Surgeons’ Disclosures of Clinical Adverse Events

Key Points

Question  How are surgeons disclosing adverse events to patients, and what is the effect of such disclosure on surgeons’ well-being?

Findings  In this observational study of surgeons at Veterans Affairs facilities who completed 62 web-based surveys about their experiences disclosing adverse events to patients, those who were less likely to follow nationally recommended elements of disclosure were more likely to report being negatively affected by the adverse event.

Meaning  Open disclosure programs are being implemented nationwide, but without training on disclosing adverse events using specific communication elements, surgeons may experience negative effects when disclosing such information.

Importance  Surgeons are frequently faced with clinical adverse events owing to the nature of their specialty, yet not all surgeons disclose these events to patients. To sustain open disclosure programs, it is essential to understand how surgeons are disclosing adverse events, factors that are associated with reporting such events, and the effect of disclosure on surgeons.

Objective  To quantitatively assess surgeons’ reports of disclosure of adverse events and aspects of their experiences with the disclosure process.

Design, Setting, and Participants  An observational study was conducted from January 1, 2011, to December 31, 2013, involving a 21-item baseline questionnaire administered to 67 of 75 surgeons (89%) representing 12 specialties at 3 Veterans Affairs medical centers. Sixty-two surveys of their communication about adverse events and experiences with disclosing such events were completed by 35 of these 67 surgeons (52%). Data were analyzed using mixed linear random-effects and logistic regression models.

Main Outcomes and Measures  Self-reports of disclosure assessed by 8 items from guidelines and pilot research, surgeons’ perceptions of the adverse event, reported personal effects from disclosure, and baseline attitudes toward disclosure.

Results  Most of the surgeons completing the web-based surveys (41 responses from men and 21 responses from women) used 5 of the 8 recommended disclosure items: explained why the event happened (55 of 60 surveys [92%]), expressed regret for what happened (52 of 60 [87%]), expressed concern for the patient’s welfare (57 of 60 [95%]), disclosed the adverse event within 24 hours (58 of 60 [97%]), and discussed steps taken to treat any subsequent problems (59 of 60 [98%]). Fewer surgeons apologized to patients (33 of 60 [55%]), discussed whether the event was preventable (33 of 60 [55%]), or how recurrences could be prevented (19 of 59 [32%]). Surgeons who were less likely to have discussed prevention (33 of 60 [55%]), those who stated the event was very or extremely serious (40 of 61 surveys [66%]), or reported very or somewhat difficult experiences discussing the event (16 of 61 [26%]) were more likely to have been negatively affected by the event. Surgeons with more negative attitudes about disclosure at baseline reported more anxiety about patients’ surgical outcomes or events following disclosure (odds ratio, 1.54; 95% CI, 1.16-2.06).

Conclusions and Relevance  Surgeons who reported they were less likely to discuss preventability of the adverse event, or who reported difficult communication experiences, were more negatively affected by disclosure than others. Quality improvement efforts focused on recognizing the association between disclosure and surgeons’ well-being may help sustain open disclosure policies.

National guidelines recommend full disclosure of adverse events or unanticipated outcomes to patients and their family members.^1,2 Evidence shows that such full disclosure involving transparent and honest communication benefits patients and families.^3,4 Nonetheless, physicians, including surgeons, frequently fail to disclose adverse events to their patients.^5-10 Even when physicians report favorable attitudes toward disclosure, they often do not tell patients all the details about a clinical adverse event.⁵

Quiz Ref IDThe National Quality Forum provides guidelines for the content of disclosures to patients, which includes providing facts about the event to patients and families, including whether or not an error or system failure occurred, and results of any event analysis to support informed decision-making by the patient; expressing regret for the outcome; and giving a formal apology if the unanticipated outcome was caused by an error or system failure.^1,7 The Department of Veterans Affairs (VA) handbook, Disclosure of Adverse Events to Patients, instructs the clinical team to provide factual information to the extent it is known, express concern for the patient’s welfare, and reassure the patient or representative that steps are being taken to investigate the situation, remedy any injury, and prevent further harm.¹¹ However, to our knowledge, few studies have examined the content of surgeons’ disclosures of adverse events, and it remains unknown which elements of disclosure are used by surgeons during these conversations, or how such disclosure affects patients and physicians.

