Post-vention, or the intervention conducted after a suicide, is a practice that has not generally been part of a formal curriculum of most psychiatric residencies. In 1988, Ellis et al. conducted a survey about post-vention practices from program directors of general and child/adolescent psychiatric residencies (1). The survey found that 18.6% of programs had post-vention procedures specified in the policies-and-procedures manual. Also, 46.5% of programs had training activities that instructed residents about what to do should a patient commit suicide. More recent surveys have found similar results. One-third of Canadian programs have a policy in place for debriefing a resident after a patient suicide (2), and one-third of American programs have a risk-management policy regarding patient suicide (3).

In didactic settings, suicide topics such as risk factors, early warning signs and recognition, standards of clinical care, and ethics of hospitalization are taught in over 90% of programs, whereas post-vention is only taught in one-quarter of programs (4). A survey of chief residents found that only 19% felt prepared for the possibility of having to manage the aftermath of a patient suicide (4). Given that anywhere from 14% to 68% of residents experience a patient suicide sometime during residency (2, 5–7), one might conclude that post-vention is a topic that has been underrepresented in training programs.

With few programs offering formal clinical exposure and didactics about post-vention and so few residents feeling prepared for the aftermath of a patient suicide, residents and program directors have been found to have differing attitudes about post-vention. A survey conducted by Digiovanni et al. (3) found that 61% of program directors felt programs dealt effectively with the impact of patient suicides on residents. Only 32% of residents felt similarly.

Objectives, Methods, and Procedures

We created an online survey for psychiatric chief residents and program directors in the United States. The survey asked about trainees' experiencing patient suicide and subsequent post-vention practices in 2008. The content of the survey was largely modeled after the Ellis et al. survey in 1988, which asked program directors about residency size, the existence of a written post-vention protocol or guideline, and whether certain post-vention activities were recommended or required. The findings of this online survey would help determine whether there has been change in residency policies regarding post-vention from 1994 to 2008.

This study differed from the Ellis et al. survey in that it also surveyed chief residents, allowing a comparison of Chief Resident and Program Director reports. The survey also asked for the total number of suspected or confirmed suicides for each program in 2008 as well as the post-vention activities that were actually implemented. This information would allow us to look at programs with and without protocols and determine whether this had an effect in actual post-vention. It was hypothesized that programs with protocols would have a higher rate of post-vention.

The protocol for this study was submitted to the Department of Clinical Investigation at Walter Reed Army Medical Center and was granted an exemption from IRB Review. A chief resident e-mail list for all U.S. psychiatric residencies was maintained by the American Psychiatric Association and was available to the primary investigator of this study. This list included the e-mails of most of the general and child/adolescent psychiatric chief residents. The e-mail information for program directors of general and child/adolescent psychiatric residencies were publicly available and were obtained primarily through the websites www.acgme.org/adspublic/ and www.residencyplace.com.

Chief Residents and Program Directors were contacted via e-mail in February 2009. The e-mail explained the survey and its purpose, as well as the voluntary, anonymous nature of the survey. A link in the e-mail took the participant to a webpage at www.surveymonkey.com, where the participant could respond. Follow-up e-mail reminders were sent every week for 3 more weeks to those who did not respond promptly.

Over the 4-week collection period, 54 of 257 Chief Residents responded (21.0%) and 94 of 302 Program Directors responded (31.1%). Of the Program Directors, 57 of 183 general program directors responded, and 37 of 119 child/adolescent program directors responded. Of the 54 Chief Resident responses, there was an average of 26.8 residents per program, as compared with the 94 Program Director responses, for an average of 19.0 residents per program.

Chief Residents reported an average of 1.44 suicides per residency, or 5.4% of residents experiencing a patient suicide in the year 2008. The numbers reported from Program Directors were 0.88 and 4.7%, respectively. The majority of programs as reported by Program Directors had no suicides (49/94), whereas only a slight plurality of programs as reported by Chief Residents had no suicides (18/54).

In terms of percentage of programs having a written post-vention protocol, both Chief Residents and Program Directors reports were similar, at 20% and 21%, respectively. However, Chief Residents reported a lower percentage of written requirements than Program Directors and recommendations for timely notification of the Program Director, case-conferences, or mortality-and-morbidity conferences, process sessions, and psychological autopsies. This trend also continued in actual post-vention in 2008, as Chief Residents reported a lower percentage of supervision sessions assessing emotional impact, therapy, or counseling, case-conferences, or mortality-and-morbidity conferences, and psychological autopsies than Program Directors.

Comparing Chief Residents from programs with versus those without written protocols, Chief Residents from programs with protocols reported a higher rate of timely notifications of the Program Director, process sessions, therapy, or counseling, and emergency leave after a patient suicide. These differences were statistically significant by two-sided Fisher's exact tests. Comparing Program Director responses from programs with versus those without protocols, Program Directors from programs with protocols reported higher rates of timely notifications of the Program Director, case-conference, or mortality-and-morbidity conferences, process sessions, and supervision sessions assessing emotional impact. These differences, however, did not reach statistical significance.

