Patient Generated Data — Moving Beyond “Ignorance is Bliss”
What crisis? The cartoon, referenced in our lab meeting, referred to regulators not being aware of what was going on around them (the Greek debt crisis) — because their heads were buried in the sand. Our fellow Dr. Mujeeb Basit was making the point, during his presentation evaluating decision support rule failures, that just because we don’t look doesn’t mean there’s nothing to see.
This prompted a recurring conversation in our lab: when is it better — and when is it unethical — to “not know.” Our recent conversations have focused on patient-generated health data, including medical history, device data, and logs. It’s becoming more practical for patients to track their own information, and transmit data directly to their care team when appropriate.
Patients can prepare for visits by providing history in advance of the face to face time with the clinician. This may help patients organize their thoughts and allow docs to jump into clarifying questions.
But as opposed to trying the technology out, imperfect as it may be, many clinicians seem like they would rather not have any access to the data.
We Need To Figure This Out
We now have the ability, facilitated by technology, to track how people are doing in their own homes, including device data, medical history, and logs. The data are likely to be more valuable and meaningful, if analyzed properly.
Technology advances are effectively disrupting how health data are being collected.
Connections through patient portals and mobile applications are increasingly going to make it easy and relatively frictionless to share data with clinicians (the visualization and analysis tools for clinicians is sorely lacking though . . . ). For example, blood pressure and weight measurements can be easily collected using electronic devices, connect by Bluetooth to mobile phones, and forwarded to clinicians.
Many of our well meaning colleagues would rather “not know” than be “responsible” for data that could carry risk. And it’s hard to blame them. They need better tools. But that may not be the only reason for their trepidation.
Many well meaning colleagues would rather not know as to not ‘be responsible’ for the data.
Warner Slack, co-founder of our lab, reflects that in the early days of the chemistry auto-analyzer, there was an important question. Chemistry auto-analyzers often run several tests on a sample of blood automatically for the most routine tests. If a doctor orders some but not all of a common panel of tests, such as a sodium and a chloride, but not a third such as potassium, and all three results are theoretically known, should all three be reported, or only the two that were specifically ordered? The logic was that if a potassium, though not specifically ordered, were critically low or high, it would be unethical to censor that result. The doctor should see it.
In another example, Dr Slack included questions about depression in a computer-based interview available for employees of our hospital. Should the computer ask about suicide? Dr. Slack felt that it would be important to ask the question in the context of a visit, so equally important for the computer to ask in this situation. However, what if the respondent said “yes?” What would or should happen? Do we have an obligation to prospectively reach out to that person now? This was solved by putting some information up about resources, such as numbers for the Samaritans.
A few other vectors . . .
Already Information Overloaded
As clinicians, we constantly juggle too many information streams. Some old fashioned — the pager for “need to reach now” type of messages and faxes from other clinicians or home care agencies. Newer technologies and electronic health records bring more data: queues of results and reports, forwarded notes, and more messages. E-mail is the defacto channel for most team communication. With more frequency, we also have secure messages from patients (while some of our work shows that the volume is not overwhelming in and of itself, it adds up). And of course we have the pesky telephone.
Liability For Missing Information
Besides being stressed about the sheer volume of data that already clutters their physical and virtual desktops, clinicians have concern about their professional liability related to all of these messages and data. Failure to follow-up data and communications can lead to harm to a patient, and is a leading area of malpractice risk.
So, what if an abnormal signal is in the data that we don’t see in time?
We can become trapped in a mindset where we are more concerned about managing professional risk/liability rather than figuring out what is most helpful to our patients.
Uncertainty In The Data
How reliable, and how valid, are the data being generated? Will the data help me as a clinician to make a diagnosis, or will it confuse the issue at hand?
Can information be harmful? Yes, sometimes. This is often unrecognized, but data considered “positive” often lead to cascading effects, including procedures and testing, or treatments.
Many clinicians are skeptical of being responsible for data where they are not compensated for its review. Changes in healthcare financing, accountable care organizations, and global payments may help with this.
We have important reasons to move forward, and figure this out.
Just because you don’t know that a patient’s blood pressure is 180/110 doesn’t mean that it’s not.
Collaborative care with patients will increasingly rely on these connections and data sharing.
Informatics groups, like ours, need to continue to work on making data actionable, and alert clinicians when they need to be alerted. Care teams, with shared but delegated responsibility for patients, are necessary in primary care clinic environments.
We have been developing a policy at BIDMC for the incorporation of patient-generated health data into our electronic health record and routine clinical practice. It starts with a conversation between patients and clinicians about the purposes, use, and review.
We’re planning on refining our thoughts on “when is it better not to know” and welcome comments and conversation.