This week’s selection is another contribution to the growing body of papers that look at physician practice outcomes at the patient level. The co-hosts discuss two major themes: When it comes to physicians in practice, what should clinician-educators focus on to make the most difference in patient outcomes? and, Does performance improve with practice, or does it decline over time?
KeyLIME Session 136 – Article under review:
View/download the abstract here.
Norcini JJ, Boulet JR, Opalek A, Dauphinee WD. Patients of doctors further from medical school graduation have poorer outcomes. Medical Education. 2017. 51(5):480.
Reviewer: Jason R. Frank (@drjfrank)
When it comes to physicians in practice, what should clinician-educators focus on to make the most difference in patient outcomes? Is it pay for performance? Big pharma’s influence? High-quality conferences? Just-in-time learning? Recertification?
Continuing professional development is that phase of meded that is concerned with physicians in practice. We educators spend an awful lot of our energies focused on UGME & residency education, though these two phases of a medical career are clearly the shortest. CPD is perhaps the neglected, but critical phase of medical education whose time for attention has come.
But what to focus on? The literature is diverse and contradictory. The KeyLIME Podcast has reviewed papers on physician outcomes a few times already. Some papers suggest that physician volumes of specific diseases predict patient outcomes. Other studies suggest it is just busier physicians that are better. Choudhry’s review suggested that distance from graduation was a major risk factor, while Asche’s review suggested that we all get better for every decade in practice. Finally, papers like Norcini’s in Episode 69 suggest it is knowledge that is key.
So, which is it? Does performance improve with practice, or does it decline over time?
Norcini’s group is back with another retrospective observational study in Pennsylvania, this time looking at the relationship of physician practice volume and time since med school graduation vs patient outcomes.
Type of Paper
Research: Retrospective administrative database observational study
Key Points on Methods
The authors accessed all Pennsylvania hospitalizations over 7 years (2003 to 2009) for 5 serious acute conditions (AMI, CHF, GI bleed, hip fracture, and pneumonia). Pennsylvania makes for a great laboratory, as all admissions are required to be reported to a central database. The resulting large anonymized database included 694,020 admissions in 185 hospitals by 5280 internists or family physicians. They looked at the physician characteristics of “recent practice volume” and time since graduation and compared these to patient mortality. Notably, “recent practice volume” was not really defined.
They also looked at hospital location, certification status, and US vs international medical school of origin.
They tried to control for hospital effects. Specialty had no effect.
The methods are retrospective, involve data mining, and rely on an enormous number of comparisons. These stats need to be taken with a grain of salt. We don’t know the influence of other factors, the learning habits of the physicians involved, or the impact of trainees or other providers.
The average year of graduation was 1987, and about 80% of MDs were specialty board certified. Mean number of admissions per physician was 132 +/- 159 (quite a range).
Average absolute risk of mortality was 4.1%.
The authors found that “recent practice volume” was associated with in-hospital mortality, but it was not statistically significant. (Each admission dropped mortality by about 0.007%).
By contrast, each decade since graduation from medical school was associated with a 0.455% increase in relative risk of death.
Board certification was also associated with a 7.7% drop in mortality, as in past studies. Non-US IMGs did better than those born or trained in the US.
The authors conclude that this study is a call to arms: there is a need to prevent the increase in patient mortality as physicians age. In their study, they indicated that volume of practice nor years of experience are enough to protect against a harmful trend. They argue that a robust CPD system is needed and is in the public’s interest.…
Spare Keys – other take home points for clinician educator
- This is another contribution to our growing body of papers that look at physician practice outcomes at the patient level. These are important papers by meded giants like Norcini, Dauphinee, Tamblyn, Asche, & Chen.
- Beware the conclusions of retrospective database studies: the expression is “garbage-in, garbage-out”. Many confounders cannot be accounted for.
- Beware small effect sizes that generate a significant p value and sexy headlines…They may not be educationally or clinically significant.
- We need more research on how to really enhance the quality of care we provide as medicine changes rapidly over the years of practice. Choudhry’s study suggested that teaching, being part of a group practice, and being female all help, for example.
To John Norcini, Jack Boulet, & Dale Dauphinee: #MedEd gurus who write well.
Access KeyLIME podcast archives here
Check us out on iTunes!