However, it has been shown that the physical examination skills of primary care physicians have clinically important deficiencies,⁴ and that the physical examination skills of medical school and residency graduates are lacking,^5,6 raising the concern that erosion of these skills may lead to a downward spiral in which those who are inadequately taught will be ill-prepared to teach the next generation of medical students and residents.⁷ It is clear that improvement in physical examination instruction should be a priority for medical educators.

To achieve this goal, we must first recognize the obstacles present in our current method of teaching. Most physical examination textbooks and instructional syllabi do not explicitly distinguish elements that have proven clinical utility from those that do not, nor do they distinguish elements that are difficult to master, such as the fundoscopic exam, from those in which it is easier to achieve skill, such as measuring blood pressure. Furthermore, the skills of those who teach physical examination may not be consistent, given that instructors may include private physicians, full-time medical school faculty, subspecialty fellows, residents, and in some cases, fourth-year medical students. The ideal physical diagnosis teacher would have both high self-confidence in skill and a belief that the physical exam has high utility, in addition to demonstrated or actual skill in the physical exam. However, it is unknown how these teachers view their own competence in physical examination skills and the utility of these skills.

Regarding which specific physical examination skills should be emphasized, JAMA has published an ongoing series of critical reviews entitled the “Rational Clinical Examination,”^8–19 and this information can help guide the selection of specific skills that deserve emphasis. To identify those specific skills which should be targeted for improvement in instruction, we should know the skills in which physicians, residents, and medical students are least confident in their ability, as well as those skills that they view as being most important. These attitudes may play an important role in how physical examination skills are both learned and taught.

From October 2003 to January 2004, the survey was distributed to Brown Medical School (as it was then known) third- and fourth-year medical students, Internal Medicine and medicine–pediatrics residents, Internal Medicine chief residents, and general Internal Medicine faculty at Rhode Island Hospital, a tertiary care teaching hospital in Providence, RI. The survey was distributed via mail, e-mail, or on a face-to-face basis, but respondents’ answers were recorded anonymously. Nonrespondents received 2 additional mailings in the same manner as originally solicited. The survey was pilot-tested on a convenience sample of second- and third-year Internal Medicine residents to gauge length, readability, comprehension, and face validity.

Self-Confidence Table 2 lists all mean self-confidence scores for overall physical examination skills and for each individual physical examination skill by training level. In terms of overall self-confidence, M3 students were lower than M4 students (p=.002), PGY-2 through 4 residents (p=.01), and faculty Internists (p<.001), but not compared to PGY-1 residents (p=.95). PGY-1 residents had lower overall self-confidence compared to faculty Internists (p=.01). There was a weak but significant positive correlation between mean overall self-confidence score and training level (r_s=.20, p<.001), and similar correlations were exhibited for most individual skills. Collectively and by training level, respondents reported the highest self-confidence in measuring blood pressure and lowest in the nondilated fundoscopic examination using an ophthalmoscope to assess retinal vasculature. Mean self-confidence of M3 students was below “Neutral” for 7 of the skills, and mean self-confidence of PGY-1 residents was below “Neutral” for 5 of the skills. Faculty Internists reported self-confidence above “Neutral” for all skills. Table 3 shows the comparison of self-confidence in individual skills by training level, focusing on M3 students and PGY-1 residents. M3 students had lower self-confidence in 6 skills compared to M4 students, in 7 skills compared to PGY-2 through 4 residents, and in 13 skills compared to faculty Internists, but in only 1 skill compared to PGY-1 residents. PGY-1 residents had lower self-confidence in 4 skills compared to M4 students, in 2 skills compared to PGY-2 through 4 residents, and in 10 skills compared to faculty Internists.

Perceived Utility Table 4 lists mean perceived utility scores for overall physical examination skills and for each physical examination skill by training level. Both groups of residents reported overall lower perceived utility compared to M3 students (PGY-1, p=.02 and PGY-2 through 4, p=.02) and M4 students (PGY-1, p=.006 and PGY-2 through 4, p=.004). Across all levels, measuring blood pressure had the highest perceived utility. For all levels except faculty Internists, detecting clubbing had the lowest perceived utility; for faculty Internists, the nondilated fundoscopic examination using an ophthalmoscope to assess retinal vasculature had the lowest perceived utility. No group rated the utility of any of the skills as less than “Neutral.” There was a weak but significant negative correlation between mean overall perceived utility and training level (r_s=−.24, p<.001).

We expected that M3 students would have less self-confidence in most individual skills compared to other levels. However, we found that M3 students had less confidence than PGY-1 residents in only one skill, and a lack of significant difference in overall self-confidence in physical examination skills between these 2 groups. Overall self-confidence was generally higher at advanced training levels, but M4 students had more self-confidence than did PGY-1 residents. One explanation may be that M4 students have an artificially inflated self-confidence because, having completed the third year of medical school, they possess a new sense of accomplishment and mastery, whereas PGY-1 residents, confronted with newfound responsibility for patient care, realize their deficiencies in skill, thereby casting lower perceptions of their abilities. Alternatively, PGY-1 residents’ lower self-confidence may be attributed to their having learned the physical exam at various different medical schools with different curricular experiences, whereas M3 and M4 students were from one institution, and PGY-2–4 residents had spent at least 1 year training in the same institution. A recent study of students in an inpatient medicine clerkship showed that students spent more time with interns than residents, and seemed more satisfied with physical examination instruction by interns compared to residents.²⁰ However, our study implies that, during a critical time in their education, medical students in clerkships may not be learning the physical examination from the most optimal teachers. Interestingly, a recent study by McMahon et al. showed that upper-level Internal Medicine residents rated their overall physical examination skills lower than those of medical students and PGY-1 residents.²¹ In our study, we found that M3 students had lower self-confidence, and PGY-1 residents had similar self-confidence, when compared to upper-level residents.

