Earlier this month, I completed the Clinical Knowledge portion of Step 2 of the United States Medical Licensing Examination (USMLE). The nine-hour slog of confirming when a squatting child has tetralogy of Fallot, diagnosing African American women with sarcoidosis, and never selecting “urine metanephrines” as an answer choice (even in the world of artificial scenarios I was denied seeing a patient with a pheochromocytoma) has made me reflect about the training I am receiving. As the practice of medicine in the United States changes rapidly, medical students like me feel a great tension between the model of the all-knowing, all-doing independent physician of the past and the highly segmented, team-based model of the future.
The Step exams emblemize the problem. As a test taker, I must identify the correct growth medium for fungal cultures, distinguish among the various dysmorphic facies of children born with chromosomal abnormalities, know how to triage care for a woman whose Pap smear shows atypical squamous cells of unknown significance, pinpoint the location of a seizure that has an aura with a smell of burnt rubber, and prescribe all-trans retinoic acid for the variant of acute myelogenous leukemia with a translocation between chromosomes 15 and 17. Useful information to someone, to be sure, but I am not surprised that the National Board of Medical Examiners (NBME) found a significant gap between what we learn in medical school and the everyday practice of medicine.
Researchers have tried to discern whether these exams (or any other metric) can predict performance in residency, though mostly in small, single-center studies. Residency program directors have probably initiated the most work on this topic — see this useful compilation and analysis done at Ohio State University and, fittingly, published in the Journal of the American Academy of Dermatology. Dermatology is arguably the most competitive residency to enter: A very high Step 1 score can be the ticket to an average salary of more than $300,000, great hours, and the highest level of job satisfaction in any medical specialty. The conclusion of the research was that high scores on previous tests predicted high scores on future tests but bore little relation to supervisors’ ratings of performance.
In a recent viewpoint article for JAMA’s themed issue on medical education, members of the NBME discuss how to make the USMLE more useful in assessing competency in clinical practice. One of their stated goals is greater emphasis on the basic sciences because students “fail to recognize” its value. Recognizing the value of basic science is undoubtedly a worthy goal that I support, but will increasing medical students’ knowledge of it improve the quality of healthcare delivery?
I sometimes wonder if it would be more valuable for medical students to memorize clinical practice guidelines. Maybe then children would receive more than 46.5% of indicated care in the ambulatory setting, we might reverse the decade-long slide away from guideline-recommended care for chronic back pain, and more than 62% of adults would receive appropriate prescriptions; in the process, we could prevent 67,996 deaths a year from heart failure. A recent effort by students to learn and implement quality-improvement initiatives and, thereby, increase adherence to screening guidelines for diabetes, dyslipidemia, HIV, and cervical cancer was probably a more worthy endeavor than if they had organized a journal club to discuss a review article on IVIG for autoimmune and inflammatory disease.
In the future, my greater basic science education may add to my core knowledge and intrinsic value as a physician. But for now, the popular advice is not to hire more physicians with all their refined qualifications, as discussed in an article in Medical Economics:
If you are having trouble finding physicians to join your practice, dismayed by their demands or expectations at interviews, or concerned about their high cost or need to be a partner, hiring a physician assistant (PA) or nurse practitioner (NP) may be your answer.
It’s actually pretty good advice. After all, NPs who work in the UK’s National Health Service have more satisfied patients than doctors do, with no difference in health outcomes. Nurse endoscopists in the U.S. are as accurate and safe at flexible sigmoidoscopy as experienced gastroenterologists. Certified registered nurse anesthetists (CRNAs) can perform the same set of anesthesia services, including open heart surgeries and organ transplants, as anesthesiologists, and can work unsupervised without increases in patient complications or deaths. Minnesotans who go to retail clinics receive the same level of care as at a physician’s office or an urgent care center, and at lower cost.
My school, in keeping with current trends and feedback from students, is cutting the pre-hospital rotation component of our curriculum from 2 years to 1.5 years. The goal is to give us more time to explore electives and make a better decision about what specialty to enter. That’s a good idea, but it also means that the lecture-based, basic-sciences portion of my education is a mere 6 months longer than what the PA curriculum sets aside for the same material. My friends who have graduated from PA programs are now expected to operate on the level of a resident and, on some surgery floors, to effectively manage all post-op patient care, while being paid well and working just 40 hours a week. I take comfort in knowing that, several years from now, my extra 6 months studying the basic sciences will blossom into the proper credentials to enter a subspecialty.
In the absence of evidence, economic factors will continue to force change. PAs and NPs represent about 30% of the current primary care workforce in the U.S., and that number is constantly increasing. Primary care PAs represent about 31% of all PAs in the U.S., and like their PCP counterparts, their median compensation of $85,000 ranks lowest among specialties. The incentive is, as with physicians, toward specialty care.
The nimbleness of the PA profession is an incredible advantage. It has few traditions and taboos, so PAs can go anywhere that has a demand and, probably, administer about the same amount of recommended care as doctors do at a lower cost. It would not be surprising if someday a study shows that PAs who had worked in a cardiology practice and had taken classes in reading ECGs and echocardiograms could obtain patient outcomes as good as the cardiologists did in this Duke study.
Imagine a more hypothetical scenario, one in which you create a healthcare system in a society that doesn’t have one. Would you decide that the proper way to train a highly specialized physician, such as a cardiothoracic surgeon, is with four years of medical school, five to seven years of general surgery, and two to three years of a fellowship? Or would you cut out all the parts involving hernia repairs, cholecystectomies, and lipoma resections? Does performing interventional radiology require four years of a radiology residency? It appears that the interventional radiologists themselves are questioning that assumption with a new pilot program.
I support the NBME’s decision to stake the value of future doctors on the basic sciences — it is a forward-thinking approach. Currently, though, we cannot accurately calculate the added value of my ability to detect a rare condition, given my additional scientific training. It would be even more difficult to calculate the risk for serious complications from a delay in diagnosis and treatment because I did not detect an underlying condition until it manifested in an obvious way — or until it was too late. I am optimistic that the day will come when the NBME can make these types of calculations. When it does, I’ll be ready for their metanephrines.
Offer your thoughts on these reflections from Nicholas. Is the training we give to medical students changing adequately to meet the new realities of healthcare delivery?