Effects of Patient Suicide on Professional Practice Among Mental Health Providers

Abstract

Introduction.

The loss of a patient to suicide has an enormous impact on clinicians, but few studies have examined its effects.

Method.

In this retrospective study, we compared clinicians who have and have not experienced a patient suicide using a survey of 2,157 outpatient clinicians from 169 New York clinics to determine differences in their suicide prevention knowledge, practices, training, and self-efficacy.

Results.

Approximately 25% of the clinician respondents lost patients to suicide; psychiatrists, nurses/nurse practitioners, and those with more years of experience were disproportionately affected. After controlling for these demographic/professional differences, clinicians who had experienced patient suicide reported feeling that they had insufficient training, despite actually having more suicide prevention training, greater knowledge of suicide prevention practices, and feeling more comfortable working with suicidal patients than clinicians who had not lost a patient to suicide. There were no differences in self-efficacy or utilization of evidence-based clinical practices.

Conclusions.

Controlling for demographic/professional differences, clinicians who experienced a patient suicide had more training, knowledge, and felt more comfortable working with suicidal patients. It is critical that sufficient training be available to clinicians, not only to reduce patient deaths, but also to help clinicians increase their comfort, knowledge, skill, and ability to support those bereaved by suicide loss.

Suicide rates have increased by over a third since 1999.¹ Over 48,000 individuals died by suicide in the United States in 2018, making suicide the tenth leading cause of death and a significant contributor to recent reductions in life expectancy.¹ For every patient who dies by suicide, there is also a treating clinician who becomes a suicide loss survivor. Death of a patient by suicide is an event that many clinicians experience.^2,3 While physicians may lose patients during surgery or due to serious illness, mental health clinicians may have an amplified reaction to the loss of a patient due to the unique emotionally vulnerable relationships they share, and the complicated and devastating nature of suicide in particular.⁴ The mourning process faced by any suicide survivor may include feelings of shock, trauma, guilt, overwhelming grief, and complex emotions.⁵ For clinicians, this process may be no less emotionally difficult and may also be further compounded by adverse professional consequences.⁴ The extent to which a clinician is able to overcome these dual professional and personal impacts and come to terms with the patient’s suicide likely influences his or her professional identity, relationships, and clinical interactions, and it also ultimately impacts the clinician’s comfort remaining in the field and treating other high-risk patients.⁵

A survey of psychiatrists found that, when asked to describe losing a patient by suicide, all 120 respondents reported painful emotions, such as sadness (71%), anxiety, worry, and fear (33%), guilt and self-blame (31%), regret (20%), and anger (19%).³ Furthermore, between 40% and 98% of those who have had a patient die by suicide report that the experience influenced their career,^2,6 with 40% of psychiatrists reporting that they struggled with clinical work for 1 week to 6 months after the suicide, 21% for 6 months to 2 years, and 13% experiencing ongoing difficulties.³ Common clinician reactions to a patient’s death by suicide include increased awareness of suicide risk, proactiveness in risk assessment, and increased caution when working with suicidal patients.⁷ Many clinicians reported more frequent consultations with colleagues and engagement in more detailed clinical record-keeping following a patient’s suicide.⁸ An increased tendency to refer patients for hospitalization,⁹ concerns about one’s professional competency,^7,8 subsequent reluctance to accept suicidal patients into one’s practice,^7,10 and greater attention to legal liability were also reported.^8,9

Given the potential personal and professional impact experienced by clinicians who lose patients to suicide, it is important to better understand their experiences. However, inconsistencies exist across previous studies with respect to patient suicide rates and the degree of impact on clinicians, which are likely attributable to differences in measures, samples, and settings. Furthermore, much of the research on clinician loss survivors is dated and may not accurately reflect the experiences of clinicians treating the increasing number of suicidal patients over the past 2 decades. In addition, much of this research focused exclusively on certain professions and may not generalize to all mental health professionals, especially those in outpatient settings. To this end, this study used survey data from over 2000 outpatient mental health providers across New York State (NYS), collected before the implementation of a statewide suicide prevention quality improvement initiative,¹⁰ to examine clinician characteristics and suicide prevention knowledge, training, and clinical practices associated with the experience of a patient suicide.

