- Email: [email protected]
Housestaff and Medical Student Attitudes Toward Medical Errors and Adverse Events
The Joint Commission Journal on Quality and Patient Safety
Health Professions Education
Housestaff and Medical Student Attitudes Toward Medical Errors and Adverse Events
Pamela D. Vohra, M.D. Julie K. Johnson, M.S.P.H., Ph.D. Christopher K. Daugherty, M.D. Ming Wen, Ph.D. Paul Barach, M.D., M.P.H.
T
he Institute of Medicine proposed that quality improvement and patient safety practices should be a formal part of medical training and that an emphasis on safety would improve the quality of medical treatment, enhance the public’s view of medicine, and decrease the costs of health care.1 Its 2001 report, Crossing the Quality Chasm, broadly addressed the need for improving professional education with recommendations that “colleges of medicine . . . should build more instruction into their curriculum on patient safety and its relationship to quality improvement.”2(p. 19) Concerns about medical errors that occur during physician training (residency) programs have directly resulted in significant changes in medical training (for example, an 80-hour work week, admission caps, and strict housestaff supervision requirements). Traditionally, adverse events were thought to be an unavoidable outcome of learning to practice medicine,3 and medical errors by trainees were viewed as teaching opportunities. The most frequent methods of dealing with errors, such as morbidity and mortality conferences, were punitive4—an approach that, from an educational point of view, may have a negative effect on learning.5 There is a growing acceptance that systemic problems, rather than poor individual performance, underlie most adverse events in health care.6 A culture of safety, as practiced, for example, in aviation and the nuclear power industry, which encourages reporting about and learning from all errors,7,8 is increasingly viewed as the most effecAugust 2007
Article-at-a-Glance Background: A lack of formal patient safety curricula has contributed to the suboptimal training of medical students and housestaff. Attitudes of physician trainees regarding medical errors and adverse events were surveyed in a pilot study. Methods: Five hundred sixty-three physician trainees were surveyed at an urban teaching hospital. Five domains were evaluated using a factor analysis as they relate to patient safety: knowledge, self-efficacy, awareness of safety culture, barriers/facilitators, and awareness of human factors. Results: One hundred fifty-eight (28%) trainees completed the survey, with 22% (n = 35) describing exposure to at least one adverse medical event. The survey showed good internal validity and reliability. Respondents who reported exposure to adverse events demonstrated a lower awareness of human factor errors (p = .0017) and lower awareness of the hospital’s approach to safety (p = .033). Older respondents scored higher on measures of self-efficacy than younger trainees (p = .042). Discussion: The exposure of physician trainees to errors and adverse events can have a negative effect on their attitudes and competencies. Exposure to adverse events and the institution’s response may decrease both error reporting and the willingness to adopt safety practices. The results support the need for implementing a sustained patient safety curriculum that promotes learning regarding adverse events.
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
493
The Joint Commission Journal on Quality and Patient Safety
tive means to enhance the reliability of care and reduce the recurrence of medical errors.9 Researchers have found that house officers who accepted and learned from their mistakes are more likely to make 10 constructive changes in their practice. Such acceptance, however, also adds a significant burden of emotional distress, and many residents became uncomfortable sharing with their attending physicians errors and near misses that they have observed or committed. This observation has strengthened the move away from a culture of blame to anonymous error reporting, systems analysis of adverse events, and collective learning opportunities.11,12 Some residency teaching departments (for example, emergency medicine13) have begun to establish a blame-free, errorbased curriculum. It has been proposed that if errors caused by systemic causes such as overwork and fatigue were addressed, house officers could focus on how their own actions contribute to adverse events.11 However, to date, housestaff have received little instruction in human factors, patient safety, continuous quality improvement, and systems thinking.12,14 Furthermore, evidence suggests that housestaff are inadequately prepared for identifying and overcoming the 5,15 organizational barriers to improving patient safety. Wu et al. examined how house officers learn from their own mistakes but not how they learn from the mistakes of others.12 Previous research has not addressed trainees’ perceptions of their self-efficacy and role in reducing medical error and patient harm. Nor has it addressed the impact of trainees’ exposure to institutional responses to adverse events. This article reports on a pilot study, undertaken at a large tertiary care medical center, which had the following two goals: 1. To describe the knowledge, self-efficacy, and attitudes of physician trainees toward adverse events 2. To explore the impact of adverse events and the local political and educational environment on the attitudes of physician trainees
at http://umdas.med.miami.edu/links/sentinelevents/— stated the following: Adverse events sometimes fall under the category of sentinel events, defined by the Joint Commission on Accreditation of Healthcare Organizations as an unexpected occurrence involving death or serious physical or psychological injury, or the events are called ‘sentinel’ because they signal the need for immediate investigation and response.* The orientation to the survey continued as follows: This study aims to examine how housestaff and medical students assess adverse events of health care at the academic hospitals. The purpose of this survey is to obtain your thoughts and perspectives regarding three basic areas: (1) Awareness of the need to reduce medical errors and improve patient safety, and the perceived barriers to awareness (2) Assessment of your role in reducing medical errors and improving patient safety, and the impact of adverse event analysis meetings (also known as root cause analysis [RCA] meetings) and outcomes on your daily practice and on patient care (3) Potential opportunities that should be considered in your department for reducing medical errors and improving patient safety
STUDY POPULATION
ORIENTATION TO THE SURVEY
After full Institutional Review Board approval, 453 medical residents and 100 senior medical students (physician trainees) at an urban teaching hospital were invited in 2004 to participate in the study. Subjects were recruited via e-mails sent by their program directors or by the medical student listserve network. Reminder e-mails to nonrespondents were sent out three months after the first invitation. The respondents anonymously completed electronic questionnaires on a secure Web site16 and provided informed online consent to participate in the study. Each completed survey was given a random computer-generated code. All identifiers were removed from the data before analysis. Respondents were compensated with an online bookstore gift certificate worth $5 (allocated by a random
The survey, “Sentinel Events at the Academic Hospitals: Evaluation of Housestaff and Medical Student Attitudes toward Adverse Medical Events”—a sample may be found
* The Joint Commission: Sentinel Event Policy and Procedures. Updated Oct. 2006. http://www.jointcommission.org/SentinelEvents/Policyand Procedures/se_pp.htm (last accessed Jun. 14, 2007).
Methods
494
August 2007
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety
number generator so that investigators could not track respondents). Respondents who indicated in the survey that they had participated in a sentinel event and a subsequent RCA meeting to analyze the incident were invited to discuss their experiences in a confidential, semistructured interview. These subjects were contacted about the interview process via e-mail. All interviewees gave written consent and were compensated with a bookstore gift certificate worth $20.
SURVEY The survey was developed and refined based on an extensive literature review. Most of the questions were adapted from previously validated tools16,17 and underwent extensive assessment by the authors. A draft of the survey was pilot-tested on five subjects and subsequently refined. The survey explored the following three areas: 1. Awareness (knowledge) of the need to reduce medical errors and barriers to improving patient safety (barrier analysis) 2. Assessment of the trainee’s role in reducing medical errors and improving patient safety 3. Assessment of the impact of institutional sentinel event RCA meetings and their outcomes on daily practice of patient care The survey used a 5-point Likert scale (“strongly disagree” to “strongly agree”) categorical response format and included several open-ended questions.
