- Email: [email protected]
Cardiac auscultatory skills of physicians-in-training: a comparison of three English-speaking countries
Cardiac Auscultatory Skills of Physicians-in-Training: A Comparison of Three English-Speaking Countries Salvatore Mangione, MD PURPOSE: Cardiac auscultation is suffering from a declining interest caused by competing diagnostic technology and, possibly, inadequate teaching and testing of physicians-in-training. Because access to technology, traditional teaching practices, and methods of trainees’ assessment vary among different countries, we speculated that trainees’ proficiency in auscultation might also vary. SUBJECTS AND METHODS: We tested the cardiac auscultatory skills of 314 internal medicine residents (189 from the United States, 89 from Canada, and 36 from England) from 14 programs. All participants were asked to listen by stethophones to 12 prerecorded cardiac events and to answer a multiplechoice questionnaire. They also completed a survey concerning attitudes toward cardiac auscultation and auscultatory teaching received during training. RESULTS: Mean (⫾ SD) identification scores for the 12 car-
diac events ranged from 0% to 58% for American trainees (mean 22% ⫾ 12%), 0% to 58% for Canadians (mean 26% ⫾ 13%), and 0% to 42% for British trainees (mean 20% ⫾ 12%). Canadians’ cumulative scores were slightly but significantly greater than those of American (P ⫽ 0.02) and British house officers (P ⫽ 0.05). British house officers improved the most during the 3 years of training (P ⬍0.05). Canadian and British trainees had received more auscultatory teaching during medical school and residency; they had also used audiotapes more frequently (all P ⬍0.001) CONCLUSIONS: Auscultatory proficiency was poor in all three countries. Although there were slight differences among countries, the most striking finding was the consistent inaccuracy of all trainees. This suggests that variables other than teaching and testing affect proficiency. Am J Med. 2001;110: 210 –216. 䉷2001 by Excerpta Medica, Inc.
T
cians and correlate directly with an increased use of subspecialty consultations (9). Errors are particularly common in the field of chest auscultation, a time-honored art and the symbol of bedside diagnostic skills (10). Even though cardiac auscultation, competently performed, remains a valuable (11,12) and cost-effective (13) tool for the evaluation of valvular diseases, generalists-in-training are quite inaccurate in the recognition of common auscultatory events, improve little with a year of training, and are not much better than third-year medical students (14). Whether deficiencies in physical diagnosis may be limited to the United States or whether they may extend to trainees of other countries is not known. Because access to technology, teaching and training practices, and testing of residents at time of board certification vary among different countries, we speculated that auscultatory proficiency might also vary. To test this hypothesis we surveyed internal medicine residents from Canada, England, and the United States. We chose these three countries because all are English-speaking and similar in cultural and medical traditions and yet quite different in organization, emphasis on clinical skills during training, and methods of final board certification examination.
here is increasing evidence that bedside clinical skills have been waning. Only one fourth of US internal medicine and family practice residencies offer any formal instruction in cardiac auscultation (1); fewer programs offer structured teaching in pulmonary auscultation (2); only a minority of residents are closely supervised while performing a physical examination (3); and attending physician rounds are increasingly conducted away from the bedside, often in conference rooms and corridors (4,5). The decreased emphasis during training has probably contributed to the high rates of error in physical examination that have been reported among house staff (6,7). In one study, errors of omission occurred two to three times more frequently than errors of commission, and serious errors leading to major changes in diagnosis and management occurred in almost two thirds of the patients examined (8). High rates of error have also been described among university-affiliated attending physi-
From the Department of Medicine and Center for Research in Medical Education and Health Care, Jefferson Medical College of Thomas Jefferson University, Philadelphia, Pennsylvania. Requests for reprints should be addressed to Salvatore Mangione, MD, Department of Medicine and Center for Research in Medical Education and Health Care, Jefferson Medical College of Thomas Jefferson University, 1025 Walnut Street, Suite 121, Philadelphia, Pennsylvania 19107. Manuscript submitted November 5, 1999, and accepted in revised form August 25, 2000. 210
䉷2001 by Excerpta Medica, Inc. All rights reserved.
METHODS We tested 314 internal medicine residents (189 American, 89 Canadian, and 36 British trainees) in 14 university 0002-9343/01/$–see front matter PII S0002-9343(00)00673-2
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
hospitals. These included 3 London hospitals (Hammersmith, Middlesex, and St. Thomas hospitals), 5 Canadian hospitals (Queens Hospital in Kingston, Ontario, McMaster University Hospital in Hamilton, Ontario, and 3 university hospitals affiliated with McGill University in Montreal, Quebec: National Jewish, Montreal General, and Royal Victoria hospitals), and 6 American hospitals (5 in Philadelphia: T. Jefferson University Hospital, Temple University Hospital, Hospital of the University of Pennsylvania, Medical College of Pennsylvania, and Hahnemann University Hospitals, as well as Georgetown University Hospital in Washington DC). Data on American trainees have been reported in part (14).
Proficiency Test The evaluation of auscultatory proficiency was arranged as a 1-hour patient management conference, incorporated in the teaching activities of the various programs. The evaluation instrument consisted of 12 cardiac events (6 extra-sounds and 6 murmurs), directly recorded from patients, and selected from a pool of more than 250 sounds. All sounds were digitized into a MacIntosh SE 30 computer (Apple Computer Inc., Cupertino, California), their waveforms were analyzed (MacRecorder, Farallon Computing, Berkeley, California), and the final 12 events were chosen based on the purity of their sounds and good reproduction of findings typical of the corresponding disease process. The recognition of these same events had been ranked in our previous survey of family practice and internal medicine program directors as very important for practicing physicians (1). Each event was recorded on a high-fidelity cassette deck (CR W85, Fisher Corp., Chatsworth, California) and played back to trainees for 1.5 minutes using wireless infrared stethophones (Cardionics Inc., Houston, Texas) and a high-fidelity audiocassette player (Marantz PMD430, Chatsworth, California). For each event, participants were told the chest area of recording and had the opportunity to listen again if they wished to do so. After listening, they answered a multiplechoice questionnaire by checking in a forced-response format whether a particular finding was present or absent and by selecting characteristics that best described it. Answers were graded using as a reference the corresponding phonocardiographic tracing. Scores reflected participants’ accuracy in the recognition of a specific event. Item scores were expressed as the percentage of respondents, for year and country of training, who correctly identified each of the sounds tested. Cumulative scores reflected the participants’ accuracy in recognizing all 6 extra-sounds, all 6 murmurs, and all 12 auscultatory events. These scores were expressed as the percentage of total number of events (murmurs, extra-sounds, or all) recognized by each participant. Whenever respondents not only selected the right finding but also reported findings that
were absent (for example, an S4 together with the right finding of an S3 gallop), a “corrected” (ie, reduced) score was calculated. A group of 10 cardiologists attending the annual meeting of the Pennsylvania Chapter of the American College of Cardiology also took our test, and both their performance (accuracy rates ranging between 80% and 90%) and their feedback confirmed the good quality of the test sounds.
