlamydia trachomatis testing) at 6 weeks postpartum. Further, all participants will receive regular postpartum and well-baby visits at 6 weeks, 10 weeks, 14 weeks and 6 months with an adolescent-friendly study nurse. During the visits, participants will complete routine assessments. At 9-months postpartum, participants will complete end line behavioral and biological assessment (HIV, HSV-2 Trichomonas vaginalis, Neisseria gonorrhea and Chlamydia trachomatis testing).

Provision to Privacy All participants will receive adolescent-friendly clinical services with our study nurse during the postnatal period. The study nurse will provide standard of care clinical services, and will have gone through training to ensure she is providing youth-friendly care. In addition, a trained research assistant will administer all behavioral assessments and be available for any queries that participants have as they complete the survey. The research assistant will complete training on multicultural sensitivity, gender-based violence, and ethical guidelines as recommended by the WHO Violence Against Women (VAW) study to engage with participants in a supportive and non-judgmental manner, deal with sensitive issues, and respond to distressed participants.

Additionally, the following systems are in place to maintain cohort retention and follow up while protecting the privacy interests of subjects: At enrollment, study participants will be asked if they are willing to share their telephone number, their physical home address and a postal address. Women will be told that providing this information is entirely voluntary, and will only be used to follow up with the women to remind them of their appointments. Study participants in the intervention trial will be contacted by text message prior to a study visit to remind them of their visit. Study participants who miss their scheduled appointment are contacted within 24 hours. The first line of contact is an SMS. The second is a reminder telephone call to the participant directly or to a close contact that is identified by the client, while maintaining confidentiality. If the study staff cannot reach the participant, she will call once per day over the next three days. If the tracer still cannot reach the participant by the end of the third day, the tracer will do a home visit if home location was provided by the participant. Home visits are done by tracers familiar with the residential areas. For the proposed study, the investigators will employ two full-time tracers. If the participant has changed her address maximum information is gathered (while maintaining confidentiality of the participant) from neighbors and relatives that form the basis of further contact. A participant will be considered lost to follow-up after 3 unsuccessful attempts of phone contact and 2 unsuccessful home visits. A SMS and phone script that will be used to contact women and remind them of their appointments.

A similar system for tracing study participants was used in the previous clinical trials implemented and has resulted in retention rates of higher than 75%, with no adverse events associated with tracing reported.

Data Management and Analysis Data will be primarily managed on-site by a data entry and management team. All risks will be minimized through staff ethics training, layered data security and by restricting access of sensitive information to staff members who have a specific need for the information. Furthermore, as part of the training that investigators conduct for all staff members, investigators will include a training module on the confidential nature of the data. Investigators have conducted similar training modules for research staff in other studies. Following the training, all staff will sign an Oath of Confidentiality.

Data Security for all Data All participants will be assigned a participant identifier (PID) after randomization. Thereafter, all biological and non-biological samples/data will only be identifiable vis-a-vis the PID. Participant's names will never appear on the samples/data. Further, consent forms and any other forms with personal identifiers will be kept separate from data files in key-locked filing cabinets. A computerized follow-up system which links identification numbers and personal information will be kept on a computer with password protection. Only the site PI, Dr. Daya Moodley and authorized study staff who have completed all required ethics training will have access to the computer. Protection against loss of data is provided through scheduled backup of data files.

Non-biological Data Data will be collected using Samsung Galaxy or equivalent Tablets running customized version of Android operating system. The investigators have multiple levels of security measures to ensure the safety of data on the tablets. With the customized version of operating system, the interviewers are allowed access to only certain apps and settings on the tablets. A password is required to activate the tablet and a second password must be entered to access the data capture software. There are also separate levels of access for interviewers, field supervisors, and system administrators. Only users with administrator level access will can delete data that had been previously captured. Further, the data captured on the device will be in a machine-readable (xml) format and is not easy to interpret without using a computer program. To add another layer of security, once the data has been transferred to the project server, it will be removed from the device. As an additional security measure, lost or stolen tablets can be tracked and have their data erased remotely by the data team when the lost or stolen tablet is connected to a Wi-Fi or cellular network.

All data captured by the tablets will be transmitted to the project server using HTTPS, which is a widely-used communications protocol for secure communication over a computer network. HTTPS provides authentication of the web app and associated web server that one is communicating with, which protects against malicious attacks. Additionally, it provides bidirectional encryption of communications between a client and server, which protects against eavesdropping and tampering with and/or forging the contents of the communication.

Once the data is on project server, it allows real time monitoring of progress in the field, reported via automated Benchmark reporting tables on a password protected website, and facilitates monitoring of the status of data collection, and helps make rapid corrections to assure timely adherence to the timeline. In addition, the data will be processed to create incremental Stata datasets every week and will be sent through email in an encrypted format to Project coordinator, who in turn will run weekly QC activities and return any queries to field staff.

Data Collection, Storage and Sharing for Biological Data Biological specimens will be stored at the CAPRISA repository in -70 refrigerators. The samples will be stored in temperature controlled and monitored environment. Specimens will be stored until all assays are complete which is dependent on the receipt of funds for the relevant assays proposed. Repository staff who maintain the samples daily will have access to the samples.

Statistical Analysis The investigators will use appropriate bivariate and multivariate analysis to test whether adolescent mothers who received the intervention have lower levels of STIs and HIV risk than those in the control arm. Multivariate analysis will be used to measure the association between treatment arm and school reenrollment. For HIV-positive participants, investigators will assess whether receiving the intervention increases retention in care. Analyses will be completed using SAS statistical programs.

Power Analysis Investigators used data from a previous trial in the same setting to help estimate the appropriate sample size for our RCT [1]. Specifically, in the previous study, investigators had a 26.5% prevalence of consistent condom use and a 21.6% prevalence of any STI in the control group at the postpartum assessment.

Outcome 1: STI. Investigators estimate that the total sample size needed at follow-up to detect a standardized effect (d) of .20 (OR=.69) of our intervention on STI with 80% power (2-sided, alpha=.05) is 350 (total across both arms). The investigators inflated this sample size assuming an attrition rate from baseline to follow-up of 25% (350/.70=467).

Outcome 2: Inconsistent condom use. The investigators estimate that the total sample size needed at follow-up to detect a standardized effect (d) of .20 (OR=.69) of our intervention on inconsistent condom use with 80% power (2-sided, alpha=.05) is 314 (total across both arms). Investigators inflated this sample size assuming an attrition rate from baseline to follow-up of 25% (314/.70=419).

While investigators estimated that 467 participants would provide sufficient power to detect an effect on STI outcomes, investigators increased our sample size to 480 participants so that each of the 12 mentor mothers would be delivering the intervention to an equal number of participants (n=10) in each cohort.