By Eric Reifsnider, Manager, Data Science, Validic™
As a software engineer, I can get thoroughly caught up in the inner workings of the systems I am building. It is easy for me to over-generalize what the folks who are using my system are actually using it for, out in the world. If you asked me as I was merrily coding away, when I looked up from my code, I might say, “I’m helping to improve people’s health by making patient-generated health data available and actionable for everyone.”
But that “everyone” is complicated in product roadmap meetings, where I join in as a manager. The most clear dichotomy from my perspective as an engineer is between the world of wellness, where patient-generated health data (PGHD) adoption and integration is very advanced, and clinical care, where it is not so much. There are some technical differences, to be sure – integration with the systems that back up wellness programs is very different from integration with the systems that back up electronic health records (EHR) that are used in clinical care. But from a technical standpoint, it is really not that different – certainly not different enough to account for the difference in rates of adoption.
It seems straightforward to me how PGHD would be used in a clinical setting. I recently went in to my primary care physician for a complaint – a minor problem with a vein in my calf. I filled out the usual, paper intake forms. I then had to explain to my doctor that since our last checkup I have continued to exercise every day (almost) for 30 minutes (usually). But, this verbal explanation was very short and not entirely objective.
My doctor referred me to a specialist, where I again filled out these paper intake forms – and again briefly explained my exercise history. I was then referred to another specialist – more paper forms – and another brief word about exercise. Daily exercise is a big part of how I try to maintain my own personal health, so it felt odd to be just check the box next to “I exercise for 30 minutes or more at least three times per week” on a paper intake form, and then add a few sentences during an interview. To me that did not seem anything like a real assessment of my exercise – and consequently of my health.
Like many people today, I use a wearable fitness tracker (in my case, a Fitbit). And, the apps for wearables like mine typically provide historical graphs showing my activity data, dating back months at a time. As I was waiting for a specialist, sitting on the crinkly paper on the exam bench, I imagined a world in which my doctor could call up that historical information from my wearable in his EHR, as a part of my chart. Even the “at-a-glance” charts that the app for my wearable provides, showing how active I have been in the last few months, would be a big improvement.
In the world of wellness, a person’s history of activity is a critical measure of healthy behavior. My wearable fitness tracker data can be used directly by a wide array of different wellness systems, which require only a few clicks to link up my PGHD with backend wellness systems. And it is commonplace for wellness systems to have the PGHD automatically compared to individual goals that promote an active lifestyle, with rewards automatically given to participants who achieve their goals.
I am not looking for an automated primary care situation where I see less of my doctor. Far from it. I value my doctor’s time with me, and I can see that my doctor is working hard for me and for other patients – always busy and always on the move. And I understand that it’s my doctor’s duty to do no harm, taking action only when they are confident of a positive outcome, including confidence in the information leading to a decision to act.
But, again, as a technologist, I am very aware that even during the actual face-to-face conversations with my doctor, heart rate data is flowing from my Fitbit into the cloud, and that data is stopping there, never reaching my doctor, who just recorded my heart rate in his office. Not the same quality of measurement, to be sure – the cuff wins – but there is so much more heart rate data available from my Fitbit. And so much activity and exercise data. And more data types being added every day, including EKG and blood pressure in the near future. And after all, the doctor is asking me about exercise, and putting it on my intake forms, for a reason. Surely, I think, the data from my wearable must be useful to my doctor.
So perhaps someday, when I make an appointment to see my doctor on my phone, I will be able to check a checkbox to authorize pulling in my Fitbit data, as a natural part of a digital intake form. And when I show up for an appointment, there will be fewer paper intake forms to fill out. And clerical staff at my doctor’s office will spend less time typing in my intake form data (at an error rate that might be near 3% ) and more time assisting my doctor in other ways. And if I’m really feeling bad, I won’t have to worry quite as much about communicating with the doctor, with my PGHD helping to explain my health issues for me.
And finally, as an engineer, I am optimistic that analytics and machine learning can support our doctors and help them to focus on the most important things in their practice. I would look forward to that future as a patient – and I am helping to build it as a technologist.
 Hong, et al., “Error rates in a clinical data repository: lessons from the transition to electronic data transfer—a descriptive study”, BMJ Open. 2013; 3(5).