Takeaways from the Africa mHealth Research Institute (AmHRI) workshop

[beatrice]

I had an awesome week at the Africa mHealth Research Institute (AmHRI) workshop held last week in Nairobi. We had an amazing faculty and a group of scholars. It was great spending the week learning and sharing with peers in mhealth research. Below are highlights from the week-long workshop.

Faculty:

Fogarty-National Institutes of Health (Laura Povlich, PhD)

Moi University School of Medicine (Profs. Violet Naanyu; Ann Mwangi; Eunice Kamaara; Juddy Wachira)

Vanderbilt University, School of Medicine (Martin Were,MD MS; Holly Cassell, MPH)

Yale School of Medicine (Saria Hassan, PhD)

Purdue University (Young Kim, PhD; Munirul Haque, MD, PhD)

University of Georgia (Juliet Sekandi, PhD)

Scholars: We were 31 scholars from not-for –profit organizations, academic institutions and technology companies. These included Medic Mobile, Partners in Health, Rwanda, KEMRI, Regional Centre of Excellence for Biomedical Engineering and e-Health, Rwanda, Makerere University, Mbarara University, Intellisoft, Mount Kenya University, University of Bergen, University of Botswana, Africa Health Research Institute, South Africa and Moi University

1. mHealth overview
During this session Martin Were and Young Kim walked us through mHealth innovation and trends between 1992 to present. Types of mHealth Applications in developing country settings as outlined by the mHealth Alliance Evidence Working Group (EWG)

The seven considerations for mHealth interventions which I find very useful to have at the back of our minds as we deploy digital health solutions for CHWs. These include (i) security & privacy , (ii) scalability versus pilotitis, (iii) conformity to health informatics standards, (iv) interoperability and information exchange (v) cost, payer vs. beneficiary (vi) ethics & equity considerations (insert article from Prof. Kamara) and (vii) evidence of impact of mHealth solution. Additionally the use of the BEACON tool to assess a site’s readiness to implement a digital health intervention. https://drive.google.com/drive/u/0/folders/1mpXtbQpjx7HkPsXdgM182iXOrYNSqUeZ

2. Research proposal writing

This session was valuable in providing guidance on writing successful medical research, grant writing tips, publishing the research findings and the NIH grant application process and components. This was very useful as we were working on our on the team science assignment. Please find a link to articles on successful grant writing and acing research aims @chris and @doreen you may find this helpful for other funding opportunities https://drive.google.com/drive/u/0/folders/1kA2BQci9_QKbCTuRX8yccPzld10A9ZQk

I would like to share this quote from Norm Braveman, NIH on grant writing “There is no grantsmanship that will turn a bad idea into a good one, but there are many ways to disguise a good idea.” Additionally Holly Cassell, MPH reminded us that “Good” proposals seldom get funded, but “Excellent” ones have a good chance!

3. Adopting implementation research approaches to evaluate mHealth interventions

I was pleased to see implementation science included in this mHealth workshop @isaacholeman We have often discussed the need to consider implementation research approaches for evaluation of mHealth interventions to classic gold-standard research approaches given the iterative nature of our tools. The focus of this session was on the steps to incorporating implementation science principles as well as theories, frameworks and models we could adopt when evaluating these interventions

Nielsen implementation science framework highlighting the five categories of theories, models and frameworks used in implementation science (https://drive.google.com/drive/u/0/folders/1FHUHHReXJXxCKcUKr2qKFWmd6cDnVHdw)

Susan michie et al. behavior change wheel –a framework used for designing interventions https://drive.google.com/drive/u/0/folders/1FHUHHReXJXxCKcUKr2qKFWmd6cDnVHdw

4. Sensors and machine learning

This was one of the most intuitive sessions for both scholars and faculty. Dr. Young Kim discussed the evolution of smartphones and sensors and how we can leverage on sensors for mHealth. Most smartphone devices have built-in sensors that measure motion, orientation, and various environmental conditions. The validity of using smartphone sensors to measure clinical outcomes .This presentation highlighted 15 smartphone sensors and examples of how these have been used for mHealth i)activity monitoring mHealth ii) Cardiovascular health; iii)Respiratory mHealth (Spirometry - Mayo Clinic); iv) Ophthalmology (Smartphone-based retinal imaging system) Undilated versus dilated monoscopic smartphone-based fundus photography for optic nerve head evaluation | Scientific Reports; v) Hematology (Virtual Hyperspectral Imaging of Eyelids - <i>mHematology</i> for Blood Hemoglobin Analysis by Michelle A. Visbal-Onufrak, Md Munirul Haque, Martin C. Were, Violet Naanyu, Sang Mok Park, Md. Kamrul Hasan, Mhawila A. Mhawila, Kit Yee Yeung, Young L. Kim :: SSRN) @isaacholeman @marc @derick

IBM Research Africa shared their work on machine learning and specifically the IBM Watson tool and its use in the field of oncology. Indeed they are doing some great work on AI, but would have appreciated hearing how they have deployed this within a real-world settings. I had hoped to learn their experiences deploying the AI tool in LMIC, challenges and the algorithm validation process.