SCALE CONSTRUCTION We used an initial affinity sort of the responses, followed by a modified Delphi process to reach consensus in identifying five domains of housestaff and medical student attitudes toward medical errors and adverse events. These domains emerged from review of the literature and from assessing and analyzing the data at the pilot phase. They were then used to measure the attitudes towards patient safety. The five domains were as follows: 1. Knowledge: awareness of the need for improvement in patient safety 2. Self-efficacy: belief in one’s ability to reduce errors and improve patient safety 3. Awareness of safety culture: error-reporting behaviors (such as willingness to discuss or take responsibility for errors) and willingness to adopt new safety practices
August 2007
4. Beliefs about barriers/facilitators: attitudes about systemic barriers to improvement in patient safety 5. Awareness of human factors: perception and knowledge of how education/training, communication, and other organizational factors affect patient safety. The items and their respective domains are shown in Table 1 (page 496). The five scales were constructed using orthogonal principal factor analysis, a data reduction technique concerned with finding a small number of common factors that linearly reconstruct the original variables.18 The factor loadings reflect the strength of the correlation between the items and the latent common factors. Table 2 (page 497) presents the ranges of factor scores and internal consistency and reliability coefficients (Cronbach’s ␣) for the five scales corresponding to the five domains. These scales are weighted on the basis of each item’s factor loadings. The use of aggregate scales allowed for the fit of parsimonious and efficient regression models to identify predictors of attitudes toward patient safety.
INTERVIEWS The confidential interview, which consisted of semistructured questions designed to be completed in 45 minutes, explored residents’ attitudes toward adverse events, their role in assessing the incident, and the institutional response. The seven interviews were recorded and transcribed verbatim; tapes were subsequently erased. The identifiers on the transcripts were removed before analysis. The interviews were coded [J.K.J., P.B.].
STATISTICAL ANALYSIS Alpha coefficients of each of the five scales were computed to measure the scales’ internal reliability. Pair-wise Pearson correlation analyses were performed for the five scales. These scales were used as the dependent variables in multiple ordinary least squares (OLS) regression analyses, which were conducted to assess the effects of age, gender, department, level and length of postgraduate training, and previous involvement in a sentinel event and a related RCA meeting on each of the five domains.
Results CHARACTERISTICS OF RESPONDENTS Of the 463 medical residents and 100 medical students
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
495
The Joint Commission Journal on Quality and Patient Safety
Table 1. Factor Analysis of the Safety Domains*
Number Agreed
Percentage Agreed
Factor Loading
Knowledge Significant efforts are needed to reduce errors and improve patent safety in my immediate professional practice/work area.
87
76
0.90
Significant efforts are needed to reduce errors and improve patent safety throughout this hospital.
97
85
0.87
In general, the residency program directors with whom I regularly work believe that significant efforts are needed to reduce errors and improve patient safety in our immediate professional practice/work areas.
74
65
0.69
Self-Efficacy I have a significant role in the reduction of errors and improvement of patient safety in this institution.
70
61
0.64
I am encouraged by department administrators and supervisors to reduce errors and improve patient safety.
68
60
0.80
When errors do occur in my professional setting, I am asked to help identify sources of error in this incident and ways of correcting for these errors.
57
50
0.78
When errors do occur in my professional setting, I am able to adopt new practices and procedural changes that may reduce error and improve patient safety.
67
59
0.69
Safety Culture If I was directly involved in a medical error, I would document it in the patient’s case record.
110
96
0.87
If I was directly involved in a medical error, I would tell the patient and/or family that I made the mistake.
99
87
0.74
If I was directly involved in a medical error, I would inform the patient’s attending physician.
109
96
0.73
Barriers/Facilitators In my immediate professional setting I feel that communications between patient care team members is sufficient to ensure safe patient care.
38
33
0.80
In my immediate professional setting I feel that technology is adequately used to promote safe patient care.
26
23
0.73
In my immediate professional setting I feel that there are adequate staffing resources to minimize errors and maintain an acceptable level of patient safety.
24
21
0.70
Human Factors When errors do occur in my professional setting, I am informed of their occurrence by colleagues, department administrators, or supervisors.
65
57
0.74
When errors do occur in my professional setting, I am given adequate information about the causes of the error and ways of improving patient safety.
46
40
0.86
When errors do occur in my professional setting, I am notified of procedural changes implemented to reduce errors and improve patient safety.
42
37
0.77
Factor
* Out of 114 responses.
496
August 2007
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety
Table 2. Internal Consistency and Reliability Coefficients of Five Scales* Scale
No. of Items
Cronbach’s alpha
Factor Loading Range
Knowledge
3
.76
.69–.90
Self-Efficacy
4
.71
.64–.80
Barriers/Facilitators
3
.60
.70–.80
Safety Culture
3
.68
.73–.87
Human Factors
3
.70
.74–.86
* Factor loading ranges and Cronbach alpha values.
Table 3. Descriptive Statistics and Factor Loadings of the Scales Corresponding to the Five Domains Independent Variables
n
%
Knowledge
Self-Efficacy
114
100
1.76
1.42
-1.24
3.12
-4.44
Male
60
53
-.14
-.01
-.05
.03
-.01
Female
54
47
.15
.02
.11
-.22
.03
Medical Student
37
32
.09
-.23
-.20
-.09
-.13
Resident
77
68
-.04
.11
.96
.05
.06
Overall
Safety Culture Barriers
Human Factors
Gender
Position
Department (Residents)* Surgery
17
15
-.06
.20
.05
-.31
-.09
Pediatrics
23
20
.21
.21
.08
-.03
.21
Internal Medicine
12
11
-.02
.19
.26
-.04
.05
Exposure
35
31
0.10
0.05
0.15
-0.39
.03
None
79
69
-0.04
-0.02
-0.07
0.017
-.01
Sentinel Event
* In addition, six residents (5%) were from the ophthalmology department, five (4%) were from the emergency medicine department, and four (4%) were from other departments.
invited to participate, 158 (28%) medical residents and medical students responded (44 responses were excluded due to incomplete responses). The study population included 114 medical residents and medical students, combined because of the small numbers. The median age of respondents was 28 years (range, 24–45), and median postgraduate medical experience was 3 years (range, 1–10). As shown in Table 3 (above), medical students made up 32% (n = 37) of the respondents, and 68% (n = 77) were residents. Thirty one percent of the respondents (45% of responding residents) reported being exposed to an adverse event, and 17% reported participating in an
August 2007
RCA meeting. Of the respondents who reported exposure to an adverse event, 22% were unaware of the hospital’s counseling services, and 19% of all respondents did not know how to report errors using the hospital reporting system. A few respondents reported having seen so many errors that it was difficult to choose which one to describe.
