Apple HealthKit, Google Fit
“We saw this as a way to solve some problems related to reliability,” he says. Blood pressure readings made in clinics are often unusually high, for example, because patients are more nervous there. But send a patient home with a cuff and ask him to track his blood pressure daily, and the figures sent back to the clinician sometimes make no sense. Patients may accidentally switch two digits or make an error when they transcribe scribbled written notes. “When we do this wirelessly, we can, in theory, make better decisions based on better values collected at home,” Bloomfield says.
Anthony Wiemelt, Penn Medicine’s chief administrative officer for precision medicine and an expert in personalized medicine, is working to ensure that appropriate genomic information is also integrated effectively into electronic health records. For example, according to the Food and Drug Administration, 2 to 14 percent of people, depending on their ethnicity, carry a gene that prevents them from metabolizing a type of platelet inhibitor known as clopidogrel — these are drugs given to patients dealing with heart disease, including during emergency care. Electronic information about your genetic makeup that is accessible when drugs are prescribed may improve patient outcomes, Wiemelt says.
“If we can identify who carries a genetic variant that prevents them from responding to the standard medication, we can put them on a more appropriate drug,” he says. In oncology, Penn Medicine providers now can evaluate a total of 80 genomic markers to inform decisions about the best possible drug for an oncology patient.
Penn and a few other institutions are going further by using electronic health records to identify groups of patients to invite into clinical trials, which can improve patient care in the long run.
Up and Down the Scales
Penn’s work illustrates what Doug Fridsma, the president and CEO of the American Medical Informatics Association, calls issues of scale or magnitude. “We want the ability to generate technical solutions to healthcare challenges that span from Fitbit to the human genome,” says Fridsma, formerly the chief scientist of the Office of the National Coordinator for Health IT. “This is what we need if we want to achieve a learning health system and use informatics effectively in health care.”
The Right Ingredients
Like a healthy body, an integrated personal health information system needs just the right balance of resources, smarts, and functionality to work. The factors described below are likely to be components of such a system, though details and brands will surely shift over time.
Icons by Ty Wilkins
Electronic Medical Records
EMRs are digital versions of a patient’s paper chart or folder, storing information in far more readily accessible formats for instant access when needed. EMRs can also be shared with other providers, such as specialists, emergency room physicians, or laboratories, far more quickly than paper records can.
Electronic Health Records
Like EMRs, EHRs are digital versions of a patient’s paper chart. But they go beyond the clinical information collected in the doctor’s office and offer a broader view of a patient’s health care. According to the Office of the National Coordinator for Health Information Technology, “EHRs are built to share information with other healthcare providers, such as laboratories and specialists, so they contain information from all the clinicians involved in the patient’s care.”
Electronic Personal Health Records
Unlike EMRs and EHRs, which are owned and operated by doctors, hospitals, and insurance companies, PHRs are created and managed by patients to help them keep up with everything from physicians’ names and contact information, allergies, family history, immunization history, exercise habits, screening results, and more. An electronic personal health record can be shared with a patient’s doctor.
The federal government has pushed healthcare providers to use EHRs in a meaningful way and has even defined levels of meaningful use. Examples include that providers use EHRs to engage patients or their families in their own health care or that EHRs enable coordination of care by multiple providers.
Health IT Vendors
Many companies are building the software that stores our electronic health data and enables quick access and assessment by providers in hospitals, physician offices, and other settings. Some of the most widely used software packages are made by Epic Systems Corporation, Medical Information Technology Inc., Cerner Corporation, and Allscripts Healthcare Solutions Inc.
Wearable devices ranging from bracelets to clothing and watches to jewelry, made by Fitbit, Jawbone, Misfit and more, track steps taken, distances walked or run, or sometimes sleeping and eating patterns and even heartbeats. Users can review their data in apps or online to gain insight into activity (and inactivity) patterns, and even share the results with friends and a broader community.
Fitness and activity tracking apps that measure miles run or calories burned are too numerous to list: MapMyFitness, Lose It!, and Strava are just a few examples. Comprehensive platforms from Apple (Health), Google (Fit), and Samsung (SAMI) can integrate metrics from many Internet-enabled sources — such as WiFi scales, Bluetooth-connected blood pressure cuffs and other health apps — in one place. Both Apple Health and Google Fit play well with other apps and allow you to visualize associations such as whether you eat more or drink less alcohol on days that you exercise or sleep more.
Start with a single patient, such as Fridsma himself. He chose his gym based on its technology, which allows him to integrate all his workout data into an app he uses to track his exercise, diet, and sleep patterns. And he chose a doctor’s office that uses electronic health records meaningfully, to engage with patients over test results and more.