Many researchers have argued that disclosure is as important for physicians as it is for patients,¹² stating that disclosure provides an outlet for the emotional toll that physicians experience as part of the adverse event process.^13,14 Physicians may fear that disclosing adverse events increases the risk of negative consequences related to the adverse event,¹⁵ with physicians being considered the “second victims.”¹⁶ After adverse events, surgeons may experience burnout, shame, guilt, and even depression, contributed to by litigation and concerns, as well as colleagues’ reactions.^17-19 A survey administered to 108 surgery, emergency medicine, and anesthesiology physicians at one hospital showed that 85 (78.7%) had experienced either a serious adverse patient event and/or a traumatic personal event within the previous year.²⁰ Many of these physicians reported several barriers to seeking support for these events, including a perceived negative effect on their careers, stigma, and lack of confidentiality in the process. An emotional toll may also be taken when no explicit medical error has occurred, or when an adverse event is potentially unpreventable.²¹

This study provides information on how surgeons describe aspects of their experiences with disclosures of adverse surgical events, in the absence of specific training on disclosure. We sought to test whether surgeons who reported adverse events using elements of disclosure advocated by the National Quality Forum and VA were less likely to experience negative effects after disclosure (eg, increased anxiety), whether surgeons who reported more negative perceptions about adverse events (eg, perceived seriousness of the event) and disclosure (eg, more difficulty discussing the event) were more likely to experience negative effects after disclosure, and whether surgeons who reported positive attitudes toward disclosure were less likely to experience negative effects after disclosure. Examining disclosure of adverse events and its effects from the surgeon’s perspective may inform efforts to improve the disclosure process as well as both surgeons’ and patients’ outcomes following disclosure.

This prospective study took place at 3 VA medical centers in the United States from January 1, 2011, through December 31, 2013. Twelve surgical specialties—cardiac, general, gynecology, neurosurgery, ophthalmology, orthopedic, otolaryngology, plastics, podiatry, thoracic, urology, and vascular—are practiced at these facilities. The study was approved by the Veterans Affairs Boston Health Care System, Veterans Affairs Connecticut Health Care System, Veterans Affairs Puget Sound Health Care System, and Edith Nourse Rogers Memorial Veterans Hospital (Bedford, Massachusetts) Institutional Review Boards, and a certificate of confidentiality from the National Institutes of Health was obtained to protect the identities of participating surgeons.

We obtained lists of surgeons from each medical center’s surgical service (n = 75) and introduced the study to these surgeons at group informational meetings. Individual meetings were then scheduled whereby a study team member further explained the study’s objectives. Participating surgeons provided written informed consent and completed a baseline 21-item questionnaire. We aimed to recruit surgeons from many specialties to increase the number of participants, decrease the risk of confidentiality breaches, and examine disclosures among a range of surgeons.

Quiz Ref IDSurgeons were asked to identify a case involving an unplanned return to the operating room or an unplanned procedural intervention following surgery, any time within 30 days of the original surgery. These criteria were kept broad to capture a multitude of events. Quality assurance information on facility adverse event reporting was unavailable to the study team, following 38 USC § 5705 (2010) (federal statutes on confidentiality of medical quality assurance records)²²; therefore, surgeons’ self-reports of disclosures were required. Surgeons contacted their site study team within 2 weeks of adverse event identification to complete a web-based survey about the disclosure. Emails were periodically sent to participating surgeons to remind them of this study process. This process adhered to strict institutional review board guidelines to protect surgeons’ confidentiality and to reduce coercion for study participation.

Attitude questions were adapted from a published survey²³ and were collected in response to 2 vignettes representing scenarios with the potential for high harm and low harm (eAppendix 1 in the Supplement). Research shows that physicians are more likely to disclose adverse events with the potential for high harm to patients rather than those with the potential for low harm.²³ The scenario with potential for high harm involved a surgical sponge retained during surgery but not discovered until postoperative day 8. Additional surgery was needed, and the patient did not fare well. The scenario with potential for low harm involved significant intraoperative blood loss that was corrected through autologous blood transfusion, and the patient recovered well. Following guidelines, scores from this questionnaire were divided into 2 subscales: a scale called positive attitudes toward disclosure, a continuous scale ranging from 4 to 20, with higher scores denoting more positive attitudes, and the other called negative attitudes toward disclosure, a categorical scale ranging from 0 to 6, with lower scores denoting more negative attitudes.