5.Gamification to improve CHW performance
As part of our team science assignment we were to evaluate an innovation leveraging on gamification to change the behavior and improve CHWs performance for HIV testing and counseling in Kenya. Gamification is the use of game elements and game-design in non-game contexts” (Werbach & Hunter, 2012) to engage and motivate people to achieve their goals by providing a whole different user experience. It aims at stimulating people’s motivation in doing an activity by trying to make it rewarding for itself (https://drive.google.com/drive/u/0/folders/1mpXtbQpjx7HkPsXdgM182iXOrYNSqUeZ) Gamification was quite a new term to me prior to this workshop, however, I realized this was something I would engage with day-to-day for example when using Duolingo language app leaderboard

6. Panel discussions with funders
On the final day of the workshop we had a panel discussion with funders (Ministry of Health, Kenya –Research and Innovation (Dr.Joyce Wamishwe rep. Dr. Nzioka), NIH (Laura Povlich), USAID-Washington, DC (Michelle Jones), Vanderbilt (Holly Cassell) on funding opportunities for mHealth research. They shared insights of what funders look for in proposal and shared a list of resources on funding opportunities https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6096297/pdf/main.pdf

NIH is shifting focus to research around new and emerging technologies such

1. Combining data science with data-driven technologies (machine and deep learning)
2. Leveraging on internet of things

USAID does not provide funding primarily for research but rather for applied research to improve development. Some of the opportunities include THE PEER program, LASER program as well as the women connect challenge. USAID will soon release it’s digital strategy and encourage those in the digital health space to provide feedback.

Next steps
We will continue to interact with the faculty through the workshop online platform and they have expressed a willingness to work with us on projects, publications and grant writing process @isaacholeman @joshnesbit . They will continually share grant calls and we could leverage on scholars’ expertise and submit joint applications. I would really recommend this course for those interested. We were the first cohort for this workshop and they intend to engage this year’s scholars in subsequent workshops.

We had a couple of scholars who are interested in our work and open-source tools within the CHT and may have already requested for a demo @francesca

Thanks,
Beatrice

[isaacholeman]

Thanks for this update @beatrice! I’m so happy to hear that it was such an interesting experience, and that there seem to be good opportunities for follow up. This Norm Braverman quote:

“ There is no grantsmanship that will turn a bad idea into a good one, but there are many ways to disguise a good idea .”

Soooo resonates with me. I also have a question about implementation science methods. One of the arguments we make in our Human-Centered Design for Global Health Equity paper is that HCD and related research methods like Action Design Research and Research Through Design are important methods for implementation research. Here’s a relevant section:

much of the interest in the poor scalability or reproducibility of digital health interventions in lower-income settings has come from the medical literature (Shuchman, 2014; Tomlinson et al., 2013) and is linked to the growing field of global health implementation research or implementation science (Kim, Farmer & Porter, 2013; Kruk et al. 2016). This research community has made important steps towards documenting the “real world” contextual factors that are overlooked or poorly captured by the randomized trials of classic biomedical research. However, this literature often treats design and implementation, and thus design research and implementation research, as separate phases. As a result, some guides to implementation research suggest focusing on “implementation strategies” for spreading existing interventions (Peters, Tran & Adam, 2013), rather than fundamental, iterative redesign of those interventions.

This artificial separation is significant from a design perspective. Growing awareness of the importance of iterative methods that integrate implementation and ongoing redesign stems directly from attention to the complexities and wicked problems that characterize the broader human context of implementation (Buchanan, 1992; Rittel & Webber, 1973; Schön, 1983). It is also a pragmatic roadblock in places where patients with interrelated health issues (e.g., HIV and TB) must navigate a range of confusingly disjointed programs. Calls for more integrated care can only be meaningfully addressed if interventions, which initially were tested in pilots that focused on a single disease, can be fundamentally redesigned in and through the course of implementation.

The tendency to separate design and implementation is understandable, given that design research is seldom discussed as one of the disciplines that informs implementation research. It is our hope that this paper will serve as a starting point for researchers who are interested in exploring human-centered design, not only as a practical approach to addressing implementation challenges, but as a body of concepts and research methods with much to offer global health implementation research.

My question for you is, what do you think the researchers at this workshop would think of these methods? Did you learn anything about what the funders might think? My sense is that implementation researchers are often open minded about drawing on methods from other disciplines, including design research, but the major funding streams still strongly favor RCTs. What do you think?

[beatrice]

Thanks for this question @isaacholeman. It was a recurring question from all of us scholars during the sessions on research designs for evaluating mHealth interventions and discussion with funders. Generally, funders acknowledge the rigour of using RCTs from the ability to clearly link cause and effect but are increasingly embracing implementation science methods to help identify what works, what doesn’t and why and then use these learnings to improve on interventions or programs . Additionally, Laura from NIH mentioned that if the grantee had a two year grant it would be more appropriate to consider implementation science approaches since there may not be adequate time as well as the budget implications may not allow for a RCT methodology.