SURVEY PSYCHOMETRICS Following the completion of the exploratory factor analysis and elimination of items with low factor loadings and/or substantial “cross-factor” loadings, a final factor structure was produced. The items, descriptive statistics,
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
497
The Joint Commission Journal on Quality and Patient Safety
Table 4. Attributed Causes of Adverse Events (Number and Percentage of Responses)* Fatigue
18 (44%)
Inexperience
17 (41%)
Faulty judgment
17 (41%)
Hesitancy Other
5 (12%) 23 (56%)
* Percentages add to more than 100% as respondents could choose several causes.
and factor loadings are shown in Table 1 (page 496). For example, the knowledge scale consists of three items, with factor loadings ranging from 0.69 to 0.90; the Cronbach’s alpha coefficient is equal 0.76. The higher the score, the more knowledgeable the trainee. Similarly, the higher the self-efficacy score, the greater the trainee’s self-efficacy; the higher the barriers/facilitators score, the greater the understanding and impact of barriers to patient safety; the higher the safety culture score, the more a culture of safety is needed; the higher the human factors score, the less knowledge of human factors the trainees had. This analysis suggests that the scales have reasonably good internal reliability.
reflected in the scales. The length of medical training was slightly (that is, nonsignificantly) negatively correlated (b = -.14, p = .10) with willingness to adopt new safety measures (safety culture). Exposure to a sentinel event was negatively associated with awareness of safety culture (b = -.55, p = .012). Further analyses of specific items in the barriers/facilitators scale indicated that respondents who described exposure to a sentinel event had lower scores on questions pertaining to adequacy of staffing resources (p = .01) and higher scores related to the impact of technology on patient outcomes (p = .04). We found no significant differences between the responses of medical students and residents.
INTERVIEW RESULTS The seven interview respondents were asked to list the underlying causes of the adverse events that they reported (Table 4, left). The causes reported included, in descending rank order, fatigue, inexperience, faulty judgment, and hesitancy. Selected responses to the open-ended interview questions are presented in Table 5 (page 499) according to the five factor domains of the questionnaire tool. These observations support the concerns represented by the items in the questionnaire’s five domains (Table 1).
SURVEY RESULTS
Discussion
The self-efficacy domain was correlated significantly and positively with three other domains: human factors (r = .78, p < .01), safety culture (r = .26, p < .01), and barriers/facilitators (r = .19, p = .04). The human factors domain was positively correlated with the barriers factor (r = .33, p < .01), and the safety culture (r = .24, p < .01). The knowledge domain was significantly and negatively correlated with safety culture (r = -.33, p < .01), suggesting that safety culture is independent from safety knowledge and requires a detailed training to acquire. As shown in Table 3 (page 497), the factor loadings of the scores are stratified by gender, years of postgraduate training (position), department, and experiences with sentinel events and related RCA meetings (exposure). Results from the regression analyses show that older respondents on average had a higher self-efficacy score than younger trainees (t = 2.24; p = .01); no age differences were shown for the four other domains. Department and positions were not
Physicians in training working in a high-stress environment are highly influenced by the attitudes and perceptions of their faculty and teaching organizations toward patient safety. The data presented in this study support the reliability and face validity of the survey developed to assess the attitudes of housestaff and medical students toward patient safety. The study, the first to report on the impact of medical errors and adverse events on medical students and residents (trainees), reinforces our knowledge about the ubiquitous nature of human error and patient harm in health care. It points to numerous opportunities for error identification and harm mitigation in these highrisk settings. (Although a greater proportion of students than residents responded to the survey, we found no significant differences between their respective responses.) Medical trainees exposed to medical errors were likely to report a significantly negative attitude toward patient safety. Furthermore, those who reported exposure to med-
498
August 2007
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety
Table 5. Selected Responses to Open-Ended Interview Questions in Five Domains Knowledge ■ “There have been so many [errors], but here is one.” ■ “I’ve written many wrong doses of medications and rely on the pharmacy to double-check me.” Self-Efficacy ■ “If I had seen him sooner, or at least notified the chief resident sooner, we could have gotten him to the operating room sooner for repair of his [aneurysm] leak, and he may have survived.” Safety Culture ■ “And he told me to pay more attention to the patient. Yes, I made the mistake, but hands-down I still and always did know that patient better than he did.” Barriers ■ “Our hospital system is so defunct; most residents find little reason to report errors.” ■ “Most errors which I have been involved with have been related to inadequate staffing, and systems inadequacies so that drugs were not administered in a timely fashion, or critical nursing functions were not performed.” Human Factors ■ “The equipment was not designed with residents in mind.”
ical errors and patient harm had a negative learning experience, that is, they learned little from the experience and also learned to not discuss or share the experience with others. Trainees who were exposed to an adverse event and its subsequent analysis responded negatively to survey issues surrounding the adequate use of technology, staffing resources, and sufficient training. These results suggest that beyond a lack of awareness and knowledge about important factors such as the role of technology in preventing patient harm, the trainees were quick to internalize a personal and nonsystems approach towards errors. This behavior was likely reinforced by explicit and implicit actions and cues that they internalized from their senior colleagues and others during analysis of these events. The positive correlation between self-efficacy and beliefs about barriers/facilitators or awareness of safety culture suggests that trainees’ confidence in addressing errors and adverse patient events affects their learning from these events. Trainees who have a higher confidence in their ability to improve patient safety may also have a more real-
August 2007
istic view of the system’s safety constraints and enablers than trainees with less confidence. One could hypothesize that this may lead to an increased awareness of the challenges to advancing patient safety in academic health centers, which, in turn, could encourage these trainees to be active change agents. Increased respondent age, but not additional years of training (medical student versus resident), was associated with an increased reported self-efficacy and confidence to deal with adverse events. We believe that this gap between age and training duration further underscores the need for a formal iterative safety curriculum across the continuum of medical education, from undergraduate through graduate medical training. Trainees exposed to medical errors and patient harm reported less awareness of hospital efforts to prevent harms and less confidence in the hospital systems than their peers who did not report such exposure. The comments provided in the interviews provide important additional insights and strongly support the survey’s scaled questions. For example, beliefs regarding hospital care, the provider’s work overload, limited support from program administrators and chief residents, and a punitive culture are concerning. These findings suggest that if impressionable physician trainees find their exposure to medical errors to be unpleasant or blameful, they may be less likely to engage in subsequent error-reporting behaviors (such as discussing or taking responsibility for errors) or to adopt new safety interventions and practices. This speaks to the importance of creating a nonpunitive setting for identifying and addressing adverse patient care early in medical training. We believe that the current culture in academic health centers, in which unclear accountability standards are promulgated, may discourage trainees from engaging and committing their energies to improve patient safety. Similarly, Wu et al.10 found that residents were less likely to make constructive changes if they attributed a mistake to job overload or if they felt that the institution was judgmental. At a minimum, our data suggest that trainees’ limited involvement in designing and implementing patient safety interventions reduces the opportunities for them to become effective change agents—or more worrisome, to even learn from their and others’ mistakes. We recognize several limitations to the study. Focusing on clinician perceptions is reliant on self-reports of current
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
499
The Joint Commission Journal on Quality and Patient Safety
and retrospective perceptions, which may be a source of richness but also bias. These events could not be verified, although many of the described events did come to our attention independently. Therefore, we can discuss only an association between perceived institutional responses and the attitudes reported by trainees. As a result, we do not have the ability to infer causality. Second, we cannot tell if there are significant differences between the responders and the nonresponders to the survey. Because of the anonymity of the subjects, nonrespondents could not be contacted (required by the Institutional Review Board; we chose an anonymous survey format to preserve the integrity of the responses and to encourage participation). Anonymous sampling is likely to result in improved veracity, and under some circumstances, improved generalizability of the findings, even if it can result in a smaller sample. Third, the overall response rate to our questionnaire was low, perhaps in part because of the use of an e-mail survey. In addition, medical students and residents have been difficult to survey in general, using e-mail in particular. It is possible that the perceived punitive culture in academic health centers may actually contribute further to the low response rates for a survey on patient safety (several comments in the open-ended interviews support this finding). The sensitive nature of this topic likely discouraged some participants from responding despite elaborate efforts to protect their anonymity. Even if the low response rate reflects a negative response bias, as suggested by Wu et al.,10 the findings may still carry strong validity. Fourth, the data represent only one teaching hospital, which limits the generalizability of our findings. Early results from a similar study at another large academic medical center support the findings of this study.19 Further development of the survey and other instruments to measure behavior and attitudes related to experiences regarding medical errors and adverse events is needed.