Once a surgeon reported disclosing an adverse event, a confidential survey web link created specifically for that surgeon and event was sent via email. Participants could complete up to 3 web-based surveys for 3 adverse events and disclosures during the study period. This 22-item survey was developed based on a qualitative pilot study of actual adverse event disclosures²⁴ and other previous work (eAppendix 2 in the Supplement).^8,9 This survey documented surgeons’ experiences of adverse events, their perceived seriousness of the events (from 1 [not at all serious] to 4 [extremely serious]), and whether they discussed the 8 different elements of disclosures identified through pilot interviews and disclosure guidelines from the National Quality Forum and VA (yes or no response). These 8 items ask (1) whether the surgeon explained why the event happened, (2) whether the event was preventable, (3) how recurrences of the event could be prevented, (4) whether the surgeon expressed regret for what happened, (5) whether the surgeon expressed concern for the patient’s welfare, (6) whether the surgeon apologized for the event, (7) whether the event was disclosed within 24 hours, and (8) whether the surgeon discussed steps taken to treat any subsequent problems. Each of these items was treated as a single-item scale in our analyses.

Several questions, validated in a previous study,⁹ also assessed the effect on surgeons from the event and its disclosure. Questions included how much disclosure of the event affected them overall on a scale of 1 (extremely) to 5 (not at all), whether there had been an overall negative effect on their life from this event (yes or no response), and questions (yes or no response) about whether disclosure affected their job satisfaction, their confidence in their ability as a physician, their professional reputation, their anxiety about future outcomes or events, and their ability to sleep.⁹ These items were treated as single-item scales in our data analysis. The survey included demographic questions (age, sex, and surgical specialty), and included a text box for a description of the event. Pilot testing determined that the survey took between 2 and 5 minutes to complete.

We performed descriptive analyses of the distribution of surgeon demographics and for the baseline and web-based surveys to determine the range of responses among surgeon participants across the 3 study sites. In addition, we examined the residual plots and normal probability plots of the studentized residuals, which revealed no evidence of nonlinearity; the normal probability plots raised no concerns about the conditional normality assumptions of the dependent variables. We then performed 2 sets of analyses on the survey data, with the disclosure event as our unit of analysis to test our hypotheses. The first set consisted of separate random-effects least-squares regressions of surgeons’ reports of the overall effect of the event on surgeons’ experience of disclosing the event, as well as their perceptions of the event. The second analysis consisted of random-effects logistic regressions of surgeons’ reports of whether they experienced an overall negative effect on their lives from the event or negative effect on specific aspects (ie, job satisfaction, confidence, reputation, anxiety, and ability to sleep) on their experience of disclosing the event, and surgeons’ baseline attitude scales by high- or low-harm scenario. Both sets of models controlled for surgeons’ age and sex and included separate normally distributed random-effects for each surgeon. All analyses were conducted using the SAS statistical package (SAS Institute); the random-effects least-squares regressions were implemented in PROC MIXED, and the random-effects logistic regressions were implemented in PROC GLIMMIX. The multiplicity of tests was addressed by performing a false discovery rate adjustment to the P values, an increasingly common procedure for multiple test corrections.^25,26

Sixty-seven of 75 surgeons (89%) across the study sites and representing each of the 12 surgical specialties consented to participate in the study and completed a baseline questionnaire assessing their attitudes toward disclosure of adverse events.²³ Of 67 eligible surgeons, 35 (52%) from 11 specialties completed 62 web-based surveys reporting on their disclosures of adverse events to patients and/or family members. eAppendix 3 in the Supplement contains descriptive statistics of study participants based on the 62 surveys (surgeons could complete up to 3 surveys). There were no significant sex differences between those who completed the web-based survey and those who did not (χ²₁ = 1.59, P = .21).

Thirty-six of 60 survey respondents (60%) indicated that the adverse event had affected them moderately, quite a bit, or extremely, but most did not report significant effects in areas such as job satisfaction, confidence, professional reputation, or ability to sleep. Seventeen of 62 survey respondents (27%) reported anxiety about future outcomes or events (eAppendix 4 in the Supplement).

