20 Mar ResearchKit Mobile Health Apps: In Conversation with Dan Webster and the Mole Mapper Journey
The introduction of ResearchKit (Apple Inc.), the open-source platform in clinical research, piqued interest in the medical community even amidst apprehensions about skewed data and anticipation of performance in the ‘real world’. It has been two years since, and some of the results trickling in from studies using researchkit are demonstrating the feasibility of conducting studies entirely on smartphones. With android libraries of ResearchDroid and ResearchStack being available, mobile health app studies can now reach out to a larger population.
Mole Mapper is a melanoma app which recently published its preliminary results. We connected up with its developer Dan Webster via an email interview to understand the mobile health app, launch of its android version and the initiative better.
In 2015, Mole Mapper, was released by Sage Bionetworks and Oregon Health & Science University. It is a personalized tool to help you map, measure and monitor the moles on your skin. Using a familiar Maps-like interface, you can measure the size of a mole using the camera and a common reference object like a coin.
It’s Android version uses ResearchStack, the android cousin of ResearchKit, developed by a team from Cornell Tech, the nonprofit group Open mHealth, and Android development firm Touchlab.
Mole Mapper has an interesting story behind its development. Outlined below is the interview with Dan and the Mole Mapper journey.
Q: Mole Mapper app was initially a mole tracking app and then it was relaunched as a ResearchKit study. Can you outline the journey for our readers.
Dan: The original Mole Mapper app was a tool that I built for my wife because she is at a high risk for developing melanoma. We wanted to quantitatively measure and map her moles to see if they were changing over time. It was a pretty simple app that kept all of the data locally, and once we had it working, I released it as a free app in case it could help people like us. It was downloaded and used by a few hundred people, which I was happy about and I felt like I achieved all I could with Mole Mapper…but then Apple released ResearchKit.
When Apple announced ResearchKit at their Spring 2015 event, I wrote to the email address they posted on the ResearchKit website on the first day telling them that I was a biomedical researcher and also had a health-tracking app. To my surprise, they got back to me right away. They encouraged me to integrate the ResearchKit framework into the existing functionality of Mole Mapper and let users contribute their data to a larger IRB-approved, institutionally-supported research study. This was when I teamed up with Sage Bionetworks and OHSU to create the Mole Mapper app that is in the App Store today.
We worked hard to finish it in time for the “second wave” of ResearchKit apps after the initial 5 that were released. We then had the privilege of presenting Mole Mapper to journalists together with 2 other ResearchKit apps, EpiWatch and Autism & Beyond. The Mole Mapper study is ongoing, and we continue to enroll participants to this day.
Q: What was the main aim of the Mole Mapper researchkit app study? (Were you expecting people to consult a dermatologist based on self collected data since the app clearly indicates that it is not a medical tool.)
Dan: What we did NOT want it to be was another app that promoted itself as having the ability to detect cancer through your phone. A number of these apps have been fined by the FTC for making false claims. Instead, we took the approach that any claim we make or interpretation we give will be backed by thorough research on a large, relevant dataset. So our first step was to accumulate a large dataset of “normal” moles, which helps us to establish what is “abnormal”. We recently published our preliminary analysis of the first 7 months of data collection, and made this data open to qualified researchers through the Mole Mapper Public Researcher Portal on Synapse.
In the long term, we’d like the app to help you make a decision about whether or not to go in to the dermatologist. Imagine someone who works in a rural area on a farm who may be highly sun-exposed. This person might be hours from the nearest dermatologist, so the decision to seek care is one that might cost them a whole day’s drive, notwithstanding potential out-of-pocket medical expenses. Mole Mapper may one day help you make that decision and then serve as a detailed record for you and your dermatologist to review when you get to the clinic. If the soonest you can get a dermatologist appointment is 3 months from now, you can measure your moles once per month until your visit and then you can review your records as if it were 3-visits-in-1.
Q: 25% of people consented to enroll and contribute data in the Mole Mapper study. What were some of the hassles you faced while enrollment ? (Drop off reasons etc.)
Dan: One of the first principles we had was that we wanted the app to be valuable to the user, whether or not they enrolled in the study. As a result, they can opt out of the study enrollment process at any time and go straight to the mole mapping functionality without contributing data for research. I think that contributes to our enrollment rate, but it is an important principle for us.
With that being said, users can always enroll at any time, even if they skipped the initial enrollment process. We have refactored our consent process for our upcoming app release to have fewer screens, but to still have a short quiz to make sure users fully understand the study.
Q: With the Android version active now for the app, have you observed a marked increase in downloads and enrollment?
Dan: The majority of users are still coming from iOS side, but we hope to see an increase in Android users over time. This will be particularly important as we expand the Mole Mapper Study outside of the US.
Q: What were some of the challenges you faced in making the Android version of Mole Mapper?
Dan: The Android version of Mole Mapper co-evolved at the same time as ResearchStack, with these efforts both being initiated and driven by Cornell Tech and Open mHealth. So it was a challenge to have the 2 projects getting to the finish line neck-and-neck. Developing for both iOS and Android will always be a major effort, but with ResearchStack in place, I think that will be much easier.
Q: What new features would you like to see in ResearchKit?
Dan: One feature I’d love to see is a photo-based medication recognition engine, like the nutrition facts barcode recognition on an app like MyFitnessPal. As ResearchKit and CareKit come together, having a quick and easy way to enter information about medication would be a huge benefit. The good thing about ResearchKit is that it is open source! So if I want to see something incorporated, it is incumbent on me to make it happen and then submit a Pull Request. I hope all of your readers will see this same opportunity.
Some salient points from results published in Nature:
- The study cohort included 2,069 participants who contributed 1,920 demographic surveys, 3,274 mole measurements, and 2,422 curated mole images.
- Red hair was found to the most significant factor in the survey data analysing association of melanoma and demographic risks.
- Participant-provided mole measurements indicate an average mole size of 3.95 mm.
- From October 15, 2015 to June 1, 2016, the Mole Mapper iOS app was downloaded by 11,056 users, of which 2,798 (25.3%) chose to enroll and contribute their data to the study.
The Mole Mapper Study takes a participant-centered approach, providing participants with control over data generation and data sharing. Future directions of the mobile health app study include implementation of automated mole detection, integration of genetic testing results, inclusion of the Android operating system, and extension outside of the United States.
The future impact of mobile health studies is going to be huge going by the trend of results of studies that have started coming in. Reaching and recruiting huge number of participants via a simple smartphone is a revolution in the making where drug development and translational research is concerned.