Conclusions The findings suggest that efforts to better incorporate a patient safety culture and enhance patient safety training are clearly needed. These findings also highlight the personal and professional burden that the ever-present recognition of errors and fear of these errors represent for physician trainees. There is a need for well-designed cur500
August 2007
ricula within medical schools and residency programs to help trainees learn from their experiences with adverse events in a constructive and nonpunitive manner.20 Promoting self-awareness and assessment in future physicians is key to lifelong learning and development. Trainees should be taught how to report errors and adverse events in a safe way and to seek counseling when adverse events occur. If trainees feel safe, they are more likely to report such events. Such curricula would be consistent with the efforts of the Accreditation Council for Graduate Medical Education and the Association of American Medical Colleges to make patient safety central to undergraduate and graduate medical education.21,22 Senior faculty need to appreciate the importance of sharing their experience with and reflections about medical errors (their own and those of others) and how they have learned to make sense of them. Learning to acknowledge and report adverse patient care should help to foster truth telling about clinical care and instill a more professional and reflective approach toward adverse care in these future physicians. J The research described in this article was supported by a Pritzker Summer Research Program Grant. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the Universities of Chicago, Utah, or Miami. The authors thank the medical students and housestaff who participated in the study; David Faber, who designed the questionnaire’s Web site and assisted in the data collection; and Dr. Richard Frankel for his editorial help.
Pamela D. Vohra, M.D., is a Second-Year Resident, Department of Medicine, Beth Israel Deaconess Medical Center, Boston. Julie K. Johnson, M.S.P.H., Ph.D., is Assistant Professor, Department of Medicine, University of Chicago, Chicago, and Director of Research, American Board of Medical Specialties, Evanston, Illinois. Christopher K. Daugherty, M.D., is Associate Professor, Department of Medicine, University of Chicago. Ming Wen, Ph.D., is Assistant Professor, Department of Sociology, University of Utah, Salt Lake City. Paul Barach, M.D., M.P.H., formerly Associate Dean for Patient Safety and Associate Professor, Departments of Anesthesia, Pediatrics and Epidemiology, University of Miami, is Associate Professor, Departments of Anesthesiology, Environmental and Occupational Health, University of South Florida College of Medicine and Public Health, Tampa, Florida. Please send correspondence to Paul Barach, [email protected].
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety References 1. Institute of Medicine: To Err Is Human: Building a Safer Health System. Washington, D.C.: National Academy Press, 1999. 2. Institute of Medicine: Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, D.C.: National Academy Press, 2001. 3. Leape L.: Error in medicine. JAMA 272:1851–1857, Dec. 21, 1994. 4. Pierluissi E., et al.: Discussion of medical errors in morbidity and mortality conferences. JAMA 3290:2838–2342, Dec. 3, 2003. 5. Bosk C.: Forgive and Remember: Managing Medical Failure. Chicago: The University of Chicago Press, 1979. 6. Reason J.: Managing the Risks of Organizational Accidents. Aldershot, U.K.: Ashgate Publishing, 1997. 7. Helmreich R., Merrit A.: Culture at Work in Aviation and Medicine: National, Organizational, and Professional Influences. Aldershot, U.K.: Ashgate Publishing, 1998. 8. Roberts K.: Research in nearly failure-free, high reliability organizations: Having the bubble. IEEE Transactions on Engineering Management 36:132–139, May 1989. 9. Hale A.: Introduction: The goals of event analysis. In After the Event: From Accident to Organizational Learning. New York: Elsevier, 1997, pp. 1–10. 10. Wu A., et al.: Do house officers learn from their mistakes? JAMA 265:2089–2094, Apr. 24, 1991. 11. Casarett D., Helms C.: Systems errors versus physicians’ errors: Finding the balance in medical education. Acad Med 74:19–22, Jul. 1999. 12. Small S., Barach P.: Statewide survey of Massachusetts physician attitudes towards policy and workplace issues of patient safety (abstract). Anesthesiology 93:A-1188, 2000.
August 2007
13. Hobgood C., Ma J., Swart G.: Emergency medicine resident errors: Identification and educational utilization. Acad Emerg Med 7:1317–1320, Nov. 2000. 14. Eliastam M., Mizrahi T.: Quality improvement, housestaff, and the role of chief residents. Acad Med 71:670–674, Jun. 1996. 15. Weingart S.: A house officer-sponsored quality improvement initiative: Leadership lessons and liabilities. Jt Comm J Qual Improv 24:371–378, Jul. 1998. 16. Sexton J., Thomas E., Helmreich R.: Error, stress, and teamwork in medicine and aviation: Cross sectional surveys. BMJ 320:745–749, Mar. 18, 2000. 17. Bognar A., et al.: Assessing the burden of error recognition on patient care in cardiac surgical teams. Paper presented at American Society of Anesthesiology Annual Meeting, Chicago, Oct. 15, 2006. 18. Harman H.: Modern Factor Analysis. Chicago: University of Chicago Press, 1976. 19. Vohra P., et al.: Medical student and housestaff attitudes towards adverse medical events. Abstract presented at the 89th Annual Clinical Congress, American College of Surgeons, Chicago, Oct. 20, 2003. 20. Gilula M., Barach P.: Designing a patient safety curriculum. In Sheikh A., Hurwitz B. (eds.): Health Care Errors and Patient Safety. London: BMJ Books, Wiley-Blackwell, forthcoming, 2007. 21. Accreditation Council for Graduate Medical Education (ACGME): ACGME Common Program Requirements, Feb. 2004. http://www.acgme.org/acWebsite/dutyHours/ dh_dutyHoursCommonPR.pdf (last accessed Jun. 7, 2007). 22. Association of American Medical Colleges: Contemporary Issues in Medicine: Quality of Care, Aug. 2001. http://www.aamc.org/meded/ msop/msop5.pdf. (last accessed Jun. 7, 2007).