Most participating surgeons reported that they used 5 of the 8 recommended disclosure items: explained why the event happened (55 of 60 surveys [92%]), expressed regret for what happened (52 of 60 [87%]), expressed concern for the patient’s welfare (57 of 60 [95%]), disclosed the adverse event within 24 hours (58 of 60 [97%]), and discussed steps taken to treat any subsequent problems (59 of 60 [98%]) (eAppendix 5 in the Supplement). Quiz Ref IDFewer surgeons discussed whether the event was preventable (33 of 60 surveys [55%]) and/or how recurrences of the event could be prevented (19 of 59 [32%]). Although VA guidelines do not mention apology, 33 of 60 survey respondents (55%) indicated that they apologized to the patient during disclosure.

Surgeons reported more positive attitudes toward disclosure in the high-harm scenario (mean [SD] score, 19.76 [0.82]) than in the low-harm scenario (mean [SD] score, 18.24 [2.57]) (t = –4.43; P < .001) (Table 1). There were no differences in negative attitudes toward disclosure reported in either scenario: surgeons’ mean (SD) score for the high-harm scenario was 2.90 (1.96), while their mean (SD) score for the low-harm scenario was 2.53 (2.13) (t = –1.01; P = .32) (Table 2).

Surgeons reported being significantly affected by disclosure when they perceived that the adverse event was more serious (β = 0.60; 95% CI, 0.31-0.90), reported very or somewhat difficult experiences when discussing the event with the patient and/or family (β = –0.61; 95% CI, –1.00 to –0.23), did not discuss with the patient or family whether the event was preventable (β = 1.00; 95% CI, 0.39-1.61), and did not discuss how recurrences of the event could be prevented (β = 1.39; 95% CI, 0.70-2.08) (Table 3). Clinically, this result means that the contribution of the predictor variable (ie, seriousness of the event, difficult experiences with disclosure, and no discussion of preventability) is significantly positive, and that each 1-point increase in the predictor variable corresponds to an additional increase (range, 0.60-1.39) in surgeons being negatively affected by the event. Quiz Ref IDMoreover, surgeons who reported very difficult experiences when discussing the event were 4 times more likely to report that the event had a negative effect on their lives (odds ratio, 4.09; 95% CI, 1.72-9.72) (Table 4). Surgeons who reported more negative attitudes toward disclosing an adverse event in a low-harm setting at baseline reported greater anxiety about future outcomes following disclosure (odds ratio, 1.54; 95% CI, 1.16-2.06). There were no significant results reported for any of the remaining items about the personal effect on surgeons.

We found that most of the surgeons in this study reported using 5 of the 8 recommended communication items during disclosure that had been identified as important by surgeons and patients in previous pilot interviews and that were part of the National Quality Forum and VA guidelines. Quiz Ref IDThe 3 items that were not as commonly discussed—apologizing to the patient, discussing whether the event was preventable, and discussing how recurrences of the event could be prevented—are viewed as more challenging communication items, but nonetheless critical elements of disclosure that patients have indicated they prefer to hear.^6,7,15,27-29 To our knowledge, VA surgeons participating in this study had not been trained on disclosure of adverse events. Therefore, it would be surprising if many surgeons were consistently including all 8 recommended items in their disclosures.

Participants indicated that they had difficulty communicating about preventability of the adverse event with patients. Surgeons may not have known whether the event was preventable (since not all reported cases were medical errors) or what steps could be taken to prevent its recurrence. Surgeons may also have thought that discussing prevention and recurrences was too difficult, perhaps because it might suggest blame for the event. However, since these elements of disclosure are specifically mentioned as key elements in guidelines, and patients have indicated wanting to know how an event occurred, not just that it did,^28,29 it is important to know that not all surgeons are disclosing this information. Furthermore, we found that not disclosing this information is associated with surgeons feeling more negatively affected by the adverse event. The Disclosure Training Program, recently disseminated across 26 VA medical centers, would be a next step for training surgeons on how to discuss challenging issues such as preventability of adverse events.³⁰ Creating a culture of professionalism that allows for self-care following disclosures may also increase surgeons’ well-being, as has been done in other large health care systems.³¹

Surgeons who perceived an adverse event to be extremely or very serious also reported being negatively affected by the event. This finding makes sense: in general, surgeons strive for perfection, and an association exists between a perfectionist personality type and choosing surgery as a medical specialty.³² Indeed, when surgeons reported experiencing difficult conversations disclosing adverse events, they also reported being more negatively affected by the event. Surgeons also may not have known what information to disclose, indicating that greater training to support surgeons may be needed.

