Newtown, CT – April 25, 2015
Appropriateness is the code word used by health plans and health care providers to describe whether a service or set of services is warranted – Unfortunately, for the most part, accurately determining whether or not a service is appropriate requires a review of a patient's medical record. To-date, plans have nevertheless tried to quantify the appropriateness of services by resorting to certain utilization measures. For example, number of procedures, treatments or specific services per 1000 plan members. However, these rates, calculated at a provider level, can lead to very misleading conclusions without the medical context of the patients managed by that provider. Consider two orthopedic practices, one that mostly handles complex fractures, and one that mostly manages bunions. The first will likely have a far greater rate of diagnostic imaging than the second. As such, looking at the rate of imaging studies for both and deeming that the first is doing an inappropriate quantity of those studies might be the wrong conclusion. The same can be said about the volume of procedures, making the determination of appropriateness very difficult for payers. And that's one of the oft-repeated concerns about procedural bundled payment programs – that they can create an incentive for excessive procedures. And yes, they might, unless they're done in the context of the patient's medical conditions.
What this means to you – Let's consider another example, with two cardiology practice, each managing a mix of patients with different cardiac conditions. The first practice's patients with coronary artery disease have an average yearly episode price of $5,500 and that price includes all modes of treatment, even angioplasties and bypasses. The price is also adjusted for severity, taking into account the mix of conditions of each patient. Using the same technique, the second practice's patients have an average episode price of $4,500. Finally, let's also assume that the two practices have a similar rate of potentially avoidable complications. The comparison of the two practices would lead us to the conclusion that the second is more efficient than the first, and that's true. A further dissection of the data, looking at rates of angioplasties and other coronary interventions for each practice, per 100 patients, shows us that the first practice has a rate twice as high as the second. And now we have the beginning of an answer about the appropriateness of those treatments. At the very least it's perfectly reasonable to ask some hard questions. As we've delved ever more deeply in understanding warranted and unwarranted variability – variability that is warranted by the differences in the clinical make-up of the patients, and variability that cannot be explained by that make-up and is therefore unwarranted – we've come to observe that there is very often a close relationship between the rate of procedures and higher episode costs. To an extent, there's nothing new here because many researchers, including our friends at Dartmouth, have long shown the unexplainable variation in procedure rates, but our new contribution to the field is to relate procedures and treatments to conditions, to account for each independently and jointly, to adjust for the severity of patients with those conditions, and to pinpoint the source of variability. We still can't say for sure if a treatment or procedure was appropriate, but we can venture a pretty good guess, and that's a lot better than what the field has been able to do to-date.