Effective patient care requires not only a working knowledge of recommended tests and therapies but also an understanding of the frequency of harms and benefits for each. To make educated decisions, patients must understand harms and benefits of treatments. Unfortunately, patients consistently overestimate benefits and underestimate harms of medical tests and procedures.1 Likewise, physicians are poor at assessing treatment effect size and other aspects of numeracy. Some have hypothesized that clinicians, similar to patients, overestimate risks and underestimate harms.2 We evaluated physician understanding of harms and benefits of common tests and therapies.
We administered a voluntary paper survey to resident and attending internal medicine physicians at 2 academic medical centers. The survey contained 18 questions, of which 10 evaluated understanding of rates of benefits and harms of common medical interventions. The remaining questions assessed confidence in responses, use of statistical terms, and awareness of high-value care campaigns. The questions and answer scale were based on previous questionnaires administered to patients to compare physician with patient knowledge.3 This study was approved by the institutional review board of the University of Maryland and the Department of Veterans Affairs–Baltimore research and development committee; written informed consent was waived because this was an anonymous survey. Participants were not compensated
Survey results were compiled, and the frequency of each answer was analyzed for 18 questions, stratified by resident vs attending physician. Differences were tested using χ2 and Fisher exact tests using SAS statistical software (version 9.3; SAS Institute).
From November 1, 2015, to March 31, 2016, 132 eligible individuals were approached, with 117 completing the survey (response rate, 88.6%). Reasons for declining to participate included inadequate time (46.7%), stating they would complete it later (46.7%), and no reason given (6.6%). Participants included 41 postgraduate year (PGY) 1 residents (35.0%), 18 PGY 2 residents (15.4%), 21 PGY 3 to 5 residents (18.0%), and 37 attending physicians (31.6%). Of attending physicians, 9 (24.3%) were primary care; 14 (37.8%), hospitalists; and 14 (37.8%), subspecialists.
Responses to the 10 questions pertaining to harms and benefits of medical treatments are displayed in the Figure. Most participants overestimated benefits (92 of 117 [78.6%] average across 7 benefit questions) and harms (77 of 117 [65.8%] average across 3 harm questions) of common medical tests and treatments. There was no difference in the number of correct responses between residents and attending physicians. Most respondents (53 [67.5%]) were unconfident, choosing 4 or below on a scale of 1 to 10, for which 1 was not at all confident and 10 very confident. Physician numeracy is presented in the Table.
We found that most clinicians overestimate harms and benefits for most treatments. Likewise, most of the clinicians in our study reported rarely or never using statistical terms to explain treatment options to patients. However, they were interested in resources to improve understanding of treatment effect size.
Our study results add to previous studies demonstrating patients overestimated benefits and underestimated harms.1 Physicians underestimate how often most treatments have no effects on patients—either harmful or beneficial. These physician beliefs likely contribute to the same misconceptions patients have for the benefits of treatments and may lead to medical overuse.2 Even when clinicians understand numeracy, expressing these terms in a way patients will understand is challenging.4 Development of more readily accessible decision-making aids may improve both clinician and patient understanding.5
Interestingly, our study did not demonstrate any difference between the accuracy of resident and attending physicians, although attending physicians were more aware of current high-value care campaigns. Poor ability to estimate frequency of benefit likely relates to medical education focusing only on treatments that are recommended, with a heavy emphasis on P values alone, rarely using more accurate assessments of treatment effect size or frequency with absolute risk reduction or number needed to treat.
Corresponding Author: Daniel J. Morgan, MD, MS, Department of Epidemiology and Public Health, University of Maryland School of Medicine, 685 W Baltimore St, MSTF 334, Baltimore, MD 21201 (firstname.lastname@example.org).
Published Online: August 29, 2016. doi:10.1001/jamainternmed.2016.5027.
Author Contributions: Dr Morgan had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Krouss, Morgan.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Krouss, Morgan.
Statistical analysis: Croft, Morgan.
Obtained funding: Morgan.
Administrative, technical, or material support: Krouss, Morgan.
Study supervision: Morgan.
Conflict of Interest Disclosures: Dr Morgan has received editorial honoraria from Springer for serving as a book and journal editor. No other disclosures are reported.
Funding/Support: Dr Morgan is supported by the Department of Veterans Affairs’ Health Services Research and Development Service and the Agency for Healthcare Research and Quality.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
TC, Del Mar
C. Patients’ expectations of the benefits and harms of treatments, screening, and tests: a systematic review. JAMA Intern Med
. 2015;175(2):274-286.PubMedGoogle ScholarCrossref
D. Patient perception of benefits and harms: the Achilles heel of high-value care. JAMA Intern Med
. 2015;175(2):287-288.PubMedGoogle ScholarCrossref
BJ. Patients’ knowledge about 9 common health conditions: the DECISIONS survey. Med Decis Making
. 2010;30(5)(suppl):35S-52S.PubMedGoogle ScholarCrossref
et al. Evidence-based risk communication: a systematic review. Ann Intern Med
. 2014;161(4):270-280.PubMedGoogle ScholarCrossref
LC. Tools to promote shared decision making in serious illness: a systematic review. JAMA Intern Med
. 2015;175(7):1213-1221.PubMedGoogle ScholarCrossref