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
501
Health Professions Education
Housestaff and Medical Student Attitudes Toward Medical Errors and Adverse Events
Pamela D. Vohra, M.D. Julie K. Johnson, M.S.P.H., Ph.D. Christopher K. Daugherty, M.D. Ming Wen, Ph.D. Paul Barach, M.D., M.P.H.
T
he Institute of Medicine proposed that quality improvement and patient safety practices should be a formal part of medical training and that an emphasis on safety would improve the quality of medical treatment, enhance the public’s view of medicine, and decrease the costs of health care.1 Its 2001 report, Crossing the Quality Chasm, broadly addressed the need for improving professional education with recommendations that “colleges of medicine . . . should build more instruction into their curriculum on patient safety and its relationship to quality improvement.”2(p. 19) Concerns about medical errors that occur during physician training (residency) programs have directly resulted in significant changes in medical training (for example, an 80-hour work week, admission caps, and strict housestaff supervision requirements). Traditionally, adverse events were thought to be an unavoidable outcome of learning to practice medicine,3 and medical errors by trainees were viewed as teaching opportunities. The most frequent methods of dealing with errors, such as morbidity and mortality conferences, were punitive4—an approach that, from an educational point of view, may have a negative effect on learning.5 There is a growing acceptance that systemic problems, rather than poor individual performance, underlie most adverse events in health care.6 A culture of safety, as practiced, for example, in aviation and the nuclear power industry, which encourages reporting about and learning from all errors,7,8 is increasingly viewed as the most effecAugust 2007
Article-at-a-Glance Background: A lack of formal patient safety curricula has contributed to the suboptimal training of medical students and housestaff. Attitudes of physician trainees regarding medical errors and adverse events were surveyed in a pilot study. Methods: Five hundred sixty-three physician trainees were surveyed at an urban teaching hospital. Five domains were evaluated using a factor analysis as they relate to patient safety: knowledge, self-efficacy, awareness of safety culture, barriers/facilitators, and awareness of human factors. Results: One hundred fifty-eight (28%) trainees completed the survey, with 22% (n = 35) describing exposure to at least one adverse medical event. The survey showed good internal validity and reliability. Respondents who reported exposure to adverse events demonstrated a lower awareness of human factor errors (p = .0017) and lower awareness of the hospital’s approach to safety (p = .033). Older respondents scored higher on measures of self-efficacy than younger trainees (p = .042). Discussion: The exposure of physician trainees to errors and adverse events can have a negative effect on their attitudes and competencies. Exposure to adverse events and the institution’s response may decrease both error reporting and the willingness to adopt safety practices. The results support the need for implementing a sustained patient safety curriculum that promotes learning regarding adverse events.
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
493
The Joint Commission Journal on Quality and Patient Safety
tive means to enhance the reliability of care and reduce the recurrence of medical errors.9 Researchers have found that house officers who accepted and learned from their mistakes are more likely to make 10 constructive changes in their practice. Such acceptance, however, also adds a significant burden of emotional distress, and many residents became uncomfortable sharing with their attending physicians errors and near misses that they have observed or committed. This observation has strengthened the move away from a culture of blame to anonymous error reporting, systems analysis of adverse events, and collective learning opportunities.11,12 Some residency teaching departments (for example, emergency medicine13) have begun to establish a blame-free, errorbased curriculum. It has been proposed that if errors caused by systemic causes such as overwork and fatigue were addressed, house officers could focus on how their own actions contribute to adverse events.11 However, to date, housestaff have received little instruction in human factors, patient safety, continuous quality improvement, and systems thinking.12,14 Furthermore, evidence suggests that housestaff are inadequately prepared for identifying and overcoming the 5,15 organizational barriers to improving patient safety. Wu et al. examined how house officers learn from their own mistakes but not how they learn from the mistakes of others.12 Previous research has not addressed trainees’ perceptions of their self-efficacy and role in reducing medical error and patient harm. Nor has it addressed the impact of trainees’ exposure to institutional responses to adverse events. This article reports on a pilot study, undertaken at a large tertiary care medical center, which had the following two goals: 1. To describe the knowledge, self-efficacy, and attitudes of physician trainees toward adverse events 2. To explore the impact of adverse events and the local political and educational environment on the attitudes of physician trainees
at http://umdas.med.miami.edu/links/sentinelevents/— stated the following: Adverse events sometimes fall under the category of sentinel events, defined by the Joint Commission on Accreditation of Healthcare Organizations as an unexpected occurrence involving death or serious physical or psychological injury, or the events are called ‘sentinel’ because they signal the need for immediate investigation and response.* The orientation to the survey continued as follows: This study aims to examine how housestaff and medical students assess adverse events of health care at the academic hospitals. The purpose of this survey is to obtain your thoughts and perspectives regarding three basic areas: (1) Awareness of the need to reduce medical errors and improve patient safety, and the perceived barriers to awareness (2) Assessment of your role in reducing medical errors and improving patient safety, and the impact of adverse event analysis meetings (also known as root cause analysis [RCA] meetings) and outcomes on your daily practice and on patient care (3) Potential opportunities that should be considered in your department for reducing medical errors and improving patient safety
STUDY POPULATION
ORIENTATION TO THE SURVEY
After full Institutional Review Board approval, 453 medical residents and 100 senior medical students (physician trainees) at an urban teaching hospital were invited in 2004 to participate in the study. Subjects were recruited via e-mails sent by their program directors or by the medical student listserve network. Reminder e-mails to nonrespondents were sent out three months after the first invitation. The respondents anonymously completed electronic questionnaires on a secure Web site16 and provided informed online consent to participate in the study. Each completed survey was given a random computer-generated code. All identifiers were removed from the data before analysis. Respondents were compensated with an online bookstore gift certificate worth $5 (allocated by a random
The survey, “Sentinel Events at the Academic Hospitals: Evaluation of Housestaff and Medical Student Attitudes toward Adverse Medical Events”—a sample may be found
* The Joint Commission: Sentinel Event Policy and Procedures. Updated Oct. 2006. http://www.jointcommission.org/SentinelEvents/Policyand Procedures/se_pp.htm (last accessed Jun. 14, 2007).
Methods
494
August 2007
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety
number generator so that investigators could not track respondents). Respondents who indicated in the survey that they had participated in a sentinel event and a subsequent RCA meeting to analyze the incident were invited to discuss their experiences in a confidential, semistructured interview. These subjects were contacted about the interview process via e-mail. All interviewees gave written consent and were compensated with a bookstore gift certificate worth $20.
SURVEY The survey was developed and refined based on an extensive literature review. Most of the questions were adapted from previously validated tools16,17 and underwent extensive assessment by the authors. A draft of the survey was pilot-tested on five subjects and subsequently refined. The survey explored the following three areas: 1. Awareness (knowledge) of the need to reduce medical errors and barriers to improving patient safety (barrier analysis) 2. Assessment of the trainee’s role in reducing medical errors and improving patient safety 3. Assessment of the impact of institutional sentinel event RCA meetings and their outcomes on daily practice of patient care The survey used a 5-point Likert scale (“strongly disagree” to “strongly agree”) categorical response format and included several open-ended questions.