METHOD

Setting and Respondents

In 2016, the NYS Office of Mental Health Bureau of Evidence-Based Services and Implementation Science initiated a large-scale Zero Suicide quality improvement project, in partnership with the Research Foundation for Mental Hygiene, the NYS Office of Mental Health Suicide Prevention Office, the NYS Psychiatric Institute, and Columbia University Vagelos College of Physicians and Surgeons. Participating programs included 169 outpatient behavioral health clinics across NYS, representing 86 agencies and serving approximately 90,000 patients. All clinician respondents were employed by clinics enrolled to participate, but data were collected before the start of the project. Of the 3715 clinicians working in participating clinics, 2257 submitted the survey (60.8%), and 2151 provided data on their experiences with patient suicide (57.9% of all clinicians in the clinics and 95.3% of the 2257 respondents). The sample was predominantly female (76.4%), evenly distributed by age group, and reflective of the ethnic and racial composition of NYS (12.5% Latinx and 22.2% racial minority; see Table 1). The sample consisted mostly of masters-level social workers and mental health counselors (64.5%), with other clinical disciplines (eg, psychiatrists, psychologists, nurses) evenly represented. Nearly half the sample had worked in behavioral health for >10 years, with 75% in direct service roles, 13% in management, and 9% providing clinical supervision.

Table 1.

Characteristics of the total sample and of those who had and had not experienced a patient suicide death