Our results also show that even the faculty Internists, although having relatively higher confidence in their physical examination skills, nonetheless still have areas where they rate their skills as only slightly above “Neutral.” It has been shown that faculty attending physicians’ observations of housestaff physical examination skills can reveal a high incidence of errors.²² However, our study raises the question of how faculty can be the “gold standard” of physical examination skills if their self-confidence is not much different than that of housestaff.

Our study demonstrates that there are also differences across levels of training in perceived utility regarding overall and individual physical examination skills, as we saw with self-confidence. It was reassuring that all groups reported greatest perceived utility in measuring blood pressure and reported perceived utility above “Neutral” for all skills. However, 2 commonly taught skills, detecting clubbing and determining vertical liver span, consistently had the lowest perceived utilities, raising concern that the clinical context of these skills may not be fully appreciated. Furthermore, our finding that housestaff have an overall decreased perceived utility of the physical examination compared to medical students is troublesome because this attitude could potentially negatively affect the students’ perspective on clinical skills.

Overall, the physical examination skills with the largest numerical differences between self-confidence and perceived utility included distinguishing between a mole and melanoma, detecting a thyroid nodule, interpreting a diastolic murmur, detecting a breast mass in a female patient, and the nondilated fundoscopic examination using an ophthalmoscope to assess retinal vasculature. Although these numerical differences have no measurable units, they indicate, nonetheless, areas in which instruction and practice in these physical examination skills need improvement at all training levels.

Regarding specific physical examination skills, a previous study of Internal Medicine and family medicine residents found high importance attributed to cardiac auscultation and relatively lower self-confidence in this skill, but without comparison to other skills.⁵ Although we have shown that, compared to other skills, interpretations of systolic and diastolic murmurs have a lower perceived utility, there still exists a gap between self-confidence and perceived utility, particularly with interpretation of a diastolic murmur. These findings combined demonstrate a continued need to improve the teaching of cardiac auscultation skills and emphasize the utility of these skills.

Interestingly, although having a relatively lower self-confidence score, the nondilated fundoscopic examination using an ophthalmoscope to assess retinal vasculature also had a relatively lower perceived utility score. In fact, among the faculty Internists, this skill had the lowest perceived utility score. Was this because it is a difficult exam to master, and so medical students and physicians are prone to de-emphasize its importance? Would our results have been different had we specified using the new technology of the PanOptic ophthalmoscope? It is also possible that respondents felt that this exam is better done by an ophthalmologist and could be accomplished by emergent consultation if necessary. This area could be further studied to see if there is a change in attitude toward this skill using this new technology.

Our study had several limitations. The study investigated a limited set of physical examination skills and was conducted at only one institution. The study was done during an academic training year, thereby including variation in skill level because of differing educational and training experiences within each medical student and resident class. The students were from one medical school, whereas the residents and faculty graduated from many different schools. The residents are also all former students who chose Internal Medicine, whereas the students may not necessarily be entering Internal Medicine. Our questions regarding self-confidence and perceived utility were inherently subjective. Furthermore, there was no objective assessment of participants’ physical exam skills to validate the self-ratings. It is well-known that physicians’ self-assessment do not correlate with actual skill when examined with external assessment, as recently summarized in a systematic review by Davis et al.²³ In addition, our use of a 5-point Likert-type scale in assessing self-confidence and perceived utility, even with “3 = Neutral,” does not provide a clear minimum acceptable standard with which to compare respondents’ ratings. Finally, it is unclear what the implication of confidence in and perceived utility of physical examination skills has on patient care and outcomes.

However, our study had strengths that should be noted. To the best of our knowledge, no study has been done investigating the attitudes toward a wide variety of physical examination skills over the spectrum of educational and training levels from medical student to faculty Internist at one institution. Furthermore, no other studies have shown a comparison of the relative perceived utilities of specific components of the physical examination. Those who view specific skills as having less utility may be less likely to be motivated to improve their competence in these skills, even if they have less self-confidence.

Knowledge and performance of the physical examination is essential to a physician. For most medical students, however, the learning of the physical examination during medical school is best described as variable. It has long been known that teaching of clinical skills requires time and patience, yet the most qualified physicians for this instruction are often unavailable because of other commitments, thereby leaving medical housestaff recruited to teach the physical examination to medical students.²⁴ Yet, few residency programs provide additional instruction in physical examination skills. Housestaff, when in doubt of a finding, typically turn to attending physicians for guidance, and it is assumed that they, given their experience, are more proficient in the physical exam. However, as we have shown, even if faculty Internists may potentially be competent in their skills, they do not have complete confidence, and this may contribute to the decline in teaching at the bedside, although they acknowledge that this will lead to decreased skills in medical students and residents.²⁵