SCALE CONSTRUCTION We used an initial affinity sort of the responses, followed by a modified Delphi process to reach consensus in identifying five domains of housestaff and medical student attitudes toward medical errors and adverse events. These domains emerged from review of the literature and from assessing and analyzing the data at the pilot phase. They were then used to measure the attitudes towards patient safety. The five domains were as follows: 1. Knowledge: awareness of the need for improvement in patient safety 2. Self-efficacy: belief in one’s ability to reduce errors and improve patient safety 3. Awareness of safety culture: error-reporting behaviors (such as willingness to discuss or take responsibility for errors) and willingness to adopt new safety practices
August 2007
4. Beliefs about barriers/facilitators: attitudes about systemic barriers to improvement in patient safety 5. Awareness of human factors: perception and knowledge of how education/training, communication, and other organizational factors affect patient safety. The items and their respective domains are shown in Table 1 (page 496). The five scales were constructed using orthogonal principal factor analysis, a data reduction technique concerned with finding a small number of common factors that linearly reconstruct the original variables.18 The factor loadings reflect the strength of the correlation between the items and the latent common factors. Table 2 (page 497) presents the ranges of factor scores and internal consistency and reliability coefficients (Cronbach’s ␣) for the five scales corresponding to the five domains. These scales are weighted on the basis of each item’s factor loadings. The use of aggregate scales allowed for the fit of parsimonious and efficient regression models to identify predictors of attitudes toward patient safety.
INTERVIEWS The confidential interview, which consisted of semistructured questions designed to be completed in 45 minutes, explored residents’ attitudes toward adverse events, their role in assessing the incident, and the institutional response. The seven interviews were recorded and transcribed verbatim; tapes were subsequently erased. The identifiers on the transcripts were removed before analysis. The interviews were coded [J.K.J., P.B.].
STATISTICAL ANALYSIS Alpha coefficients of each of the five scales were computed to measure the scales’ internal reliability. Pair-wise Pearson correlation analyses were performed for the five scales. These scales were used as the dependent variables in multiple ordinary least squares (OLS) regression analyses, which were conducted to assess the effects of age, gender, department, level and length of postgraduate training, and previous involvement in a sentinel event and a related RCA meeting on each of the five domains.
Results CHARACTERISTICS OF RESPONDENTS Of the 463 medical residents and 100 medical students
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
495
The Joint Commission Journal on Quality and Patient Safety
Table 1. Factor Analysis of the Safety Domains*
Number Agreed
Percentage Agreed
Factor Loading
Knowledge Significant efforts are needed to reduce errors and improve patent safety in my immediate professional practice/work area.
87
76
0.90
Significant efforts are needed to reduce errors and improve patent safety throughout this hospital.
97
85
0.87
In general, the residency program directors with whom I regularly work believe that significant efforts are needed to reduce errors and improve patient safety in our immediate professional practice/work areas.
74
65
0.69
Self-Efficacy I have a significant role in the reduction of errors and improvement of patient safety in this institution.
70
61
0.64
I am encouraged by department administrators and supervisors to reduce errors and improve patient safety.
68
60
0.80
When errors do occur in my professional setting, I am asked to help identify sources of error in this incident and ways of correcting for these errors.
57
50
0.78
When errors do occur in my professional setting, I am able to adopt new practices and procedural changes that may reduce error and improve patient safety.
67
59
0.69
Safety Culture If I was directly involved in a medical error, I would document it in the patient’s case record.
110
96
0.87
If I was directly involved in a medical error, I would tell the patient and/or family that I made the mistake.
99
87
0.74
If I was directly involved in a medical error, I would inform the patient’s attending physician.
109
96
0.73
Barriers/Facilitators In my immediate professional setting I feel that communications between patient care team members is sufficient to ensure safe patient care.
38
33
0.80
In my immediate professional setting I feel that technology is adequately used to promote safe patient care.
26
23
0.73
In my immediate professional setting I feel that there are adequate staffing resources to minimize errors and maintain an acceptable level of patient safety.
24
21
0.70
Human Factors When errors do occur in my professional setting, I am informed of their occurrence by colleagues, department administrators, or supervisors.
65
57
0.74
When errors do occur in my professional setting, I am given adequate information about the causes of the error and ways of improving patient safety.
46
40
0.86
When errors do occur in my professional setting, I am notified of procedural changes implemented to reduce errors and improve patient safety.
42
37
0.77
Factor
* Out of 114 responses.
496
August 2007
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety
Table 2. Internal Consistency and Reliability Coefficients of Five Scales* Scale
No. of Items
Cronbach’s alpha
Factor Loading Range
Knowledge
3
.76
.69–.90
Self-Efficacy
4
.71
.64–.80
Barriers/Facilitators
3
.60
.70–.80
Safety Culture
3
.68
.73–.87
Human Factors
3
.70
.74–.86
* Factor loading ranges and Cronbach alpha values.
Table 3. Descriptive Statistics and Factor Loadings of the Scales Corresponding to the Five Domains Independent Variables
n
%
Knowledge
Self-Efficacy
114
100
1.76
1.42
-1.24
3.12
-4.44
Male
60
53
-.14
-.01
-.05
.03
-.01
Female
54
47
.15
.02
.11
-.22
.03
Medical Student
37
32
.09
-.23
-.20
-.09
-.13
Resident
77
68
-.04
.11
.96
.05
.06
Overall
Safety Culture Barriers
Human Factors
Gender
Position
Department (Residents)* Surgery
17
15
-.06
.20
.05
-.31
-.09
Pediatrics
23
20
.21
.21
.08
-.03
.21
Internal Medicine
12
11
-.02
.19
.26
-.04
.05
Exposure
35
31
0.10
0.05
0.15
-0.39
.03
None
79
69
-0.04
-0.02
-0.07
0.017
-.01
Sentinel Event
* In addition, six residents (5%) were from the ophthalmology department, five (4%) were from the emergency medicine department, and four (4%) were from other departments.
invited to participate, 158 (28%) medical residents and medical students responded (44 responses were excluded due to incomplete responses). The study population included 114 medical residents and medical students, combined because of the small numbers. The median age of respondents was 28 years (range, 24–45), and median postgraduate medical experience was 3 years (range, 1–10). As shown in Table 3 (above), medical students made up 32% (n = 37) of the respondents, and 68% (n = 77) were residents. Thirty one percent of the respondents (45% of responding residents) reported being exposed to an adverse event, and 17% reported participating in an
August 2007
RCA meeting. Of the respondents who reported exposure to an adverse event, 22% were unaware of the hospital’s counseling services, and 19% of all respondents did not know how to report errors using the hospital reporting system. A few respondents reported having seen so many errors that it was difficult to choose which one to describe.
