In my last blog posting I noted that the thousands of health apps available today are beginning to generate good, accurate patient data. But just because the data is accurate doesn’t mean it’s meaningful. Especially when it collides with the real world of the healthcare professional.
There are three important issues that need to be addressed before this surfeit of personalized patient data becomes useful and meaningful to both consumers and physicians.
Data Overload
The first concern is just data overload. As patient devices become interoperable with each other and with EMR systems (a good thing), they will dump raw data, whether it’s heart rate, blood pressure, glucose level, etc. right into the physician’s office. And frankly, doctors just don’t have enough hours in a day to be able to look at and process that information.
If you follow most internists today, they’re in the office all day seeing 15, 20, even 25 patients and then in the evening they’re spending three hours reviewing their notes and lab reports or they’re logged onto their patient portal site to respond to the two dozen patient emails they received that day. We can’t ask them to now review and respond to potentially dozens of patient data streams.
Patient Expectations
Secondly, we may be setting ourselves up for unmet patient expectations. Once a patient feels like, “Oh great, my doctor is getting my data, she’s tracking me now,” the natural conclusion will be that, “My doc is going to let me know if my numbers are out of whack and she will call me up if they’re bad so I can relax a little bit and I don’t have to be quite as focused on a personal level.”
But in fact, there is currently no system that enables doctors to realistically track, prioritize and easily respond to individual patient data feeds. The average primary care doctor in the US cares for 2,300 patients. If even a fraction of this patient load begins to sync their personal activity or symptom data to the physician’s database, then how is the physician going to respond on a timely basis? We are setting up the doctor for failure because of an “expectation disconnect” with their patients.
Non-personal Data
Lastly, much this new patient data is essentially non-actionable because it’s not actually personalized data. It’s personal data, but it’s not personalized. Most mobile health apps are capturing and reporting raw data that is unique to the individual but not necessarily meaningful for the specific individual.
There is a difference between raw data and actionable intelligence. They aren’t the same thing.
A standard blood glucose monitor will report raw data, but it can’t tell whether this particular patient’s glucose level is in or out of their normal range. The data needs to be interpreted. Very few devices are able to establish a personalized baseline against which it can measure variability. For example, few monitors will baseline normal blood pressure when a consumer goes up a flight of stairs or will capture a normal waking heart rate and then highlight any unexplained variations. Data by itself can’t provide intelligence such as a green-yellow-red light alert, or differentiate between readings that are serious or not.
Accurate data doesn’t always reach the level of meaningful data.
So what’s the answer?
An important success criteria of any new health technology is whether it makes patients smarter and whether it truly saves physicians time or helps them be more efficient in the office.
Achieving Meaningful Data
Two ways to achieve this goal are to 1) more meaningfully engage the patient upstream in self-management through personalized feedback, and 2) filter for diagnostic intelligence downstream to the physician’s office.
For patients, this means a more engaging user experience with their health apps with tighter feedback loops so they begin to recognize symptom triggers or the health implications of their diet, sleep and exercise choices. But it needs to be personalized. Counting daily steps is a nice start, but it says nothing about an individual’s fitness. Smart health apps will establish a personalized baseline and track variances.
For the physician, patient health apps need to get smarter and more personalized by learning the general range of “normal” for individual patients and alerting meaningful changes. In some cases, this may mean that the apps will need to become Class II or III FDA medical devices, but eventually apps will need to graduate from toys to clinical tools if they want to stay relevant.
This is an exciting time for mobile health apps, and I applaud the rapid uptake by both consumers and healthcare professionals. Now that we know we can collect accurate personal data, let’s take the next step and make it medically meaningful for patients and doctors alike.