But scale these technologies from a patient or practice to a regional health system, and things sometimes begin to fall apart, he says. Hospitals and primary care and specialists’ offices may all use different electronic systems, for example, which cannot reliably connect with one another.
And ideally, healthcare technologies would integrate at even bigger scales, enabling researchers and healthcare providers to consider populations of millions of people. Kaiser Permanente innovates at this scale, Fridsma says, capturing data on the millions of patients in its system and using smart technologies to identify healthcare trends and issues in a way that can trigger sensible decisions.
“Then take it up to 1 billion people, 9 billion, and you’re talking about the human genome and a basic understanding of the human condition,” Fridsma says. “At this scale, with clinical research, we can begin to understand the basic science to make inferences that are generalizable to the public,” he says.
For example, research at this broad scale may help doctors identify new gene variants that increase risk for Alzheimer’s disease. An integrated electronic health records system could help clinicians then quickly identify people with that gene variant, track their disease progression, and test interventions that may slow it, a painstakingly sluggish and costly process in the current health system. Scaling back down to the individual again, a provider could engage electronically with a patient who was dealing with early signs of Alzheimer’s, tracking critical symptoms.
“You could ask, ‘Is my patient getting enough exercise, enough cognitive stimulation?’” Fridsma says. “This is the promise of the learning health system and integration across these different scales.”
Challenge to Integration
If the benefits to individual patients and public health seem obvious, why don’t we have integrated systems yet?
“There are lots of reasons,” Fridsma says. “Some are social — I like the saying that ‘information moves at the speed of trust.’ So if you trust that your provider will be a good steward of that information or that by sharing your information with a researcher you’ll be contributing to something larger than yourself, well, you’ll be more likely to share than if you worry they will monetize and sell
This trust goes hand in hand with issues of security. Although some feel that certain data-gathering devices, such as fitness trackers, generate information that hackers would consider relatively mundane by itself, when that information is integrated into a larger system, such as a patient portal, the additional context might prove more tempting to hackers and thus pose more of a threat to a consumer. His or her information might be used for the wrong reasons.
Money matters, too, says Denver’s Schwartz. Government organizations such as the National Institutes of Health (where he formerly worked) do a tremendous job of funding innovations in understanding disease and health, he says, but to develop technological systems that support and advance those innovations, Schwartz imagines a more diverse system, which includes private investment. “We have good computational resources, but we still have very crude devices, and we just don’t yet see the venture capital investment necessary for the development of more sophisticated devices.”
And there are mundane challenges, as well, from the technological to the legal. At Penn Medicine, for example, researchers and healthcare providers need to understand the timing of readings, yet many personal health trackers and even electronic health records lack a database field for time of measurement versus time of receipt by the EMR. A blood glucose level sent in from a patient might be fine if it was taken soon after breakfast but worrisome if taken just before lunch. And for certain conditions — women at risk of postpartum high blood pressure, for example — healthcare providers might need measurements taken at the same time every morning and evening.
“It can be incredibly important to have that sense of time for key events, and it’s not a big deal to build it in, but it hasn’t been done yet,” in some systems, Wells says. “There’s no place to record time of measurement or identify the proximity to other events (like meals or sleep start and end). Apple’s recent ResearchKit announcement is starting to bring survey data and biometric data together, but it is targeted at clinical research and not clinical care.”
He and his colleagues bring up legal concerns, too. What if a patient struggles to get a Bluetooth connection between a home blood pressure cuff and his phone, or tests the system on his wife and accidentally uploads her metrics to his own electronic records?
Worse, what if an electronic health system picks up an abnormal glucose reading from a diabetic patient who engages with her providers electronically, but then a live person — a doctor or nurse or other care provider — fails to act on the reading because of a scheduling glitch, a lost cell phone, or another emergency. “What if that patient went into shock or had a bad outcome?” Wells asks. “This is all so new, the industry is still trying to figure it out.”
In working toward a future of smart, integrated health systems that benefit the patient as well as larger populations of people, Fridsma says leaders need to think strategically and shouldn’t focus on creating a single, perfect, integrated system. “This will have to be iterative, incremental,” Fridsma says, “and we can’t get there if we create systems that can’t evolve.”
Schwartz, one of those leaders, has little doubt about the future. He’s focused on integrated systems in his own research, where he sees tremendous potential for technology to help identify people at genomic or environmental risk for disease, and to better intervene. And he envisions a time when technology will allow his health data to be transmitted to his doctor and then information from the doctor will be sent back to him.
“I’ll be able to regulate the dose and frequency of any medicine or my level of activity, diet, and exposures, in a way that maximizes my healthcare status and functional activity,” says Schwartz. “I imagine a future where I won’t actually have to go in to see the doctor.”