As I look at healthcare, it seems like everyone is chasing the data. Every health IT company is a data company or at least they think they are. And in many ways, every health IT company is a data company. Everything from medical devices to wearables to EHR to labs to phones to scrubs are collecting hordes of data in healthcare. The future of health IT will be built on the back of data.
What’s amazing when I talk to people is how they feel they’re limited in what they can do with the data they have. They all want more data. I often hear, “If we just had all the data, we could do [insert cool idea].” It’s fair. If we had all the healthcare data in one place in a way that was nicely structured and easily accessible, healthcare would be better for it. The innovation we could do on top of that data would be extraordinary.
Here’s the problem:
You’ll Never Have All the Data You Want in Healthcare
Every AI and machine learning algorithm I know wants more data. The more quality data you have available, the more you can do with that data. It’s going to be hard to solve revenue cycle issues if you don’t have revenue data. However, it may not be as intuitive to think that you need employee timesheet data if you want to really work on drug diversion problems. That’s why you see so many organization trying to aggregate all the data they can legally (and sometimes not legally) get their hands on. They assume the more data they have, the more they’ll be able to discover.
The reality is that you’ll never have enough data for your data science team. The AI and machine learning efforts always want more data. They want cleaner data. They want more diverse data. However, the fact that you don’t have all the data doesn’t mean you don’t have enough data to do something valuable.
Sure, we all know the AI and analytics gets better with more data. We want to make sure the AI and analytics are as effective as possible. However, we also need to adjust our expectations for it. AI and other analytics doesn’t have to be perfect. We’d love for it to be perfect, but that’s an impossible standard. If we’re waiting for perfect, we might as well stop waiting.
Comparing analytics to perfection is the wrong measurement. The right measurement is comparing an AI or analytics solution to the status quo. The status quo is humans and we’re certainly not perfect. Just hop on social media if you don’t believe me.
The reality in healthcare is that we’ve been making really important healthcare decisions with imperfect information forever. Has a doctor ever seen a patient with perfect information? Of course not, and yet we trust doctors to make the best decision possible with the imperfect information they have available to them. For the most part, they get incredible results with imperfect information.
If we apply that to analytics and AI, we need to have the same understanding. We know the AI is imperfect and the data it has available is imperfect. That doesn’t mean they can’t provide a lot of value to patients.