I saw this tweet from the famous @HealthcareWen which asks a really good question:
— Wen Dombrowski MD (@HealthcareWen) April 4, 2014
While I enjoy the humor of the tweet as much as the next person (everyone who knows me knows I’m all about the humor), this conversation reminds me a lot of what was done with ICD-10. The “funny ICD-10 codes” got all the attention and made ICD-10 a joke in the minds of so many people. This was highlighted by this guest post on EMR and HIPAA called “Why Do People Find ICD-10 So Amusing?” Those who support the shift to ICD-10 did a poor job explaining why ICD-10 was valuable to the quality of care a patient gets. Talking about all the funny ICD-10 codes (and they are funny) goes against the goals of those who see value in the move to ICD-10.
I bring this up because the same thing could easily happen with big data in healthcare. While it’s funny to think about how a doctor might treat us if they know we had a donut for breakfast, there are really meaningful data sources beyond the EMR. If we focus too much on the periphery of the data, then we’re going to miss out on a lot of the value that comes from the not so funny parts of big data.
Right now our EMR systems can’t support most of the data that could come from outside the EMR. However, that shift is going to happen and it’s going to happen quickly. My gut tells me that it will start with the wave of consumer centric medical sensors. Then, I see genomic and social data getting integrated next (both really large projects). These three areas will set the baseline for how outside data is integrated with the EMR data.
Let me offer the key points to consider in these data integrations:
–Automated: The data must pass seamlessly without the need for user interaction
–Smart Data: The user of the system needs the system to be smart. The user should only be notified with what’s actionable, but with the ability to drill into the data as needed.
–Bi Directional: The data needs to be seen and updated by both provider and patient. The system will need to have a great way to track who updated which data. However, we need both the patient and providers eyes on the data with the ability to update incorrect data.
These points should illustrate why integrating outside data is going to be such a challenge. However, it’s also why it holds such promise.