Attitudes Survey Participants also completed a one-page attitude questionnaire about self-motivated learning of cardiac auscultation, auscultatory teaching they had received during medical school and internal medicine residency, their opinions of the clinical importance of cardiac auscultation and the need to devote more time to its teaching, and confidence with their own cardiac auscultatory skills.
Statistical Analysis We determined the differences in auscultatory proficiency among trainees of different countries and the degree to which third-year residents have greater skills than first-year residents. Chi-square tests were used to compare trainees of different countries, and the between-year data were analyzed by a trend-analysis chi-square test in the three countries by the 3 years of medical residency. Trainees’ attitudes toward cardiac auscultation were analyzed by Student’s t test and chi-square tests. Relations between trainees’ proficiency and attitudes or practices toward cardiac auscultation were analyzed by Spearman rank correlations. Continuous results are reported as mean ⫾ SD.
RESULTS Proficiency The most striking finding was the overall poor performance of all house-officers, independently of their country of training (Table 1). Yet, Canadian trainees had significantly higher cumulative noncorrected scores than American or British trainees. Compared with Americans, they also had higher identification scores for the 6 murmurs, both corrected and noncorrected. American trainees had higher corrected identification scores for the 6 extra-sounds. There were no significant differences in cumulative scores when comparing first- and third-year residents of either the United States or Canada. Among British trainees, however, third-year residents had significantly higher noncorrected scores (29% ⫾ 9% versus 15% ⫾ 12% for all 12 events, P ⫽ 0.008; and 27% ⫾ 18% versus 12% ⫾ 16% for the 6 extra-sounds, P ⫽ 0.04). Canadians had higher noncorrected identification scores for the murmurs of mitral regurgitation, aortic regurgitation, and mitral stenosis than did Americans, and they also identified the murmur of patent ductus arte-
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110 211
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
Table 1. Scores for the Various Auscultatory Findings among Internal Medicine Residents of Three English-Speaking Countries US Trainees (n ⫽ 189)
UK Trainees (n ⫽ 36) Mean ⫾ SD
Scores Total Noncorrected score† Corrected score Murmurs Noncorrected score Corrected score Sounds Noncorrected score Corrected score PGY1 versus PGY3 Noncorrected (total) score Noncorrected (sounds) score
Canadian Trainees (n ⫽ 89)
22 ⫾ 12 10 ⫾ 9
20 ⫾ 12 10 ⫾ 9
26 ⫾ 13* 12 ⫾ 10
16 ⫾ 15 12 ⫾ 14
22 ⫾ 19 16 ⫾ 16
27 ⫾ 19† 19 ⫾ 17†
28 ⫾ 18‡ 9 ⫾ 13‡
20 ⫾ 18 3⫾8
25 ⫾ 17 5 ⫾ 10
20 ⫾ 10 vs 24 ⫾ 14
15 ⫾ 12 vs 29 ⫾ 9*
22 ⫾ 11 vs 28 ⫾ 14
26 ⫾ 17 vs 31 ⫾ 19
12 ⫾ 16 vs 27 ⫾ 19*
22 ⫾ 14 vs 28 ⫾ 18
* P ⬍0.05 by comparison with other countries or other years of training. † P ⬍0.001 by comparison with other countries or other years of training. PGY ⫽ postgraduate year.
rious significantly more frequently than both American and British trainees. Compared with British trainees, Americans identified more frequently the S4 gallop and less frequently the murmurs of mitral regurgitation and stenosis. When corrected scores were analyzed, Canadian trainees still maintained higher identification rates for the murmurs of mitral regurgitation, aortic regurgitation, and mitral stenosis. Differences between American and British trainees were unchanged (Figure 1). Only the recognition of the S3 gallop and the murmur of mitral stenosis improved significantly among American trainees; the identification of the early ejection sound actually worsened. For Canadian trainees there was significant improvement by year of training only for the opening snap and the murmur of mitral regurgitation (Figure 2). None of these differences persisted when corrected scores were analyzed. Among British trainees, identification rates more than doubled for 7 of the 12 events tested, but only 3 of these 7 improvements approached statistical significance (P ⫽ 0.07).
