In my blogs over the last few weeks, I took a high level look at what is included under the hood in newly introduced Apple ResearchKit. I also touched upon functional details on the three out of box modules provided in the ResearchKit software framework – Surveys, Informed Consent and Active Tasks. You can get into more implementation level details for these modules here.
Today in the final blog on this topic, I am going to touch upon issues and concerns that have been raised about ResearchKit in the last three months. Many medical researchers as well as some eminent healthcare ethicist have raised these concerns. I want to acknowledge my reference to these as a source for this blog.
This has been raised as a significant concern by many commentators since ResearchKit is currently only supported on iPhone’s.It is argued that the population of iPhone owners is not representative enough of the general smart phone user population. There is certainly some truth to the notion that participants in any study that includes only iPhone users will predominately be educated high earners with very narrow racial and gender diversity.
One easy way around this concern is to not release research app based on ResearchKit to general population (on the app store) but to release it to a controlled group via the Apple Enterprise account.Naturally this takes away from the potential of having a large sample size if anyone could download the app and participate in the study but it still provides an opportunity for the researcher to leverage the phone as a data collection mechanism. One thing to note is that Apple so far hasn’t commented on the possibility of expanding the ResearchKit to Android and Windows phones – which will go a long way in getting over this issue.
High Attrition Potential
This is the flip side of the high participation potential because anyone can chose to participate in the study as long as they have an iPhone (and hopefully honestly self qualify). But as many in medical research have experienced, keeping the participants interested and committed to a study is always a challenge. This is likely to be even more pronounced when the participation was self-motivated. Signees can lose that motivation very quickly and the study may end up with a high attrition rate this severely impacting the quality and efficacy of the research. Researchers need to keep the participants engaged and motivated throughout the study period and also should plan for a higher level of attrition.
Though ResearchKit does provide means to qualify the participants, it is self-qualification if the app is publicly available on the app store and thus anyone can download it to participate. What stops anyone from agreeing to be a healthy 25-year-old female when really being a 49-year-old overweight male – just to participate in the study? This may be a somewhat overblown concern but it does feed into another a concern about reliability of data collected through ResearchKit apps.
Reliability of data
Limited Generalizability, selection bias (which can be introduced by multiple factors based on the study subject), possible high attrition and participation bias all lead to the concerns about accuracy and reliability of the data collected and thus to the inferences drawn from the analysis of that data.These apps could potentially generate real large dataset – question that many have is whether that will help or hurt the quality of research?
Ethical and privacy concerns
This has probably been the most talked about concern since ResearchKit was introduced. Do the participants who have not actually talked to a researcher when they download an app and sign in to participate in a research study fully understand the study and it’s objectives, the kind of data that is being collected during the study, who besides the primary researcher may potentially have access to this data? Though the ResearchKit framework provides for many ways to communicate this information to the potential participant, it’s not always very intuitive. And then there is the concern about privacy for the participants even though Apple has said that it will have a third party company strip out any identifiable PHI data before passing on the collected data to the researchers.
Of course many of the same concerns and issues exist even today even with the high level of process complexity and bureaucracy built to avoid them. Even with the limitations of ResearchKit, it has the potential to significantly enhance the quality of medical research and also make it more cost effective. The key is really to understand those limitations and build safeguards that will allow to fully exploit the functionality that it provides.
Is your company building apps using Apple’s ResearchKit? If yes, I would love to hear about your experience.
Here is your chance to learn how to leverage the capabilities of Researckit.Reserve your spot for this informative webinar now!