Yesterday, Apple held one of its Keynote events. Amongst talk of super-thin minimalist laptops and $10,000 luxury smart watches, we were also very excited by the announcement of ResearchKit. This interests us both as Science Practice and as part of SP+EE where we work on healthcare-related projects, including design and development of software for medical research.
We spent some time today delving a bit deeper into ResearchKit, so here are a some initial notes along with a few open questions that we hope will be answered soon:
After six hours we have 7406 people enrolled in our Parkinson's study. Largest one ever before was 1700 people. #ResearchKit— John Wilbanks (@wilbanks) March 10, 2015
Importantly, the code will be open source meaning that, when it is released, anyone will be able to download and start building on top of the platform. Hopefully this will encourage a community of developers and researchers to form and share useful modules. These might not be limited to software alone, but also reusable design patterns, ethically approved content and best-practices.
ResearchKit has a technical overview document which can be found here. This document contains information about the main modules ResearchKit will have. These are designed to reflect the most common elements within a clinical study: a survey module, an informed consent module and an ‘active tasks’ module.
Interestingly, the technical overview documentation also has a list of things that ResearchKit doesn’t include, such as the ability to schedule tasks for participants. We’ve found scheduling tasks and reminders to be a more-or-less essential feature in similar types of study we’ve been involved in, so this is probably on Apple’s todo list.
Apple also states that it doesn’t include automatic compliance with research regulations and HIPAA guidelines. This places the responsibility firmly on the researchers. However, as the developer and research community around ResearchKit establishes itself, perhaps best-practices will be developed and shared to help streamline compliance in a range of international regulatory contexts.
In an interview in Nature, Stephen Friend, president of Sage Bionetworks states that “At any time, participants can also choose to stop. The data that they have contributed stays in, because you don’t know who they are”. This standpoint doesn’t quite chime with us. Perhaps other consent models will be possible where a participant can optionally remove their data from a study if they decide to leave. It’s their data after all.
In the same article, Ray Dorsey from University of Rochester in New York comments on the the most obvious problem with ResearchKit: sampling bias. He points out that “the study is only open to individuals who have an iPhone, specifically the more current editions of the iPhone”. We can imagine a port of ResearchKit to Android popping up fairly quickly once the source code is released, but bias will always be an important consideration in studies that require participants to have smartphones, regardless of the manufacturer.
It will also be interesting to see how the peer reviewers will interpret the quality of the data collected by the ResearchKit enabled apps when the results begin to be published.
It looks like a fascinating platform and we will be watching eagerly as it develops and maybe even jumping on-board too.