Pediatric Blog
One of the cool things about this blog is that I get to meet interesting people. People reach out, comment, or challenge me from time to time, which I actually enjoy. It keeps me on my toes. Many of the people I meet have fascinating stories and backgrounds. One of them is Thompson Aderinkomi.
Thompson is an Economist, he has a masters in Statistics, and he also has an MBA (I hear he is a pretty good musician too). Guess what industry Thompson is passionate about? That’s right, healthcare.
A while back, Thompson and I started talking about how to demonstrate the value that a medical home can provide. Anecdotally, we know this is the case. But I wanted to see data to support it. Moreover, I wanted to see data from our practice that supported this notion that preventive care was beneficial to both patients and the practice.
Thus, I turned over… what is the technical term? Oh yeah, a boat load of data from our practice to Thompson ( for the record, Thompson didn’t get any private patient data) too see what he could uncover.
I was going to summarize his finding, but he did such a fine job of telling me his finding, that I decided to post it just as he described it to me. Below is Thompson’s report.
Background:
It is commonly held that primary care medical homes can lead to lower health care costs. Therefore it should follow that the more preventative care a patient receives the fewer non preventive services they will receive in the primary care setting.
Suppose we have two patients that are exactly the same. If patent A has 5 preventative services and patient B has 10 preventative services how many more non preventive services will A have than B?
The reason this question is of interest is because it is often argued by opponents of primary care driven medicine that more primary care leads to more primary care services being provided in the primary care setting. This analysis seeks to test this hypothesis.
Data:
The purpose of the analysis was to test the hypothesis that more preventative care leads to less non preventative care in the primary care setting. The services provided to more than 1,600 patients from a pediatric clinic in Chicago with two doctors were analyzed.
Each service for each patient was classified as either preventative or non preventative using the BETOS code classification system. Then the count of preventative services and the the count of non preventative services were tabulated for each patient. (Note services does not mean visit. In a given visit there may be multiple services provided, also preventative and non preventative services may occur in the same visit.)
Results:
The number of preventive services provided had a statistically significant impact on the number of non preventative services provided. Namely, for every preventative service provided, on average there were 0.8 fewer non preventative services.
In other words, if we looked at two identical patients; same gender, age, length of enrollment at clinic, and health plan where patient A had 5 preventative services and patient B had 10 preventative services our regression model predicts that patient A would have received 4 more non preventative services than patient B.
Conclusion:
The results of this analysis support the common belief that more preventive care services leads to fewer non preventative care services. Furthermore, fewer non preventative care services may be an indication of better outcomes.
Limitations:
This analysis does not take quality into account nor does it take into account the services provided to the patients outside of the primary care setting. Furthermore, a more robust analysis would utilize a matched treatment control design to further isolate the effect the number of preventative services has on the number of non preventative services. Matched treatment control design analysis is the gold standard in analyses of this type.
Methods:
In order to isolate the effect that the number of preventative services have on the number of non preventive services, 5 controlling variables were included in a linear regression analysis. The controlling variables were: length of enrollment, gender, total visit count, age, and health plan.
The reason this is done can be explained with an example. Take for instance enrollment which is the length of time the patient has been with the clinic. It follows that the longer a patient is with a clinic the more care, both preventative and non preventative they will receive.
So the enrollment variable must be included in the analysis to control for this known phenomena. The same explanation applies to the other variables. Also, a future analysis should control for each pediatrician as there were two MDs at this clinic.
End of Thompson’s report
I find this data fascinating. Especially because it is data from our practice. Of course, we aren’t making a huge discovery here. For those of us on this side of the equation (the provider side), we’ve known that the best medicine is preventive medicine.
If less sick visits or non-preventive visits is the goal, then there is not doubt that preventive medicine ought to be held as a more effective way keep patients healthier than they would otherwise.
Here is the big conundrum
If a patient with 5 preventive visits is likely to have 4 more sick visits than a patient with 10 preventive well visits, from a mere financial stand point, the patient with 5 well visits and 4 sick visits is far less expensive than the patient with 10 well visits.
Let’s say you have a business, and your objective is not wellness, but rather profits, and returning value to shareholder, how would you design a health insurance plan to maximize your return on investment? Perhaps one could limit the number of preventive wellness visits?
How about if you encourage patients to go to Retail Clinics instead of the more expensive doctor’s office for the simple runny nose? Perhaps you would not pay for ancillary preventive services such as vision and hearing screens which drive cost up? After all, you know that less preventive wellness visits, will end up being cheaper in the long run.
And here lies the huge conflict between the insurance company and doctors.
Doctors would rather prevent illnesses than to treat it. But insurance companies would rather (from an economical stand point) pay for illnesses than pay for wellness/prevention medicine. If our practice’s data is like other practices, then financially speaking, it is far better for the insurance companies to limit the number for preventive wellness visits, even though the patient might end up seeing the doctor more often for sick or non-preventive visits than to encourage more preventive wellness visits.
Do you have any thoughts on Thompson’s finding?
(This blog was originally posted on Pediatric Inc)
Brandon Betancourt is a business director for a pediatric practice in Chicago. He is a speaker, consultant and blogger. You can follow him on Twitter @PediatricInc or visit his blog at PediatricInc.com