Attitudes American and British trainees attributed significantly greater importance to cardiac auscultation than Canadians (3.8 ⫾ 0.4 for Americans, 3.7 ⫾ 0.5 for British, and 3.4 ⫾ 0.6 for Canadians on a scale from 1 [obsolete and useless] to 4 [extremely important]; both P ⬎0.001). British and Canadian trainees had both pursued and received significantly more teaching of this skill during medical school and residency training; they had also used
212
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
cardiac audiotapes significantly more frequently than their American counterparts (all P ⬍0.001; Table 2). Yet British trainees had less confidence than Americans in their own auscultatory skills (1.9 ⫾ 0.7 for British trainees, 2.5 ⫾ 0.7 for Americans on a scale from 1 [poor] to 5 [excellent]; P ⬍0.001). British and Canadians trainees wished for more auscultatory teaching during training than did American trainees (P ⫽ 0.05). Trainees who had received structured teaching during residency had greater confidence (on a 1 to 5 scale) in their own auscultatory skills than those who had not been taught (2.7 ⫾ 0.7 versus 2.3 ⫾ 0.7, P ⫽ 0.02). Structured teaching was also associated with greater accuracy (cumulative corrected scores for murmurs of 19% ⫾ 18% versus 12% ⫾ 13%, P ⫽ 0.01; cumulative noncorrected scores of 23% ⫾ 20% versus 16% ⫾ 14%, P ⫽ 0.03). Trainees who had used cardiac audiotapes had no greater confidence in their own auscultatory skills (2.5 ⫾ 0.7 versus 2.4 ⫾ 0.7, P ⫽ 0.3), and they attributed significantly less importance to auscultation (3.8 ⫾ 0.4 versus 3.5 ⫾ 0.5 on a 1 to 4 scale, P ⫽ 0.001). However, they had significantly higher cumulative noncorrected scores (26% ⫾ 12% versus 20% ⫾ 10% for all 12 events, and 24% ⫾ 18% versus 15% ⫾ 14% for the 6 murmurs; both P ⫽ 0.001) Greater self-confidence in auscultation did not correlate with higher identification scores among British and Canadian trainees. Among American trainees, there was a significant correlation between skill self-perception and
Volume 110
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
Figure 1. Accuracy rates are shown by country of training. Corrected scores were calculated whenever the respondents selected not only the right finding but also findings acoustically similar and yet absent; the corrected score considered these types of answers to be invalid. Significance is reported for improvements in each of the three participating groups when compared with each other; asterisks indicate P values ⬍0.05. MR ⫽ mitral regurgitation; AS ⫽ aortic stenosis; AS/AR ⫽ aortic stenosis and regurgitation; AR ⫽ aortic regurgitation; MS ⫽ mitral stenosis rumble; PDA ⫽ patent ductus arteriosus; RUB ⫽ pericardial rub; S4 ⫽ S4 gallop; S3 ⫽ S3 gallop; OS ⫽ opening snap of mitral stenosis; MSCLK ⫽ mid-systolic click; EES ⫽ early ejection sound.
auscultatory accuracy (r ⫽ 0.21, P ⫽ 0.02 for cumulative corrected identification score, and r ⫽ 0.24, P ⫽ 0.006 for the corrected murmur identification score).
DISCUSSION We sought to ascertain whether there were differences in cardiac auscultation skills among trainees of different countries, perhaps as a result of disparities in the assessment of bedside skills at time of board examination. We
could not, however, survey residents who were truly representative of the three countries. Thus, extrapolation of the findings is difficult. Still, within these inevitable limitations, British trainees, who had received more auscultatory teaching and who were expected to undergo an objective assessment of physical examination skills at the time of board certification, improved the most over 3 years of training. Canadian residents, similarly trained and tested, had the greatest accuracy. Yet, although significant, these differences were slight.
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110 213
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
Figure 2. Differences in diagnostic accuracy are shown by country and year of training. Gains (or losses) in diagnostic accuracy between third-year and first-year residents were calculated as the difference between the noncorrected scores. Significance is reported for improvement (or worsening) by year of training; asterisks indicate P values ⬍0.05. MR ⫽ mitral regurgitation; AS ⫽ aortic stenosis; AS/AR ⫽ aortic stenosis and regurgitation; AR ⫽ aortic regurgitation; MS ⫽ mitral stenosis rumble; PDA ⫽ patent ductus arteriosus; RUB ⫽ pericardial rub; S4 ⫽ S4 gallop; S3 ⫽ S3 gallop; OS ⫽ opening snap of mitral stenosis; MSCLK ⫽ mid-systolic click; EES ⫽ early ejection sound.
More importantly, the overall performance of all housestaff officers, independent of their country of training, was poor. This suggests that other variables beyond teaching and testing affect proficiency. One of these variables may be the availability of diagnostic technology. Although advanced technologies may be more widespread in the United States than elsewhere (by 1993, for example, Philadelphia hospitals had 63 magnetic resonance imaging scanners, compared with 57 in all of Germany and 12 in all of Canada [15]), all three countries that we tested share an equally sophisticated and technologically advanced medicine. Easy access to technology may correlate inversely with the time and attention de-
voted to physical diagnosis during training. In a nationwide survey of US internal medicine residencies, for example, we found that small community-hospitals, presumably less equipped with technologic resources, provided significantly more cardiac auscultatory teaching than large university-affiliated tertiary care centers (16). Unlimited access to technology, combined with its glamour and sophistication, may give trainees an easy alternative to painstaking and time-consuming bedside examinations. The issue of time in medical care is very important (17,18). As the time for physician-patient interactions continues to decrease—the mean visit length
Table 2. Teaching of Cardiac Auscultation Received during Medical School and Residency, and Attitudes of Participating Residents toward Auscultation US Trainees Characteristic or Attitude Teaching received During medical school During residency Independent use of audiotapes More teaching needed in medical school More teaching needed in residency
214
February 15, 2001
UK Trainees
Canadian Trainees
P Value
Numerator/Denominator (Percent) 93/137 (68) 19/121 (16) 43/121 (36) 119/126 (94) 111/119 (93)
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110
27/27 (100) 13/13 (100) 7/7 (100) 22/22 (100) 23/23 (100)
58/60 (97) 25/29 (86) 36/39 (92) 64/64 (100) 67/67 (100)
⬍0.001 ⬍0.001 ⬍0.001 0.09 ⬍0.05
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
