Adaptive evaluation of mHealth and conventional adherence support interventions to optimize outcomes with new treatment regimens for drug-resistant tuberculosis and HIV in South Africa (ADAP-TIV): Study protocol for an adaptive randomized controlled trial

Jesse Ross, Rubeshan Perumal, Allison Wolf, Mbali Zulu, Kevin Guzman, Boitumelo Seepamore, Karl Reis, Hlengiwe Nyilana, Senzo Hlathi, Radhamoney Narasimmulu, Ying Keun K Cheung, K. Rivet Amico, Gerald Friedland, Amrita Daftary, Jennifer Zelnick, Kogieleum Naidoo, Max R. O’DonnellResearch Square 2023 · DOI

TL;DR

The ADAP-TIV trial is a pioneering study protocol designed to optimize treatment adherence for patients co-infected with drug-resistant tuberculosis (MDR-TB) and HIV in South Africa. By utilizing a Bayesian adaptive trial design and a Differentiated Service Delivery (DSD) framework, the study evaluates whether combining mHealth (electronic pill boxes) with psychosocial support can significantly improve long-term clinical outcomes compared to standard care.

What Problem Does This Paper Address?

We are currently in a "golden age" of therapeutics for MDR-TB and HIV. The introduction of all-oral, shorter-course regimens containing bedaquiline and highly effective integrase inhibitors (INSTIs) has radically transformed the prognosis for these diseases. However, these medical breakthroughs face a human hurdle: adherence.

In high-burden settings like KwaZulu-Natal, South Africa, the "syndemic" of MDR-TB and HIV requires patients to manage complex, long-term medication schedules. Traditional "Directly Observed Therapy" (DOT)—where a healthcare worker watches a patient swallow pills—is often rigid, paternalistic, and resource-heavy. The gap the authors address is the lack of a scalable, patient-centered model that provides the right amount of support to the right patient at the right time.

Methods

The ADAP-TIV study is a prospective, four-arm adaptive randomized controlled trial (ARCT). The researchers are moving away from the "one-size-fits-all" approach in favor of two novel strategies:

  1. Differentiated Service Delivery (DSD): Instead of static interventions, the intensity of support is "titrated." If a patient’s adherence scores (measured via questionnaires or pill box data) drop, the intervention automatically scales up—for example, moving from monthly check-ins to biweekly home visits.
  2. Bayesian Adaptive Randomization: This is a departure from standard fixed-ratio randomization. After an initial "run-in" period with 40 participants, the study uses interim data to modify randomization weights. Arms that show better preliminary outcomes (at 4 months) will receive a higher proportion of subsequent participants. This makes the trial more efficient and ethically sound by favoring the more effective treatments mid-study.

The Four Arms: * Arm 1: Enhanced standard of care (literacy and training). * Arm 2: Psychosocial support (counseling and support groups). * Arm 3: mHealth (using Wisepill RT3000 cellular-enabled electronic pill boxes that trigger SMS reminders). * Arm 4: Combined mHealth and psychosocial support.

Datasets

As this is a study protocol, the paper does not analyze a pre-existing dataset but defines the parameters for a new one.

  • Target Population: 360 adult patients in Durban, South Africa, co-infected with MDR-TB and HIV.
  • Data Points: The primary "composite" dataset will include survival status, TB culture conversion (negative results), retention in care, and HIV viral load suppression at 12 months.
  • Adherence Data: High-resolution longitudinal data will be generated from the Wisepill devices, providing a surrogate for actual ingestion behavior.

Key Insights and Ideas

  • Titration as a Resource Optimizer: The authors argue that not every patient needs intensive intervention. By using a DSD framework, they can reserve expensive resources (like home visits) for those struggling most, potentially making the model more sustainable for public health systems.
  • The mHealth Surrogate: The use of electronic dose monitoring (EDM) allows for real-time intervention. Instead of waiting for a patient to fail a blood test months later, the system can trigger an SMS or a call the moment a pill box isn't opened.
  • Ethical Trial Design: By using adaptive randomization, the researchers minimize the number of participants assigned to a potentially inferior "standard of care" arm once evidence of a superior intervention begins to emerge.

Key Results

While the trial is ongoing (recruitment began in March 2023), the protocol establishes the statistical power needed to detect success.

  • Success Benchmarks: The researchers estimate a baseline success rate of 70% for standard care. They have powered the study to detect if the combined intervention (Arm 4) can push that success rate to 90%.
  • Primary Endpoint: A composite of clinical and biological markers at 12 months, providing a holistic view of patient health rather than just a single lab value.

Limitations and Future Work

The authors candidly acknowledge several challenges:

  • Cost Barriers: The Wisepill devices cost approximately $100 USD each. While they can be reused, this price point may be prohibitive for immediate wide-scale rollout in low- and middle-income countries (LMICs).
  • Generalizability: The unique intensity of the TB/HIV syndemic in South Africa means results might not translate perfectly to other regions, such as Southeast Asia, where the healthcare infrastructure and cultural drivers of adherence differ.
  • Behavioral "Underdosing": There is a risk that even the most intensive support won't overcome deep-seated structural barriers like extreme poverty or severe mental health challenges.

Future Work: If successful, the team plans to conduct a formal cost-effectiveness analysis and explore the use of less expensive electronic dose monitoring tools to make the "ADAP-TIV" model viable for national health programs.

Summary updated

Abstract

This trial will evaluate the contribution of different modes of adherence support on MDR-TB and HIV outcomes with WHO recommended all-oral MDR-TB regimens and ART in a high-burden operational setting. We will also assess the utility of a DSD framework to pragmatically adjust levels of MDR-TB and HIV treatment support.
BibTeX
@inproceedings{ross2023adaptive,
  author = {Jesse Ross and Rubeshan Perumal and Allison Wolf and Mbali Zulu and Kevin Guzman and Boitumelo Seepamore and Karl Reis and Hlengiwe Nyilana and Senzo Hlathi and Radhamoney Narasimmulu and Ying Keun K Cheung and K. Rivet Amico and Gerald Friedland and Amrita Daftary and Jennifer Zelnick and Kogieleum Naidoo and Max R. O’Donnell},
  title = {{Adaptive evaluation of mHealth and conventional adherence support interventions to optimize outcomes with new treatment regimens for drug-resistant tuberculosis and HIV in South Africa (ADAP-TIV): Study protocol for an adaptive randomized controlled trial}},
  booktitle = {Research Square},
  year = {2023},
  doi = {10.21203/rs.3.rs-2841179/v1},
}
Figure