As I sit here on a weekend day amidst multiple MCQ banks of USMLE Step 3 prep, I take a break and read with interest a new research article on a head to head comparison of physicians to nurse practitioners (NP)/physician assistants (PA) (1). I’m interested as I use to be an NP prior to becoming an MD. I want to like this article. I mean, no study is perfect, but it does provide support for increased graduate medical education (GME) funding on the notion that hospitalist-resident teams were more economically efficient than hospitalist-midlevel provider (MLP) teams. This study found that while relative expected mortality (REM) was higher on the hospitalist-resident services (P ,.001):
- The expected direct costs per hospital discharge were significantly higher for hospitalist-MLP teams
- Shortened length of stay (LOS) for Hospitalist-resident teams which saved 1.26 days per patient.
- Patient satisfaction scores were higher than
In other words, despite caring for patients with higher REM and longer critical care requirements, LOS was shorter by 1.26 days, and per-patient direct costs derivedfrom hospital charges were lower by $617. This is all good data right? More or less a head to head comparison study on outcomes for two groups supposedly doing the same role and function. This article, as with a similar Authentic Medicine blog article attempts to compare physicians to MLPs on the variables of direct patient costs and length of stay (current article)(1) and intermediate diabetes outcomes (previous article)(2). But can you really compare two different disciplines? The major assumption of both studies is that both physicians and MLPs function fundamentally in similar roles thus you can measure similar outcomes and obtain efficacy data for each group. This assumption may be a fallacy. You decide. Can you really take a football team against a soccer team and put them in a football game and measure the efficacy of each group in the amount of touchdowns? The two groups are analogous to Apples & Oranges. Enter the Apples & Oranges Fallacy which one could say invalidates the current study and the previous study. The current study looks at these variables between resident physician and MLP inpatient teams. The prior study compared outcomes of NP’s and PA’s to physicians in primary care. Are these studies valid research? Can you really take apples and oranges and measure the outcomes of apples? Can you compare apples (resident physicians) doing apple things to oranges (MLPs) doing apple things? The current article is based on this assumption that MLPs practice medicine. The prior study says “this study adds additional support to the conclusions of previous studies that patient outcomes are similar for those with physicians, NPs, and PA as PCPs (2).” The conclusion reported: “More research on the effect of each role on access, cost, quality of care, and satisfaction of providers and patients is needed before answers will become clear (2).” Hence why I’m calling it the Apples & Oranges Fallacy. While I think the data is noteworthy, it’s the premise on which the study is built that is suspect. You decide.
Lastly, if you look closely at the current study, what is common in both groups being studied? Both groups have an attending physician hospitalist. How does this affect the overall results? This is The Banana Factor I’ll save for another time.