was 7 to 9 minutes in a 1991 study (19)– better skills at history-taking and physical diagnosis have become more crucial for reducing the cost of care without decreasing its quality (13). These fundamental crafts alone can still deliver a diagnosis for a high proportion of patients, at least in the outpatient setting, where most care occurs (20,21). Another factor that might have contributed to the trainees’ inaccuracy in this study is the changing prevalence of valvular diseases in the Western world. Rheumatic fever, for example, has almost completely, but not entirely (22), disappeared in technologically advanced societies. Conversely, it remains quite prevalent in developing countries (23) where access to technology is limited and strong bedside skills, such as physical diagnosis, are still an essential part of medical students’ training and of patient’s evaluation. It would have been interesting to compare the auscultatory proficiency of trainees from Western world countries with that of trainees from developing nations. A possible limitation of our study is that we used recorded sounds and did not provide any medical history. Physicians’ ability to recognize taped sounds may correlate poorly with bedside skills, and clinical history provides important guidance. Our data, however, confirm other observations. Paauw et al (24) assessed the physical examination skills of 134 primary care providers using real patients, all of whom had physical findings. Even though physicians’ attention was directed to sites of patients’ lesions, their performance remained poor. Only 38% of participants recognized two murmurs of aortic insufficiency, only 17% detected diffuse lymphadenopathy, and only about half noticed any abnormality in patients with Kaposi’s sarcoma or oral hairy leukoplakia. Innovative educational methods may be necessary to rejuvenate the art of physical diagnosis. However, teaching and testing may not suffice if widespread use of diagnostic technology continues to preempt efforts at the bedside. We believe that the best way to revive skills such as auscultation would be to have a cadre of clinicianteachers who could serve as role models for physiciansin-training, revive the fascination with a Sherlockian approach to bedside diagnosis, and show how to use technology more judiciously and cost effectively. There are already indications that this strategy can be highly successful with students (25). After all, it was a brilliantly deductive and charismatic professor, Dr. Joseph Bell, who captured the imagination of young Arthur Conan Doyle and inspired the creation of Sherlock Holmes. Rekindling interest and competence in physical diagnosis will therefore require a substantial effort at the bedside as well as a considerable shift in attitudes. This shift may include readdressing issues of physicians’ liability, patients’ demands, and promotion and incentives for clinical faculty willing to invest time and effort in teach-
ing these skills. The challenge to educators is whether such investments are warranted in times of increasing technologic sophistication. To this end, the value of physical diagnosis will need to be redefined and supported continuously by ascertaining which elements remain important in the context of contemporary medicine.
ACKNOWLEDGMENTS The author wishes to express his thanks to Dr. Graham Leech of London and to Messrs. Keith Johnson and David Gibbons for their help with the English trainees’ testing.
REFERENCES 1. Mangione S. The teaching of cardiac auscultation during internal medicine and family medicine training: a nationwide comparison. Acad Med. 1998;73:S10 –12. 2. Mangione S, Loudon RG, Fiel SB. Lung auscultation during internal medicine and pulmonary training: a nationwide survey. Chest. 1993;104:70S. 3. Burdick WP, Schoffstall J. Observation of emergency medicine residents at the bedside: how often does it happen? Acad Emerg Med. 1995;2:909 –913. 4. Shankel SW, Mazzaferri EL. Teaching the resident in internal medicine: present practices and suggestions for the future. JAMA. 1986;256:725–729. 5. Collins GF, Cassie JM, Daggett CJ. The role of the attending physician in clinical training. J Med Educ. 1978;53:429 – 431. 6. Wiener S, Nathanson M. Physical examination: frequently observed errors. JAMA. 1976;236:852– 855. 7. Mangione S, Burdick WP, Peitzman S. Physical diagnosis skills of physicians-in-training: a focused assessment. Acad Emerg Med. 1995;2:622– 629. 8. Wray NP, Friedland JA. Detection and correction of house staff error in physical diagnosis. JAMA. 1983;249:1035–1037. 9. Goetzl EJ, Cohen P, Downing E, et al. Quality of diagnostic examinations in a university hospital outpatient clinic. Ann Intern Med. 1973;78:481– 489. 10. St. Clair EW, Oddone EZ, Waugh RA, et al. Assessing house staff diagnostic skills using a cardiology patient simulator. Ann Intern Med. 1992;117:751–756. 11. Lembo NJ, Dell’Italia LJ, Crawford MH, O’Rourke RA. Bedside diagnosis of systolic murmurs. N Engl J Med. 1988;318:1572–1578. 12. Roldan CA, Shively BK, Crawford MH. Value of the cardiovascular physical examination for detecting valvular heart disease in asymptomatic subjects. Am J Cardiol. 1996;77:1327–1331. 13. Danford DA, Nasir A, Gumbiner C. Cost assessment for evaluation of heart murmurs in children. Pediatrics. 1993;91:365–368. 14. Mangione S, Nieman LZ. Cardiac auscultatory skills of internal medicine and family practice trainees. JAMA. 1997;278:717–722. 15. Drake D, Fitzgerald S, Jaffe M. Hard Choices: Health Care at What Costs. Kansas City: Andrews McNeel; 1993:80. 16. Mangione S, Nieman LZ, Gracely E, Kaye D. The teaching and practice of cardiac auscultation during internal medicine and cardiology training. Ann Intern Med. 1993;119:47–54. 17. Tamblyn R, Berkson L, Dauphinee WD, et al. Unnecessary prescribing of NSAIDs and the management of NSAID-related gastropathy in medical practice. Ann Intern Med. 1997;127:429 – 438. 18. Davidoff F. Time. Ann Intern Med. 1997;127:483– 485. 19. Howie JG, Porter AM, Heaney DJ, Hopton JL. Long to short consultation ratio: a proxy measure of quality of care for general practice. Br J Gen Pract. 1991;41:48 –54.
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110 215
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione 20. Hampton R, Harrison MJ, Mitchell JR, et al. Relative contributions of history-taking, physical examination, and laboratory investigation to diagnosis and management of medical outpatients. BMJ. 1975;2:486 – 489. 21. Peterson MC, Holbrook JH, Hales DV, et al. Contributions of the history, physical examination, and laboratory investigation in making medical diagnoses. West J Med. 1992;156:163–165. 22. Bisno AL. Acute rheumatic fever: forgotten but not gone. N Engl J Med. 316:476 – 478.
