Reading ONC’s recent summary of patient engagement capabilities at US hospitals left me feeling both hopeful and wistful. The ONC, as usual, is trying to show off how much progress the field of health IT has made since Meaningful Use started, and the statistics in this dashboard meet those goals. On the other hand, I look at the statistics and wonder when real patient empowerment will emerge from these isolated gains.
The ONC dashboard includes information both on raw data exchange–what Meaningful Use called view, download, and transmit (VDT)–and the uses of that data, which ultimately mean much more than exchange.
I considered at first how important I would find it to download hospital information. I certainly would like my doctors to get the results of tests performed there, and other information related to my status upon discharge, but these supposedly are sent to the primary care physician in a Continuity of Care Document (CCD). If I or a close relative of mine had a difficult or chronic condition, I would certainly benefit from VDT because I would have to be an active advocate and would need the documentation. My point here is that our real goal in health reform is coordinated care, rather than data transfer, and while VDT is an important first step, we must always ask who is using that information.
The ONC did not ask the hospitals how much of their data patients can download. God is in the details, and I am not confident that an affirmative answer to the question of downloading data means patients can get everything that is in their records. For instance, my primary care physician has a patient portal running on eClinicalWorks (not his choice, but the choice of the hospital to which he is affiliated). From this portal I can get only a few pieces of information, such as medications (which I happen to know already, since I am taking them) and lab results. Furthermore, I downloaded the CCD and ran it through a checker provided online by the ONC for a lark, and found that it earned D grades for accurate format. This dismal rating suggests that I couldn’t successfully upload the CCD to another doctor’s EHR.
Still, I don’t want to dismiss the successes in the report. VDT is officially enabled in 7 out of 10 hospitals, a 7-fold growth between 2013 and 2015. Although the dashboard laments that “Critical Access, medium, and small hospitals lag,” the lag is not all that bad. And the dashboard also shows advances in the crucial uses of that data, such as submitting amendments to the data
A critical question in evaluating patient engagement is how the Congress and ONC define it. A summary of the new MACRA law lists several aspects of patient engagement measured under the new system:
Viewing, downloading, and transmitting, as defined before. As with the later Meaningful Use requirements, MACRO requires EHRs to offer an API, so that downloading can be done automatically.
Secure messaging. Many advances in treating chronic conditions depend on regular communications with patients, and messaging is currently the simplest means toward that goal. Some examples of these advances can be found in my article about a health app challenge. Conventional text messaging is all in plain text, and health care messaging must be secure to meet HIPAA requirements.
Educational materials. I discount the impact of static educational materials offered to patients with chronic conditions, whether in the form print brochures or online. But educational materials are part of a coordinated care plan.
Incorporating patient-generated data. The MACRA requirements “ask providers to incorporate data contributed by the patient from at least one unique patient.” Lucky little bugger. How will he or she leverage this unprecedented advantage?
That last question is really the nub of the patient engagement issue. In Meaningful Use and MACRA, regulators often require a single instance of some important capability, because they know that once the health care provider has gone through the trouble of setting up that capability, extending it to all patients is less difficult. And it’s heartening to see that 37 percent of hospitals allowed patients to submit patient-generated data in 2015.
Before you accept data from a patient, you need extra infrastructure to make the data useful. For instance:
You can check for warning signals that call for intervention, such as an elevated glucose level. This capability suggests a background program running through all the data that comes in and flagging such warning signals.
You can evaluate device data to see progress or backsliding in the patient’s treatment program. This requires analytics that understand the meaning of the data (and that can handle noise) so as to produce useful reports.
You can create a population health program that incorporates the patient-generated data into activities such as monitoring epidemics. This is also a big analytical capability.
Yes, I’m happy we’ve made progress in using data for patient engagement. A lot of other infrastructure also needs to be created so we can benefit from the big investment these advances required.