Further statistical analysis found that, according to Chief Residents and Program Directors, there was a tendency for programs with a post-vention protocol to have more patient suicides. According to chief residents, of the 11 programs with a post-vention protocol, 9% had no suicides; 36% had one suicide; and 55% had two-or-more suicides. Of the 39 programs without a post-vention protocol, 36% had no suicides; 23% had one suicide; and 41% had two-or-more suicides. This trend was not statistically significant by Pearson's chi-square test.

However, when analyzing Program Directors' responses comparing programs with and without protocols, the results were statistically significant. Of the 20 programs with a post-vention protocol, 45% had no suicides; 10% had one suicide; and 45% had two-or-more suicides. Of the 73 programs without a post-vention protocol, 55% had no suicides; 33% had one suicide; and 12% had two-or-more suicides. Pearson's chi-square test (χ²)=11.88, with a significance level of p=0.003.

The low response rate in the current survey limits a meaningful comparison with the Ellis et al. survey, which had 166/296 (56.1%) program directors responding. Ellis et al.'s method of using mailed packets may account for the higher response rate than the current study, which relied on e-mail communication. The collection period for the two surveys likely also played a role in response rate, as the current study was limited to 4 weeks, whereas the data collection period for the Ellis et al. study was not specified.

Despite these limitations, comparisons with the survey from 1988 would indicate that there has not been substantial change in the number of programs documenting post-vention procedures in residency policies-and-procedures manuals over the last 14 years. The number of programs with written protocols remains approximately 1 in 5, and the percentage of recommendations and requirements for certain post-vention activities may even have decreased. In the survey by Ellis et al., timely notification of the program director (66.5%), quality assurance, or mortality-and-morbidity review (58.9%), supervision regarding emotional impact (76.0%), and counseling (46.8%) were required and recommended at much higher rates than in the current survey. A possible explanation for this stagnation could be the lack of consensus in the psychiatric community on how to teach and implement post-vention in residencies, and, therefore, a concomitant shift away from these post-vention activities. The fact that the process may be further complicated by participation from multitudes of third parties, including hospital administration, public affairs, legal affairs, malpractice insurance, family members, and peer-to-peer reviews only makes it more difficult to develop a standard post-vention strategy that includes elements that are useful, fair, and supportive.

With 5% of residents experiencing a patient suicide in 2008, this suggests that 20% of residents would experience a patient suicide over a full 4-year residency, which would be on the lower end, but consistent with what was seen in previous reports (2, 5–7). This study may have had a selection bias for a lower patient suicide rate because chief residents and program directors may not have been completely aware of all the suicides that go on in an entire residency, especially in residencies where residents rotate through multiple hospital systems.

With program directors reporting no suicides in their program more often than chief residents, there may have been another selection bias, where chief residents from programs with a patient suicide were more likely to respond, and program directors from programs with one-or-more patient suicides were less likely to respond. Chief residents may have been disgruntled by the lack of adequate post-vention in their residency and were more motivated to respond, whereas program directors who experienced suicides in their residencies may have opted not to respond because it would require them to report unfavorable information.

With chief residents reporting a smaller percentage than program directors on both the written requirements and recommendations as well as the actual implementation of post-vention, one might consider that chief residents may not have been as aware of existing written policies and protocols and/or may not have been privy to the implementation of post-vention activities. Program directors who were less comfortable in dealing with the fallout of a suicide may have manifested a more “paranoid” culture where information about the event was kept out of the public eye, and the topic was avoided. Hence, chief residents would have been less likely to be included in the process. Also, chief residents from programs that did not require or recommend post-vention activities may have been more likely to respond to the survey, again, as a result of dissatisfaction with the level of post-vention support from their programs.

Discrepancies between chief residents' and program directors' reports might suggest that many of the survey findings are invalid. However, one might take the findings to suggest that chief residents and program directors perceive the post-vention process differently, as was found in the Digiovanni survey in 2008 (3). Program directors may overestimate and chief residents may underestimate the training program's ability to deal effectively with the impact of patient suicide on the resident.

The hypothesis that having a post-vention protocol would increase the rate of actual post-vention activities within a program was supported by the survey results. However, the results were only statistically significant for four post-vention activities and only for the survey results from chief residents. When the presence of a protocol was compared with patient suicides, a new pattern emerged. The findings from the program directors' responses suggest that programs with a higher frequency of suicides may have adopted post-vention protocols to improve residency training and support. Residents and staff may have experienced past suicides with antagonizing perspectives, which hurt the morale and cohesion of a residency, creating strong emotional countertransference. The protocols may have been developed to address administrative needs of the programs, and educational, training, and emotional needs of the residents. Future efforts should be made in developing content for a post-vention protocol and curriculum that is standardized and widely accepted by the academic-psychiatry community, such as those proposed by Lomax (8) and Brown (6).