216
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
23. Vijaykumar M, Narula J, Reddy KS, Kaplan EL. Incidence of rheumatic fever and prevalence of rheumatic heart disease in India. Int J Cardiol. 1994;43:221–228. 24. Paauw DS, Wenrich MD, Curtis JR, et al. Ability of primary care physicians to recognize physical findings associated with HIV infection. JAMA. 1995;274:1380 –1382. 25. Berg D, Sebastian J, Heudebert G. Development, implementation, and evaluation of an advanced physical diagnosis course for senior medical students. Acad Med. 1994;69:758 –764.
Volume 110
diac events ranged from 0% to 58% for American trainees (mean 22% ⫾ 12%), 0% to 58% for Canadians (mean 26% ⫾ 13%), and 0% to 42% for British trainees (mean 20% ⫾ 12%). Canadians’ cumulative scores were slightly but significantly greater than those of American (P ⫽ 0.02) and British house officers (P ⫽ 0.05). British house officers improved the most during the 3 years of training (P ⬍0.05). Canadian and British trainees had received more auscultatory teaching during medical school and residency; they had also used audiotapes more frequently (all P ⬍0.001) CONCLUSIONS: Auscultatory proficiency was poor in all three countries. Although there were slight differences among countries, the most striking finding was the consistent inaccuracy of all trainees. This suggests that variables other than teaching and testing affect proficiency. Am J Med. 2001;110: 210 –216. 䉷2001 by Excerpta Medica, Inc.
T
cians and correlate directly with an increased use of subspecialty consultations (9). Errors are particularly common in the field of chest auscultation, a time-honored art and the symbol of bedside diagnostic skills (10). Even though cardiac auscultation, competently performed, remains a valuable (11,12) and cost-effective (13) tool for the evaluation of valvular diseases, generalists-in-training are quite inaccurate in the recognition of common auscultatory events, improve little with a year of training, and are not much better than third-year medical students (14). Whether deficiencies in physical diagnosis may be limited to the United States or whether they may extend to trainees of other countries is not known. Because access to technology, teaching and training practices, and testing of residents at time of board certification vary among different countries, we speculated that auscultatory proficiency might also vary. To test this hypothesis we surveyed internal medicine residents from Canada, England, and the United States. We chose these three countries because all are English-speaking and similar in cultural and medical traditions and yet quite different in organization, emphasis on clinical skills during training, and methods of final board certification examination.
here is increasing evidence that bedside clinical skills have been waning. Only one fourth of US internal medicine and family practice residencies offer any formal instruction in cardiac auscultation (1); fewer programs offer structured teaching in pulmonary auscultation (2); only a minority of residents are closely supervised while performing a physical examination (3); and attending physician rounds are increasingly conducted away from the bedside, often in conference rooms and corridors (4,5). The decreased emphasis during training has probably contributed to the high rates of error in physical examination that have been reported among house staff (6,7). In one study, errors of omission occurred two to three times more frequently than errors of commission, and serious errors leading to major changes in diagnosis and management occurred in almost two thirds of the patients examined (8). High rates of error have also been described among university-affiliated attending physi-
From the Department of Medicine and Center for Research in Medical Education and Health Care, Jefferson Medical College of Thomas Jefferson University, Philadelphia, Pennsylvania. Requests for reprints should be addressed to Salvatore Mangione, MD, Department of Medicine and Center for Research in Medical Education and Health Care, Jefferson Medical College of Thomas Jefferson University, 1025 Walnut Street, Suite 121, Philadelphia, Pennsylvania 19107. Manuscript submitted November 5, 1999, and accepted in revised form August 25, 2000. 210
䉷2001 by Excerpta Medica, Inc. All rights reserved.
METHODS We tested 314 internal medicine residents (189 American, 89 Canadian, and 36 British trainees) in 14 university 0002-9343/01/$–see front matter PII S0002-9343(00)00673-2
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
hospitals. These included 3 London hospitals (Hammersmith, Middlesex, and St. Thomas hospitals), 5 Canadian hospitals (Queens Hospital in Kingston, Ontario, McMaster University Hospital in Hamilton, Ontario, and 3 university hospitals affiliated with McGill University in Montreal, Quebec: National Jewish, Montreal General, and Royal Victoria hospitals), and 6 American hospitals (5 in Philadelphia: T. Jefferson University Hospital, Temple University Hospital, Hospital of the University of Pennsylvania, Medical College of Pennsylvania, and Hahnemann University Hospitals, as well as Georgetown University Hospital in Washington DC). Data on American trainees have been reported in part (14).
Proficiency Test The evaluation of auscultatory proficiency was arranged as a 1-hour patient management conference, incorporated in the teaching activities of the various programs. The evaluation instrument consisted of 12 cardiac events (6 extra-sounds and 6 murmurs), directly recorded from patients, and selected from a pool of more than 250 sounds. All sounds were digitized into a MacIntosh SE 30 computer (Apple Computer Inc., Cupertino, California), their waveforms were analyzed (MacRecorder, Farallon Computing, Berkeley, California), and the final 12 events were chosen based on the purity of their sounds and good reproduction of findings typical of the corresponding disease process. The recognition of these same events had been ranked in our previous survey of family practice and internal medicine program directors as very important for practicing physicians (1). Each event was recorded on a high-fidelity cassette deck (CR W85, Fisher Corp., Chatsworth, California) and played back to trainees for 1.5 minutes using wireless infrared stethophones (Cardionics Inc., Houston, Texas) and a high-fidelity audiocassette player (Marantz PMD430, Chatsworth, California). For each event, participants were told the chest area of recording and had the opportunity to listen again if they wished to do so. After listening, they answered a multiplechoice questionnaire by checking in a forced-response format whether a particular finding was present or absent and by selecting characteristics that best described it. Answers were graded using as a reference the corresponding phonocardiographic tracing. Scores reflected participants’ accuracy in the recognition of a specific event. Item scores were expressed as the percentage of respondents, for year and country of training, who correctly identified each of the sounds tested. Cumulative scores reflected the participants’ accuracy in recognizing all 6 extra-sounds, all 6 murmurs, and all 12 auscultatory events. These scores were expressed as the percentage of total number of events (murmurs, extra-sounds, or all) recognized by each participant. Whenever respondents not only selected the right finding but also reported findings that
were absent (for example, an S4 together with the right finding of an S3 gallop), a “corrected” (ie, reduced) score was calculated. A group of 10 cardiologists attending the annual meeting of the Pennsylvania Chapter of the American College of Cardiology also took our test, and both their performance (accuracy rates ranging between 80% and 90%) and their feedback confirmed the good quality of the test sounds.
