Take Two Aspirin and Call Me When Our Healthcare Systems are Better Connected
Have you ever been in a situation where you are so baffled and amazed at a process that your only response is, "Really!?!"
As in: many health care providers are still not using analytics as the basis for decision-making? Really!?!
If you're with a fan of Saturday Night Live, Seth Meyers performs a sketch called "Really!?!" He takes a surprising news story asking a multitude of questions at the expense of the subject to humorously attempt to explain why people would take such ludicrous actions.
My "Really!?!" moment was recently directed towards a large health system.
A relative of mine needed some tests, possibly a procedure, and was looking for a second opinion. He received a recommendation for another physician, made an appointment and was told it would take two days until he could be seen. Not because the doctor was too busy, (it was not an emergency, but fairly urgent) but because it would take at least two days for the courier service to deliver his charts, images, and test results from one hospital to the other.
Really!?! Courier service!?! A carrier pigeon might actually be more efficient.
We live in a paperless society, and we are still hand delivering records at the individual patient level. Granted, this is a much better option for everyone involved than the inconvenience and expense of repeating the tests, but shocking nonetheless.
Regardless of what acronym you want to use or what you choose to call them, it turns out that electronic health records (EHR), electronic medical records (EMR) or health information exchanges (HIE), in this scenario, are only accessible within the same health system. Healthcare facilities are in the process of implementing electronic systems (and business intelligence), getting away from paper based charts, but still not at the practical/useful level that we need to get to as a society.
What does this have to do with analytics? By having all of these records in a digital format, it allows patterns and relationships to be identified through the process of data mining.
For example, let’s say a doctor is trying to diagnose a 53 year-old Caucasian male, who is 5'11, weighs 183 pounds, with a blood pressure of 140/90, a history of high cholesterol and stomach ulcers, and is complaining of shortness of breath. The lab tests reveal anemia, but how confident is the doctor in these results and what’s the next step?
To the layperson this may sound like nonsense but to the trained professional there are a few possible scenarios that come to mind. The same is true, if not more so, for a computer.
But, what if this is a new doctor? Or a rural clinic that sees five patients on a busy day? Or a rare illness with a much more complex medical history than I outlined above?
What if you could utilize electronic medical records from an entire health system, state, region, country or (dare I say) the world? The more data that is available the more accurate these models become. Think again about my example above, I came up with nine variables for my sample patient, but with a little more effort it’s easy to conceive 30+ variables per patient, not to mention all of the data that is unstructured text in clinician notes. Multiply that by all of the patients in a hospital (or all of the citizens in a country) and there is easily petabytes, if not exabytes, of data. How’s that for big data?
This reduces misdiagnosis, increases utilization of optimal treatment protocols, creates fewer harmful drug interactions, decreases readmission rates and hospital stays, and is less expensive for everyone.
That is exactly what the following clients did.
California Pacific Medical Center had an eight percent reduction in mortality rates for heart attack patients, shorter lengths of stay and improved cost savings. Centerstone Research Institute recognized a 30 percent increase in revenue and 25 percent increase in treatment plan compliance, signifying improved quality. Seton Healthcare identified patients likely for re-admission and introduced early interventions to reduce cost, mortality rates, and improved patient quality of life.
And analytics aren't only useful for recommending treatments for optimal patient outcomes. Analytics is also being used to:
· Increase patient satisfaction and loyalty – not only if the treatment of their illness/injury is successful, but understanding items such as: was the staff friendly and responsive, what did they think of the food, did their visitors have a hard time finding parking, etc.
· Improve disease management and population health – thinking of broader outcomes and satisfaction issues that take into account general health and wellness, such as the impact of diet and exercise on overall health or the impact on specific illnesses.
· Maximize resources for hospital operations – how does a hospital determine the number of nurses or technicians to staff in a certain unit at a specific time, or when do you schedule maintenance on your critical equipment (x-ray, MRI, CT scans, ultrasound machines) so it’s not a busy time in the department?
That's the power of analytics and Smarter Healthcare.