Surgeons in our study who reported more negative attitudes toward disclosure in a situation where it would be possible to not inform patients or family members of the adverse event (eg, low-harm events where the patient recovers well) were more likely to report anxiety following disclosure. There are certain events that mandate a disclosure, and the high-harm (retained sponge) scenario is one of them. Nearly all surgeons reported that they would disclose this event. In the low-harm (blood loss) scenario, the patient recovered well, and discussing this event may have been perceived as unnecessary by many surgeons in our study. It is possible that many adverse events are also categorized by surgeons as low harm vs high harm. However, if disclosure is deemed necessary by a hospital’s patient safety or risk management team, surgeons with more negative attitudes toward disclosure would find this communication more difficult, and, as in our study, might report greater anxiety about future outcomes or events as a result of having had this unwelcome conversation with patients.

Our study has some limitations. This is not a study of the entire VA surgical service but only of surgeons employed at 3 facilities who voluntarily participated and informed study staff when an adverse event and disclosure occurred. Title 38 of USC § 5705 (2010) protects quality assurance information on adverse events²²; therefore, we were not able to capture the range of adverse events occurring in our study. Moreover, our heterogeneous sample does not allow us to examine how surgeons in certain specialties are disclosing adverse events. Although our 52% response rate is higher than in other published surveys of surgeons (eg, 32%),¹⁷ achieving a greater response rate across more sites is desirable. We would expect that a wider range of surgeons’ experiences with disclosure would be more variable and potentially include more negative events. Another limitation is that although VA physicians are protected under the Federal Tort Claim Act, they are not protected from reporting to the National Practitioner Data Bank. Examining disclosures among surgeons in other health care systems where greater protection for one’s reputation may be provided is another necessary step in this line of research.

Much progress has been made toward physicians responding to demands for transparency and openness with patients, no matter what events take place during their medical care. However, very little has been done, overall, to assess physicians’ experiences with disclosing actual adverse events to patients, a situation that requires skills in immediate, transparent, open communication, and to determine whether these disclosures are following recommended guidelines. By emphasizing the potential for surgeons being negatively affected after adverse events and disclosures, and recognizing the association between attitudes, perceived seriousness of events, surgeons’ experiences with disclosures, and training on how to include specific elements of disclosure in these difficult conversations, future quality improvement efforts may be able to help sustain the implementation of open disclosure programs nationwide^33,34 while also ensuring a healthy surgeon workforce.

Accepted for Publication: May 3, 2016.

Corresponding Author: A. Rani Elwy, PhD, Center for Healthcare Organization and Implementation Research, Veterans Affairs Boston Healthcare System, Mailstop 152, 150 S Huntington Ave, Boston, MA 02130 (rani.elwy@va.gov).

Published Online: July 20, 2016. doi:10.1001/jamasurg.2016.1787

Author Contributions: Dr Elwy had full access to all 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: Elwy, Itani, Bokhour, Mueller, Glickman, Rosen, Brotschi, Sanchez, Gallagher.

Acquisition, analysis, or interpretation of data: Elwy, Itani, Bokhour, Mueller, Glickman, Zhao, Lynge, Perkal, Sanchez, Gallagher.

Drafting of the manuscript: Elwy, Mueller, Glickman, Gallagher.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Mueller, Glickman, Zhao.

Obtained funding: Elwy, Bokhour, Glickman, Rosen.

Administrative, technical, or material support: Elwy, Itani, Mueller, Lynge, Perkal, Sanchez.

Study supervision: Elwy, Mueller, Lynge, Gallagher.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was funded by grant IIR 07-199 from the Department of Veterans Affairs, Health Services Research and Development Service.

Role of the Funder/Sponsor: The funding source approved the design and conduct of the study. It had no role in the collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent those of the Department of Veterans Affairs or the US government.

Additional Contributions: We thank the many surgeons who participated in this study for trusting us with their information and for agreeing with us that these study questions were worth pursuing.