SURVEY PSYCHOMETRICS Following the completion of the exploratory factor analysis and elimination of items with low factor loadings and/or substantial “cross-factor” loadings, a final factor structure was produced. The items, descriptive statistics,
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
497
The Joint Commission Journal on Quality and Patient Safety
Table 4. Attributed Causes of Adverse Events (Number and Percentage of Responses)* Fatigue
18 (44%)
Inexperience
17 (41%)
Faulty judgment
17 (41%)
Hesitancy Other
5 (12%) 23 (56%)
* Percentages add to more than 100% as respondents could choose several causes.
and factor loadings are shown in Table 1 (page 496). For example, the knowledge scale consists of three items, with factor loadings ranging from 0.69 to 0.90; the Cronbach’s alpha coefficient is equal 0.76. The higher the score, the more knowledgeable the trainee. Similarly, the higher the self-efficacy score, the greater the trainee’s self-efficacy; the higher the barriers/facilitators score, the greater the understanding and impact of barriers to patient safety; the higher the safety culture score, the more a culture of safety is needed; the higher the human factors score, the less knowledge of human factors the trainees had. This analysis suggests that the scales have reasonably good internal reliability.
reflected in the scales. The length of medical training was slightly (that is, nonsignificantly) negatively correlated (b = -.14, p = .10) with willingness to adopt new safety measures (safety culture). Exposure to a sentinel event was negatively associated with awareness of safety culture (b = -.55, p = .012). Further analyses of specific items in the barriers/facilitators scale indicated that respondents who described exposure to a sentinel event had lower scores on questions pertaining to adequacy of staffing resources (p = .01) and higher scores related to the impact of technology on patient outcomes (p = .04). We found no significant differences between the responses of medical students and residents.
INTERVIEW RESULTS The seven interview respondents were asked to list the underlying causes of the adverse events that they reported (Table 4, left). The causes reported included, in descending rank order, fatigue, inexperience, faulty judgment, and hesitancy. Selected responses to the open-ended interview questions are presented in Table 5 (page 499) according to the five factor domains of the questionnaire tool. These observations support the concerns represented by the items in the questionnaire’s five domains (Table 1).
SURVEY RESULTS
Discussion
The self-efficacy domain was correlated significantly and positively with three other domains: human factors (r = .78, p < .01), safety culture (r = .26, p < .01), and barriers/facilitators (r = .19, p = .04). The human factors domain was positively correlated with the barriers factor (r = .33, p < .01), and the safety culture (r = .24, p < .01). The knowledge domain was significantly and negatively correlated with safety culture (r = -.33, p < .01), suggesting that safety culture is independent from safety knowledge and requires a detailed training to acquire. As shown in Table 3 (page 497), the factor loadings of the scores are stratified by gender, years of postgraduate training (position), department, and experiences with sentinel events and related RCA meetings (exposure). Results from the regression analyses show that older respondents on average had a higher self-efficacy score than younger trainees (t = 2.24; p = .01); no age differences were shown for the four other domains. Department and positions were not
Physicians in training working in a high-stress environment are highly influenced by the attitudes and perceptions of their faculty and teaching organizations toward patient safety. The data presented in this study support the reliability and face validity of the survey developed to assess the attitudes of housestaff and medical students toward patient safety. The study, the first to report on the impact of medical errors and adverse events on medical students and residents (trainees), reinforces our knowledge about the ubiquitous nature of human error and patient harm in health care. It points to numerous opportunities for error identification and harm mitigation in these highrisk settings. (Although a greater proportion of students than residents responded to the survey, we found no significant differences between their respective responses.) Medical trainees exposed to medical errors were likely to report a significantly negative attitude toward patient safety. Furthermore, those who reported exposure to med-
498
August 2007
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety
Table 5. Selected Responses to Open-Ended Interview Questions in Five Domains Knowledge ■ “There have been so many [errors], but here is one.” ■ “I’ve written many wrong doses of medications and rely on the pharmacy to double-check me.” Self-Efficacy ■ “If I had seen him sooner, or at least notified the chief resident sooner, we could have gotten him to the operating room sooner for repair of his [aneurysm] leak, and he may have survived.” Safety Culture ■ “And he told me to pay more attention to the patient. Yes, I made the mistake, but hands-down I still and always did know that patient better than he did.” Barriers ■ “Our hospital system is so defunct; most residents find little reason to report errors.” ■ “Most errors which I have been involved with have been related to inadequate staffing, and systems inadequacies so that drugs were not administered in a timely fashion, or critical nursing functions were not performed.” Human Factors ■ “The equipment was not designed with residents in mind.”
ical errors and patient harm had a negative learning experience, that is, they learned little from the experience and also learned to not discuss or share the experience with others. Trainees who were exposed to an adverse event and its subsequent analysis responded negatively to survey issues surrounding the adequate use of technology, staffing resources, and sufficient training. These results suggest that beyond a lack of awareness and knowledge about important factors such as the role of technology in preventing patient harm, the trainees were quick to internalize a personal and nonsystems approach towards errors. This behavior was likely reinforced by explicit and implicit actions and cues that they internalized from their senior colleagues and others during analysis of these events. The positive correlation between self-efficacy and beliefs about barriers/facilitators or awareness of safety culture suggests that trainees’ confidence in addressing errors and adverse patient events affects their learning from these events. Trainees who have a higher confidence in their ability to improve patient safety may also have a more real-
August 2007
istic view of the system’s safety constraints and enablers than trainees with less confidence. One could hypothesize that this may lead to an increased awareness of the challenges to advancing patient safety in academic health centers, which, in turn, could encourage these trainees to be active change agents. Increased respondent age, but not additional years of training (medical student versus resident), was associated with an increased reported self-efficacy and confidence to deal with adverse events. We believe that this gap between age and training duration further underscores the need for a formal iterative safety curriculum across the continuum of medical education, from undergraduate through graduate medical training. Trainees exposed to medical errors and patient harm reported less awareness of hospital efforts to prevent harms and less confidence in the hospital systems than their peers who did not report such exposure. The comments provided in the interviews provide important additional insights and strongly support the survey’s scaled questions. For example, beliefs regarding hospital care, the provider’s work overload, limited support from program administrators and chief residents, and a punitive culture are concerning. These findings suggest that if impressionable physician trainees find their exposure to medical errors to be unpleasant or blameful, they may be less likely to engage in subsequent error-reporting behaviors (such as discussing or taking responsibility for errors) or to adopt new safety interventions and practices. This speaks to the importance of creating a nonpunitive setting for identifying and addressing adverse patient care early in medical training. We believe that the current culture in academic health centers, in which unclear accountability standards are promulgated, may discourage trainees from engaging and committing their energies to improve patient safety. Similarly, Wu et al.10 found that residents were less likely to make constructive changes if they attributed a mistake to job overload or if they felt that the institution was judgmental. At a minimum, our data suggest that trainees’ limited involvement in designing and implementing patient safety interventions reduces the opportunities for them to become effective change agents—or more worrisome, to even learn from their and others’ mistakes. We recognize several limitations to the study. Focusing on clinician perceptions is reliant on self-reports of current
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
499
The Joint Commission Journal on Quality and Patient Safety
and retrospective perceptions, which may be a source of richness but also bias. These events could not be verified, although many of the described events did come to our attention independently. Therefore, we can discuss only an association between perceived institutional responses and the attitudes reported by trainees. As a result, we do not have the ability to infer causality. Second, we cannot tell if there are significant differences between the responders and the nonresponders to the survey. Because of the anonymity of the subjects, nonrespondents could not be contacted (required by the Institutional Review Board; we chose an anonymous survey format to preserve the integrity of the responses and to encourage participation). Anonymous sampling is likely to result in improved veracity, and under some circumstances, improved generalizability of the findings, even if it can result in a smaller sample. Third, the overall response rate to our questionnaire was low, perhaps in part because of the use of an e-mail survey. In addition, medical students and residents have been difficult to survey in general, using e-mail in particular. It is possible that the perceived punitive culture in academic health centers may actually contribute further to the low response rates for a survey on patient safety (several comments in the open-ended interviews support this finding). The sensitive nature of this topic likely discouraged some participants from responding despite elaborate efforts to protect their anonymity. Even if the low response rate reflects a negative response bias, as suggested by Wu et al.,10 the findings may still carry strong validity. Fourth, the data represent only one teaching hospital, which limits the generalizability of our findings. Early results from a similar study at another large academic medical center support the findings of this study.19 Further development of the survey and other instruments to measure behavior and attitudes related to experiences regarding medical errors and adverse events is needed.
