Two-way Texting (2wT) to Improve Antiretroviral (ART) Patient Retention in Malawi

In sub-Saharan Africa (SSA), antiretroviral therapy (ART) retention challenges threatens both individual health and HIV epidemic control. With significant healthcare system constraints and formidable healthcare worker (HCW) shortages, ART services in SSA are struggling to keep up with demand for, and continuity in, ART care. Within large-volume public ART clinics in SSA, mobile health (mHealth) innovations may prove critical to support retention of more patients on ART with reduced workload and lower cost.

Lighthouse Trust (LT) is a national ART Centre of Excellence in Lilongwe, Malawi, that operates with the Malawi Ministry of Health (MoH) to provide HIV care, treatment, and support across Malawi. In its two Lilongwe flagship clinics, 35,000 patients are on ART: 24,000 at Martin Preuse Centre (MPC) and 11,000 at Lighthouse (LH). Patients at both are managed via a real-time, point-of-care, electronic medical records system (EMRS). Despite its prominence, retention at MPC and LH clinics falls short: an average of 67% of ART patients are retained at 12-months, well below the 90% threshold needed for epidemic control. Since 2006, LH and MPC implemented a patient retention program, "Back-To-Care" (B2C), that traces ART patients by phone or home visit who miss a clinic visit by >14 days. B2C has been well-recognized for its success returning patients to care . Although early patient tracing increases likelihood of patient reengagement, tracing-based retention efforts are reactive -- waiting for clients to miss visits before intervention. B2C and other tracing programs are highly resource intensive, limiting scale. At MPC, there are over 7800 scheduled ART visits each month: of these, over 800 clients are presumed lost to follow up (LTFU) and referred to the 8 full-time B2C tracers. Program resource constraints cause B2C delays and tracing gaps, greatly diminishing B2C effectiveness and retention success. In July, 2019 only 33% of eligible clients were traced at MPC, a tracing gap found in other ART programs in SSA. Additional proactive retention efforts, within the existing limited resources, are needed to reduce LTFU before it happens.

Therefore, the University of Washington's International Training and Education Center for Health (I-TECH) and LT, with support from MoH and Medic Mobile implemented an innovative, proactive, patient retention system using two-way texting (2wT) between new ART patients and staff. 2wT aims to resolve potential retention issues before they occur and improve data quality (e.g. identifying transfers), reducing B2C workload. A comparison of retention (ART outcomes, visit adherence, VL) and B2C referrals among new MPC ART clients showed improved retention with possible lower costs.

The Malawi Ministry of Health Department of HIV/AIDS (DHA) requested a small-scale randomized control trial (RCT) to provide more rigorous evidence of the 2wT benefit for ART retention.

Study arms:

ARM 1: Routine new ART client and B2C care procedures (control):

At ART enrollment, the vast majority of clients (>90%) consent to potential B2C patient tracing and complete B2C patient locater forms with phones numbers, household map, and next of kin contacts. All new initiates receive augmented adherence counseling as part of ART initiation. New ART patient visits are scheduled monthly during the first six months on ART and every two or three months thereafter if the patient is stable and adherent to therapy. Those clients randomized to routine care will follow routine B2C processes. For routine B2C, patients with any ART drug dispensing record (including initiation) who missed a scheduled appointment by 14 days or more are identified using the electronic medical record system (EMRS) and considered LTFU. A team of B2C tracers manually verifies the LTFU list to rule out EMRS data errors and creates a paper-based B2C tracing list. Then, patients with completed locator forms are first traced by phone (SMS or call) up to five phone attempts and/or then up to three home visit attempts. The EMRS at ART patient registration is used daily to determine if/when a client returns to care and to stop B2C client follow-up for that specific event/month. For all patients reached, B2C interviews clients to ascertain ART outcomes and records data on paper B2C forms; data clerks enter it into the EMRS.

Arm 2:

Intervention (2wT):

Same new ART initiation and care as described above. New ART patients who are consented and randomized to 2wT will receive additional texting information and response instruction at 2wT enrollment. 2wT paper clients files will be marked to alert reception of participation. 2WT data clerk will manually enter client unique identification (ID), client phone number and scheduled visit dates into 2WT system. The 2WT system will automatically send varying, weekly adherence support texts and visit reminders (3 and 1 day before) for 12 months, with opt-out at any time. Clients will be asked to respond to visit reminder messages with a "1=will attend" or "0=no." A "yes" will trigger, "See you soon!" A "no" will trigger hybrid automated and interactive SMS follow-up where clients can request an appointment date change, report transfers, note other outcome, or request a call. These modifiable, scripted texts between 2wT staff and the client will complete ART outcome data (transfer, stop, etc.) or determine the next steps. To support visit verification, the 2wT system will produce a set of daily "tasks" (IDs of client with expected visits) for 2WT clerks to monitor visit attendance. Using these daily lists of presumed visits, a clerk will confirm visits using the EMRS, recording attended visits in the 2WT system, updating subsequent visit dates if needed. Confirmed attendance stops visit reminders for that specific instance; completion of other ART outcomes ascertained via 2WT (stop, transfer, dead) also stops all messaging. If an expected client visit was not confirmed, and outcome otherwise not updated, the 2wT system will create another alert after 5 days, triggering 2wT staff to launch an automated a series of "missed visit" SMS with similar 2WT response options every 3 days until day 14. If no visit, no text, or otherwise updated ART outcome by day 14, the 2wT system will create another alert for referral to routine B2C. B2C referrals will be compiled weekly. The 2WT-based generation of the B2C list with reported transfers removed and visits verified, will reduce delays in identification and tracing of those presumed LTFU. The system is free for clients.