Attitudes Survey Participants also completed a one-page attitude questionnaire about self-motivated learning of cardiac auscultation, auscultatory teaching they had received during medical school and internal medicine residency, their opinions of the clinical importance of cardiac auscultation and the need to devote more time to its teaching, and confidence with their own cardiac auscultatory skills.
Statistical Analysis We determined the differences in auscultatory proficiency among trainees of different countries and the degree to which third-year residents have greater skills than first-year residents. Chi-square tests were used to compare trainees of different countries, and the between-year data were analyzed by a trend-analysis chi-square test in the three countries by the 3 years of medical residency. Trainees’ attitudes toward cardiac auscultation were analyzed by Student’s t test and chi-square tests. Relations between trainees’ proficiency and attitudes or practices toward cardiac auscultation were analyzed by Spearman rank correlations. Continuous results are reported as mean ⫾ SD.
RESULTS Proficiency The most striking finding was the overall poor performance of all house-officers, independently of their country of training (Table 1). Yet, Canadian trainees had significantly higher cumulative noncorrected scores than American or British trainees. Compared with Americans, they also had higher identification scores for the 6 murmurs, both corrected and noncorrected. American trainees had higher corrected identification scores for the 6 extra-sounds. There were no significant differences in cumulative scores when comparing first- and third-year residents of either the United States or Canada. Among British trainees, however, third-year residents had significantly higher noncorrected scores (29% ⫾ 9% versus 15% ⫾ 12% for all 12 events, P ⫽ 0.008; and 27% ⫾ 18% versus 12% ⫾ 16% for the 6 extra-sounds, P ⫽ 0.04). Canadians had higher noncorrected identification scores for the murmurs of mitral regurgitation, aortic regurgitation, and mitral stenosis than did Americans, and they also identified the murmur of patent ductus arte-
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110 211
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
Table 1. Scores for the Various Auscultatory Findings among Internal Medicine Residents of Three English-Speaking Countries US Trainees (n ⫽ 189)
UK Trainees (n ⫽ 36) Mean ⫾ SD
Scores Total Noncorrected score† Corrected score Murmurs Noncorrected score Corrected score Sounds Noncorrected score Corrected score PGY1 versus PGY3 Noncorrected (total) score Noncorrected (sounds) score
Canadian Trainees (n ⫽ 89)
22 ⫾ 12 10 ⫾ 9
20 ⫾ 12 10 ⫾ 9
26 ⫾ 13* 12 ⫾ 10
16 ⫾ 15 12 ⫾ 14
22 ⫾ 19 16 ⫾ 16
27 ⫾ 19† 19 ⫾ 17†
28 ⫾ 18‡ 9 ⫾ 13‡
20 ⫾ 18 3⫾8
25 ⫾ 17 5 ⫾ 10
20 ⫾ 10 vs 24 ⫾ 14
15 ⫾ 12 vs 29 ⫾ 9*
22 ⫾ 11 vs 28 ⫾ 14
26 ⫾ 17 vs 31 ⫾ 19
12 ⫾ 16 vs 27 ⫾ 19*
22 ⫾ 14 vs 28 ⫾ 18
* P ⬍0.05 by comparison with other countries or other years of training. † P ⬍0.001 by comparison with other countries or other years of training. PGY ⫽ postgraduate year.
rious significantly more frequently than both American and British trainees. Compared with British trainees, Americans identified more frequently the S4 gallop and less frequently the murmurs of mitral regurgitation and stenosis. When corrected scores were analyzed, Canadian trainees still maintained higher identification rates for the murmurs of mitral regurgitation, aortic regurgitation, and mitral stenosis. Differences between American and British trainees were unchanged (Figure 1). Only the recognition of the S3 gallop and the murmur of mitral stenosis improved significantly among American trainees; the identification of the early ejection sound actually worsened. For Canadian trainees there was significant improvement by year of training only for the opening snap and the murmur of mitral regurgitation (Figure 2). None of these differences persisted when corrected scores were analyzed. Among British trainees, identification rates more than doubled for 7 of the 12 events tested, but only 3 of these 7 improvements approached statistical significance (P ⫽ 0.07).