Conclusions The findings suggest that efforts to better incorporate a patient safety culture and enhance patient safety training are clearly needed. These findings also highlight the personal and professional burden that the ever-present recognition of errors and fear of these errors represent for physician trainees. There is a need for well-designed cur500
August 2007
ricula within medical schools and residency programs to help trainees learn from their experiences with adverse events in a constructive and nonpunitive manner.20 Promoting self-awareness and assessment in future physicians is key to lifelong learning and development. Trainees should be taught how to report errors and adverse events in a safe way and to seek counseling when adverse events occur. If trainees feel safe, they are more likely to report such events. Such curricula would be consistent with the efforts of the Accreditation Council for Graduate Medical Education and the Association of American Medical Colleges to make patient safety central to undergraduate and graduate medical education.21,22 Senior faculty need to appreciate the importance of sharing their experience with and reflections about medical errors (their own and those of others) and how they have learned to make sense of them. Learning to acknowledge and report adverse patient care should help to foster truth telling about clinical care and instill a more professional and reflective approach toward adverse care in these future physicians. J The research described in this article was supported by a Pritzker Summer Research Program Grant. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the Universities of Chicago, Utah, or Miami. The authors thank the medical students and housestaff who participated in the study; David Faber, who designed the questionnaire’s Web site and assisted in the data collection; and Dr. Richard Frankel for his editorial help.
Pamela D. Vohra, M.D., is a Second-Year Resident, Department of Medicine, Beth Israel Deaconess Medical Center, Boston. Julie K. Johnson, M.S.P.H., Ph.D., is Assistant Professor, Department of Medicine, University of Chicago, Chicago, and Director of Research, American Board of Medical Specialties, Evanston, Illinois. Christopher K. Daugherty, M.D., is Associate Professor, Department of Medicine, University of Chicago. Ming Wen, Ph.D., is Assistant Professor, Department of Sociology, University of Utah, Salt Lake City. Paul Barach, M.D., M.P.H., formerly Associate Dean for Patient Safety and Associate Professor, Departments of Anesthesia, Pediatrics and Epidemiology, University of Miami, is Associate Professor, Departments of Anesthesiology, Environmental and Occupational Health, University of South Florida College of Medicine and Public Health, Tampa, Florida. Please send correspondence to Paul Barach, [email protected].
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
The Joint Commission Journal on Quality and Patient Safety References 1. Institute of Medicine: To Err Is Human: Building a Safer Health System. Washington, D.C.: National Academy Press, 1999. 2. Institute of Medicine: Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, D.C.: National Academy Press, 2001. 3. Leape L.: Error in medicine. JAMA 272:1851–1857, Dec. 21, 1994. 4. Pierluissi E., et al.: Discussion of medical errors in morbidity and mortality conferences. JAMA 3290:2838–2342, Dec. 3, 2003. 5. Bosk C.: Forgive and Remember: Managing Medical Failure. Chicago: The University of Chicago Press, 1979. 6. Reason J.: Managing the Risks of Organizational Accidents. Aldershot, U.K.: Ashgate Publishing, 1997. 7. Helmreich R., Merrit A.: Culture at Work in Aviation and Medicine: National, Organizational, and Professional Influences. Aldershot, U.K.: Ashgate Publishing, 1998. 8. Roberts K.: Research in nearly failure-free, high reliability organizations: Having the bubble. IEEE Transactions on Engineering Management 36:132–139, May 1989. 9. Hale A.: Introduction: The goals of event analysis. In After the Event: From Accident to Organizational Learning. New York: Elsevier, 1997, pp. 1–10. 10. Wu A., et al.: Do house officers learn from their mistakes? JAMA 265:2089–2094, Apr. 24, 1991. 11. Casarett D., Helms C.: Systems errors versus physicians’ errors: Finding the balance in medical education. Acad Med 74:19–22, Jul. 1999. 12. Small S., Barach P.: Statewide survey of Massachusetts physician attitudes towards policy and workplace issues of patient safety (abstract). Anesthesiology 93:A-1188, 2000.
August 2007
13. Hobgood C., Ma J., Swart G.: Emergency medicine resident errors: Identification and educational utilization. Acad Emerg Med 7:1317–1320, Nov. 2000. 14. Eliastam M., Mizrahi T.: Quality improvement, housestaff, and the role of chief residents. Acad Med 71:670–674, Jun. 1996. 15. Weingart S.: A house officer-sponsored quality improvement initiative: Leadership lessons and liabilities. Jt Comm J Qual Improv 24:371–378, Jul. 1998. 16. Sexton J., Thomas E., Helmreich R.: Error, stress, and teamwork in medicine and aviation: Cross sectional surveys. BMJ 320:745–749, Mar. 18, 2000. 17. Bognar A., et al.: Assessing the burden of error recognition on patient care in cardiac surgical teams. Paper presented at American Society of Anesthesiology Annual Meeting, Chicago, Oct. 15, 2006. 18. Harman H.: Modern Factor Analysis. Chicago: University of Chicago Press, 1976. 19. Vohra P., et al.: Medical student and housestaff attitudes towards adverse medical events. Abstract presented at the 89th Annual Clinical Congress, American College of Surgeons, Chicago, Oct. 20, 2003. 20. Gilula M., Barach P.: Designing a patient safety curriculum. In Sheikh A., Hurwitz B. (eds.): Health Care Errors and Patient Safety. London: BMJ Books, Wiley-Blackwell, forthcoming, 2007. 21. Accreditation Council for Graduate Medical Education (ACGME): ACGME Common Program Requirements, Feb. 2004. http://www.acgme.org/acWebsite/dutyHours/ dh_dutyHoursCommonPR.pdf (last accessed Jun. 7, 2007). 22. Association of American Medical Colleges: Contemporary Issues in Medicine: Quality of Care, Aug. 2001. http://www.aamc.org/meded/ msop/msop5.pdf. (last accessed Jun. 7, 2007).
Volume 33 Number 8
Copyright 2007 Joint Commission on Accreditation of Healthcare Organizations
501