Aug 29, 2023
When the Houston Astros famously became World Series Champions only a few years after being ranked one of the worst teams in baseball, they did so in part by embracing the latest science in predictive analytics, including the concept of Wins Above Replacement (WAR). In simple terms, this metric calculates the value of a player relative to others who play the same position. If you replaced me, how many wins would you get?
Could predictive analytics be used in the same way when it comes to healthcare? Would it be possible to assess the ‘value’ of any physician relative to others? And if so, could this improve care delivery?
The answer is a clear yes.
Bringing the same analytics technology used by leading banks and retailers like JPMorgan and Amazon, together with data from the Centers for Medicare Services (CMS) and electronic health records, reveals a fascinating new picture of the quality, resource effectiveness, and practice patterns of individual physicians and hospitals. Those who embrace this new model of provider performance benchmarking will have a clear advantage as the healthcare industry increasingly shifts from payments based on fee-for-service to those based on performance, also known as fee-for-value.
Which doctor is best for me or my loved one? Which care team can handle my complex needs most effectively? Which hospital should we choose? Thousands of patients, parents, daughters, sons, caregivers, and friends face these questions each day. Coming up with answers remains a process fraught with anxiety caused by a lack of transparency and objective measures. We end up relying on the opinion of our family doctors, friends, and reputation, much in the same way that sports teams once relied on scouts to assess players. But what if we did have reliable measures to augment human opinion?
It’s also important to explore a popular myth that doctors and nurses are genetically resistant to greater objectivity and transparency about their performance. Clinicians have, in fact, had solid reasons for objecting to the prior generation of comparative metrics. Their chief complaint has been, “my patients are different.” By that, they mean unless you can precisely account for the level of difficulty associated with caring for their specific patients, you will likely arrive at the wrong conclusion about which doctor is the best performer. A classic example is a surgeon who accepts the cases that all others have refused. It stands to reason that the toughest cases have a mortality rate that would be higher than average. The question is, exactly how much higher than average should we expect this rate to be?
The reality is that, historically, the methodologies used to adjust for the difficulty of a doctor’s or hospital’s patients, also known as the case mix index, have not been precise enough to provide a truly fair assessment. This has understandably created resistance to both publishing performance metrics and using these to differentially reimburse doctors and hospitals.
Getting to the level of precision needed to compare providers more fairly requires large data sets that include not only claims and electronic health record information but also information on patients such as whether they live alone or with other people, their education status, and their prior medical history. These social determinants of health, which are very similar in nature to the information used by credit agencies to assess credit risk when approving credit cards and mortgages, add a critical element of precision to case-mix adjustment.
However, more data alone does not solve the problem. What’s also required is the ability to clean, crunch, and process that data quickly – just as rapidly as companies outside of healthcare, namely financial services companies, clean, crunch, and process data. Doing this requires moving away from electronic data warehouses, which are expensive to maintain, to cloud-based approaches where any piece of data is tagged and added to a data lake. Algorithms are then created to automatically pick, clean, and sort the data elements required for any calculation. This enables, for example, the ability to adjust the weight of any factor used to assess performance for local variations.
If the standard of care in New York happens to be different than in San Francisco, we can account for this. Similarly, a different and precise weight can be assigned to the impact of high blood pressure on stroke mortality as opposed to total joint replacement mortality. While these may seem to be blindingly obvious examples, today’s case mix and risk adjustment methodologies often fail to take these realities into account.
So what’s possible when big data and the latest analytics technology are brought together? For virtually any metric that a primary care physician might consider relevant when referring a patient to another doctor or hospital – including infection rates, readmission rates, mortality rates, cost of care (impacting co-pays), and satisfaction – it is now possible to provide apples-to-apples comparisons. As a result, we have an opportunity to create our own healthcare Wins Above Replacement or aggregate metrics that, in turn, provide people with important additional information they can use to make decisions when selecting a doctor or hospital.
Is this new generation of healthcare analytics perfect? By no means. There is still a crucial role to play for healthcare professionals in applying their expertise and judgment in guiding patients and families to the best choices for them. The baseball analogy highlights an important point: While human beings ultimately make the final recommendation, analytics open up a new level of understanding about individual provider performance.
For clinicians and hospitals, the choice is to embrace these new insights to learn, improve, and provide superior care, or to ignore them and risk being outclassed by the Houston Astros of the healthcare world. Why take that risk? Better referrals, provider selection, and reimbursement will increasingly be enabled by advanced and precise analytics. For patients and their families, there is finally the promise of the greater transparency we all deserve.
Meet the Clarify team at HIMSS21 in booth 1821 and join our Lunch & Learn session, “A single stack of truth: What can healthcare learn from financial services?” to learn how Clarify has adopted the big-data efficiencies of the banking industry and the analytics methodologies of baseball to create the most advanced analytics platform and applications in healthcare today.
This post is adapted from a MedCity News article originally published in 2019.