Attitudes American and British trainees attributed significantly greater importance to cardiac auscultation than Canadians (3.8 ⫾ 0.4 for Americans, 3.7 ⫾ 0.5 for British, and 3.4 ⫾ 0.6 for Canadians on a scale from 1 [obsolete and useless] to 4 [extremely important]; both P ⬎0.001). British and Canadian trainees had both pursued and received significantly more teaching of this skill during medical school and residency training; they had also used
212
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
cardiac audiotapes significantly more frequently than their American counterparts (all P ⬍0.001; Table 2). Yet British trainees had less confidence than Americans in their own auscultatory skills (1.9 ⫾ 0.7 for British trainees, 2.5 ⫾ 0.7 for Americans on a scale from 1 [poor] to 5 [excellent]; P ⬍0.001). British and Canadians trainees wished for more auscultatory teaching during training than did American trainees (P ⫽ 0.05). Trainees who had received structured teaching during residency had greater confidence (on a 1 to 5 scale) in their own auscultatory skills than those who had not been taught (2.7 ⫾ 0.7 versus 2.3 ⫾ 0.7, P ⫽ 0.02). Structured teaching was also associated with greater accuracy (cumulative corrected scores for murmurs of 19% ⫾ 18% versus 12% ⫾ 13%, P ⫽ 0.01; cumulative noncorrected scores of 23% ⫾ 20% versus 16% ⫾ 14%, P ⫽ 0.03). Trainees who had used cardiac audiotapes had no greater confidence in their own auscultatory skills (2.5 ⫾ 0.7 versus 2.4 ⫾ 0.7, P ⫽ 0.3), and they attributed significantly less importance to auscultation (3.8 ⫾ 0.4 versus 3.5 ⫾ 0.5 on a 1 to 4 scale, P ⫽ 0.001). However, they had significantly higher cumulative noncorrected scores (26% ⫾ 12% versus 20% ⫾ 10% for all 12 events, and 24% ⫾ 18% versus 15% ⫾ 14% for the 6 murmurs; both P ⫽ 0.001) Greater self-confidence in auscultation did not correlate with higher identification scores among British and Canadian trainees. Among American trainees, there was a significant correlation between skill self-perception and
Volume 110
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
Figure 1. Accuracy rates are shown by country of training. Corrected scores were calculated whenever the respondents selected not only the right finding but also findings acoustically similar and yet absent; the corrected score considered these types of answers to be invalid. Significance is reported for improvements in each of the three participating groups when compared with each other; asterisks indicate P values ⬍0.05. MR ⫽ mitral regurgitation; AS ⫽ aortic stenosis; AS/AR ⫽ aortic stenosis and regurgitation; AR ⫽ aortic regurgitation; MS ⫽ mitral stenosis rumble; PDA ⫽ patent ductus arteriosus; RUB ⫽ pericardial rub; S4 ⫽ S4 gallop; S3 ⫽ S3 gallop; OS ⫽ opening snap of mitral stenosis; MSCLK ⫽ mid-systolic click; EES ⫽ early ejection sound.
auscultatory accuracy (r ⫽ 0.21, P ⫽ 0.02 for cumulative corrected identification score, and r ⫽ 0.24, P ⫽ 0.006 for the corrected murmur identification score).
DISCUSSION We sought to ascertain whether there were differences in cardiac auscultation skills among trainees of different countries, perhaps as a result of disparities in the assessment of bedside skills at time of board examination. We
could not, however, survey residents who were truly representative of the three countries. Thus, extrapolation of the findings is difficult. Still, within these inevitable limitations, British trainees, who had received more auscultatory teaching and who were expected to undergo an objective assessment of physical examination skills at the time of board certification, improved the most over 3 years of training. Canadian residents, similarly trained and tested, had the greatest accuracy. Yet, although significant, these differences were slight.
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110 213
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
Figure 2. Differences in diagnostic accuracy are shown by country and year of training. Gains (or losses) in diagnostic accuracy between third-year and first-year residents were calculated as the difference between the noncorrected scores. Significance is reported for improvement (or worsening) by year of training; asterisks indicate P values ⬍0.05. MR ⫽ mitral regurgitation; AS ⫽ aortic stenosis; AS/AR ⫽ aortic stenosis and regurgitation; AR ⫽ aortic regurgitation; MS ⫽ mitral stenosis rumble; PDA ⫽ patent ductus arteriosus; RUB ⫽ pericardial rub; S4 ⫽ S4 gallop; S3 ⫽ S3 gallop; OS ⫽ opening snap of mitral stenosis; MSCLK ⫽ mid-systolic click; EES ⫽ early ejection sound.
More importantly, the overall performance of all housestaff officers, independent of their country of training, was poor. This suggests that other variables beyond teaching and testing affect proficiency. One of these variables may be the availability of diagnostic technology. Although advanced technologies may be more widespread in the United States than elsewhere (by 1993, for example, Philadelphia hospitals had 63 magnetic resonance imaging scanners, compared with 57 in all of Germany and 12 in all of Canada [15]), all three countries that we tested share an equally sophisticated and technologically advanced medicine. Easy access to technology may correlate inversely with the time and attention de-
voted to physical diagnosis during training. In a nationwide survey of US internal medicine residencies, for example, we found that small community-hospitals, presumably less equipped with technologic resources, provided significantly more cardiac auscultatory teaching than large university-affiliated tertiary care centers (16). Unlimited access to technology, combined with its glamour and sophistication, may give trainees an easy alternative to painstaking and time-consuming bedside examinations. The issue of time in medical care is very important (17,18). As the time for physician-patient interactions continues to decrease—the mean visit length
Table 2. Teaching of Cardiac Auscultation Received during Medical School and Residency, and Attitudes of Participating Residents toward Auscultation US Trainees Characteristic or Attitude Teaching received During medical school During residency Independent use of audiotapes More teaching needed in medical school More teaching needed in residency
214
February 15, 2001
UK Trainees
Canadian Trainees
P Value
Numerator/Denominator (Percent) 93/137 (68) 19/121 (16) 43/121 (36) 119/126 (94) 111/119 (93)
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110
27/27 (100) 13/13 (100) 7/7 (100) 22/22 (100) 23/23 (100)
58/60 (97) 25/29 (86) 36/39 (92) 64/64 (100) 67/67 (100)
⬍0.001 ⬍0.001 ⬍0.001 0.09 ⬍0.05
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione
was 7 to 9 minutes in a 1991 study (19)– better skills at history-taking and physical diagnosis have become more crucial for reducing the cost of care without decreasing its quality (13). These fundamental crafts alone can still deliver a diagnosis for a high proportion of patients, at least in the outpatient setting, where most care occurs (20,21). Another factor that might have contributed to the trainees’ inaccuracy in this study is the changing prevalence of valvular diseases in the Western world. Rheumatic fever, for example, has almost completely, but not entirely (22), disappeared in technologically advanced societies. Conversely, it remains quite prevalent in developing countries (23) where access to technology is limited and strong bedside skills, such as physical diagnosis, are still an essential part of medical students’ training and of patient’s evaluation. It would have been interesting to compare the auscultatory proficiency of trainees from Western world countries with that of trainees from developing nations. A possible limitation of our study is that we used recorded sounds and did not provide any medical history. Physicians’ ability to recognize taped sounds may correlate poorly with bedside skills, and clinical history provides important guidance. Our data, however, confirm other observations. Paauw et al (24) assessed the physical examination skills of 134 primary care providers using real patients, all of whom had physical findings. Even though physicians’ attention was directed to sites of patients’ lesions, their performance remained poor. Only 38% of participants recognized two murmurs of aortic insufficiency, only 17% detected diffuse lymphadenopathy, and only about half noticed any abnormality in patients with Kaposi’s sarcoma or oral hairy leukoplakia. Innovative educational methods may be necessary to rejuvenate the art of physical diagnosis. However, teaching and testing may not suffice if widespread use of diagnostic technology continues to preempt efforts at the bedside. We believe that the best way to revive skills such as auscultation would be to have a cadre of clinicianteachers who could serve as role models for physiciansin-training, revive the fascination with a Sherlockian approach to bedside diagnosis, and show how to use technology more judiciously and cost effectively. There are already indications that this strategy can be highly successful with students (25). After all, it was a brilliantly deductive and charismatic professor, Dr. Joseph Bell, who captured the imagination of young Arthur Conan Doyle and inspired the creation of Sherlock Holmes. Rekindling interest and competence in physical diagnosis will therefore require a substantial effort at the bedside as well as a considerable shift in attitudes. This shift may include readdressing issues of physicians’ liability, patients’ demands, and promotion and incentives for clinical faculty willing to invest time and effort in teach-
ing these skills. The challenge to educators is whether such investments are warranted in times of increasing technologic sophistication. To this end, the value of physical diagnosis will need to be redefined and supported continuously by ascertaining which elements remain important in the context of contemporary medicine.
