Paid content supplied by John Lynn
We’ve all had the experience where our smart phone or a consumer email list sends us an alert “customized to our preferences” that misses the mark. Experiences like these and many others teach us that we should be careful about how we use and trust data since the data could be wrong. That’s an extremely important lesson for healthcare as we enter a new world of healthcare reimbursement where what we get paid will largely be driven by data. Done correctly, healthcare will lower costs, improve care, and better serve even the most vulnerable patients on the back of high quality data analytics. Done incorrectly, healthcare will treat the wrong patients, increase costs and miss helping the patients that need healthcare the most. This is the high stakes game of health data analytics.
Healthcare is currently being engulfed by data from every angle. While this is extremely challenging, we’re seeing examples of successfully using the data to improve healthcare. For example, one university medical center used predictive analytics developed from a dataset of over 250,000 patient admissions to create an early warning system which can identify high-risk hospital ward patients and improve ICU triage decisions as much as 48 hours in advance. This is why healthcare data analytics is so important and essential to the future of healthcare. The right data analytics at the right place at the right time is going to save lives and money.
Stan Huff, CMIO at Intermountain, likes to share an experience they had in their hospital. When evaluating different treatments for patients at their hospital they found that one treatment would cause problems in 4 in 100 patients while an alternate treatment would only cause problems in 3 in 100 patients. It turns out that the human mind can’t comprehend a difference in quality of 4 in 100 versus 3 in 100. However, computers can tell that difference. This is the heart of health data analytics and illustrates why we need appropriately research health data analytics to assist in care.
The challenge is that there are thousands of treatments, protocols, approaches, analytics, etc that need to be tested and evaluated for their efficacy. No one organization will be able to evaluate every healthcare analytic out there. We’re going to need to create a way to share health data analytics findings across organizations so that everyone benefits.
We’re starting to see this type of sharing happening between healthcare organizations. Some organizations are doing it in an open source manner on the back of the FHIR protocol where any organization can take their work and implement it in their organization. Others are creating commercial platforms where a healthcare organization’s research can be commercialized and shared with other organizations. Both models can work, but we need hundreds and thousands of more organizations and people involved in this health data research and sharing if we really want to extract all of the benefits health data analytics can provide. While cognitive computing and neural networks is showing promise, we still need humans to assist in the process.
What’s particularly interesting about this high stakes “game” of healthcare data analytics is that those that are most successful are going to define what the future of healthcare will look like. Health data scientists’ analytics discoveries are going to create a standard of care that will be required of every healthcare organization. It could literally be considered malpractice for someone to practice medicine contrary to what the health data says about a patient. Sure, there will be exceptions to the data analytic, but there will have to be some strong mitigating reasons to ignore the analytics and proceed down a different care path.
The beauty of health data analytics is that we have a tremendous opportunity to improve care and lower healthcare costs. The scary part of health data analytics is that we could get it wrong and patients could lose their lives.