ACKNOWLEDGMENTS The author wishes to express his thanks to Dr. Graham Leech of London and to Messrs. Keith Johnson and David Gibbons for their help with the English trainees’ testing.
REFERENCES 1. Mangione S. The teaching of cardiac auscultation during internal medicine and family medicine training: a nationwide comparison. Acad Med. 1998;73:S10 –12. 2. Mangione S, Loudon RG, Fiel SB. Lung auscultation during internal medicine and pulmonary training: a nationwide survey. Chest. 1993;104:70S. 3. Burdick WP, Schoffstall J. Observation of emergency medicine residents at the bedside: how often does it happen? Acad Emerg Med. 1995;2:909 –913. 4. Shankel SW, Mazzaferri EL. Teaching the resident in internal medicine: present practices and suggestions for the future. JAMA. 1986;256:725–729. 5. Collins GF, Cassie JM, Daggett CJ. The role of the attending physician in clinical training. J Med Educ. 1978;53:429 – 431. 6. Wiener S, Nathanson M. Physical examination: frequently observed errors. JAMA. 1976;236:852– 855. 7. Mangione S, Burdick WP, Peitzman S. Physical diagnosis skills of physicians-in-training: a focused assessment. Acad Emerg Med. 1995;2:622– 629. 8. Wray NP, Friedland JA. Detection and correction of house staff error in physical diagnosis. JAMA. 1983;249:1035–1037. 9. Goetzl EJ, Cohen P, Downing E, et al. Quality of diagnostic examinations in a university hospital outpatient clinic. Ann Intern Med. 1973;78:481– 489. 10. St. Clair EW, Oddone EZ, Waugh RA, et al. Assessing house staff diagnostic skills using a cardiology patient simulator. Ann Intern Med. 1992;117:751–756. 11. Lembo NJ, Dell’Italia LJ, Crawford MH, O’Rourke RA. Bedside diagnosis of systolic murmurs. N Engl J Med. 1988;318:1572–1578. 12. Roldan CA, Shively BK, Crawford MH. Value of the cardiovascular physical examination for detecting valvular heart disease in asymptomatic subjects. Am J Cardiol. 1996;77:1327–1331. 13. Danford DA, Nasir A, Gumbiner C. Cost assessment for evaluation of heart murmurs in children. Pediatrics. 1993;91:365–368. 14. Mangione S, Nieman LZ. Cardiac auscultatory skills of internal medicine and family practice trainees. JAMA. 1997;278:717–722. 15. Drake D, Fitzgerald S, Jaffe M. Hard Choices: Health Care at What Costs. Kansas City: Andrews McNeel; 1993:80. 16. Mangione S, Nieman LZ, Gracely E, Kaye D. The teaching and practice of cardiac auscultation during internal medicine and cardiology training. Ann Intern Med. 1993;119:47–54. 17. Tamblyn R, Berkson L, Dauphinee WD, et al. Unnecessary prescribing of NSAIDs and the management of NSAID-related gastropathy in medical practice. Ann Intern Med. 1997;127:429 – 438. 18. Davidoff F. Time. Ann Intern Med. 1997;127:483– 485. 19. Howie JG, Porter AM, Heaney DJ, Hopton JL. Long to short consultation ratio: a proxy measure of quality of care for general practice. Br J Gen Pract. 1991;41:48 –54.
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 110 215
Cardiac Auscultatory Skills of Physicians-in-Training/Mangione 20. Hampton R, Harrison MJ, Mitchell JR, et al. Relative contributions of history-taking, physical examination, and laboratory investigation to diagnosis and management of medical outpatients. BMJ. 1975;2:486 – 489. 21. Peterson MC, Holbrook JH, Hales DV, et al. Contributions of the history, physical examination, and laboratory investigation in making medical diagnoses. West J Med. 1992;156:163–165. 22. Bisno AL. Acute rheumatic fever: forgotten but not gone. N Engl J Med. 316:476 – 478.
216
February 15, 2001
THE AMERICAN JOURNAL OF MEDICINE威
23. Vijaykumar M, Narula J, Reddy KS, Kaplan EL. Incidence of rheumatic fever and prevalence of rheumatic heart disease in India. Int J Cardiol. 1994;43:221–228. 24. Paauw DS, Wenrich MD, Curtis JR, et al. Ability of primary care physicians to recognize physical findings associated with HIV infection. JAMA. 1995;274:1380 –1382. 25. Berg D, Sebastian J, Heudebert G. Development, implementation, and evaluation of an advanced physical diagnosis course for senior medical students. Acad Med. 1994;69:758 –764.
Volume 110