	Total (N=2151) % (95% CI)	Patient Suicide Death		B (SE)	Wald	df	p	OR	95% CI
		Yes (n=535) % (95% CI)	No (n=1616) % (95% CI)
Gender					7.14	2	0.03
Female⧫	76.4% (74.6–78.2%)	68.5% (64.6–72.4%)	78.9% (76.9–80.9%)	-	-	-	-	-	-
Male	22.7% (20.9–24.5%)	30.9% (27.0–34.8%)	20.1% (18.1–22.1%)	0.35 (0.13)	7.14	1	0.01	1.42	1.10–1.84
Transgender or other	0.9% (0.5–1.3%)	0.6% (0–1.3%)	0.9% (7.6–10.4%)	0.03 (1.25)	0.001	1	0.98	1.03	0.09–11.99
Age (y)					0.31	2	0.86
≤ 34	27.3% (25.4–29.2%)	13.9% (11.0–16.8%)	32.0% (29.7–34.3%)	−0.07 (0.21)	0.10	1	0.75	0.93	0.62–1.42
35–54	43.1% (41.0–45.2%)	45.5% (41.3–49.7%)	42.4% (40.0–44.8%)	0.03 (0.13)	0.06	1	0.81	1.03	0.80–1.34
55+⧫	29.6% (27.7–31.5%)	40.6% (36.4–44.8%)	25.6% (23.5–27.7%)	-	-	-	-	-	-
Race					12.32	4	0.02
White⧫	77.9% (76.1–79.7%)	81.8% (78.5–85.1%)	76.5% (74.4–78.6%)	-	-	-	-	-	-
Black/African American	10.2% (8.9–11.5%)	6.6% (4.5–8.7%)	11.4% (9.9–12.9%)	−0.74 (0.22)	11.70	1	0.001*	0.48	0.31–0.73
Asian/Asian-American	5.3% (4.4–6.2%)	7.4% (5.2–9.6%)	4.5% (3.5–5.5%)	−0.16 (0.26)	0.39	1	0.54	0.85	0.51–1.42
More than one race	3.4% (2.6–4.2%)	2.4% (1.1–3.7%)	3.8% (2.9–4.7%)	−0.17 (0.38)	0.20	1	0.66	0.85	0.41–1.77
Other	3.2% (2.5–3.9%)	1.8% (0.7–2.9%)	3.8% (2.9–4.7%)	−0.38 (0.44)	0.74	1	0.39	0.68	0.29–1.63
Ethnicity
Hispanic or Latinx	12.5% (11.1–13.9%)	7.1% (4.9–9.3%)	14.4% (12.7–16.1%)	−0.54 (0.25)	4.75	1	0.03	0.58	0.36–0.95
Not Hispanic or Latinx⧫	87.5% (86.1–88.9%)	92.9% (90.7–95.1%)	85.6% (82.9–87.3%)	-	-	-	-	-	-
Discipline					49.18	4	<0.001*
Social worker or counselor⧫	64.5% (62.5–66.5%)	52.7% (48.5–56.9%)	68.5% (66.2–70.8%)	-	-	-	-	-	-
Psychologist	5.4% (4.4–6.4%)	6.5% (4.4–8.6%)	5.1% (4.0–6.2%)	0.16 (0.23)	0.52	1	0.47	1.18	0.76–1.84
Nurse or nurse practitioner	9.8% (8.5–11.1%)	14.2% (11.2–17.2%)	8.4% (7.0–9.8%)	0.83 (0.18)	21.30	1	<0.001*	2.29	1.61–3.25
Psychiatrist	7.5% (6.4–8.6%)	17.0% (13.8–20.2%)	4.3% (3.3–5.3%)	1.25 (0.21)	35.49	1	<0.001*	3.50	2.32–5.28
Other	12.7% (11.3–14.1%)	9.5% (7.0–12.0%)	13.7% (12.0–15.4%)	0.23 (0.20)	1.34	1	0.25	1.26	0.85–1.85
Years working in behavioral health					63.18	2	<0.001*
≤ 3 years	20.0% (18.3–21.7%)	5.2% (3.3–7.1%)	25.1% (23.0–27.2%)	−1.87 (0.25)	56.84	1	<0.001*	0.15	0.10–0.25
4–10 years	31.7% (29.7–33.7%)	23.7% (20.1–27.3%)	34.3% (32.0–36.6%)	−0.81 (0.15)	28.27	1	<0.001*	0.44	0.33–0.60
11+ years⧫	48.3% (46.2–50.4%)	71.1% (67.3–74.9%)	40.6% (38.2–43.0%)	-	-	-	-	-	-

Note:

^⧫Reference category for logistic regression.

^*Statistically significant after modified Bonferroni correction.

Measures

The Zero Suicide Workforce Survey (ZSWS)¹² is an 80-item measure created for the Zero Suicide project. It assesses clinician demographics, exposure to patient suicide, formal training in suicide prevention, knowledge of suicide prevention best practices, clinical assessment and intervention practices, intentions to utilize suicide prevention clinical practices, self-efficacy for engaging in suicide prevention clinical practices, comfort working with suicidal patients, suicide prevention attitudes, and organizational climate (copy of measure available by request). The ZSWS was adapted from a previous measure developed by the National Action Alliance for the Zero Suicide Initiative,¹³ with additional items added based on the Theory of Planned Behavior^14,15 and theoretical factors associated with successful implementation¹⁶ (see citation 12 for more information). Items assessing exposure to suicide, training, knowledge, self-efficacy, comfort working with suicidal patients, and clinical practices are presented here.

Exposure to suicide was assessed by 2 questions asking whether the respondent had: 1) worked directly with a patient who died by suicide during or after treatment and 2) worked directly with a patient who made a suicide attempt during treatment, with response options of: No; Yes–One patient; or Yes–More than one patient. Two follow-up questions assessed the respondent’s perceptions of support and/or blame on a Likert scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). Training was assessed by seven dichotomous (yes/no) items measuring formal training in different suicide prevention techniques (screening, risk assessment, safety planning, means reduction counseling, suicide-specific treatment planning, evidence-based suicide-specific psychotherapies, and monitoring procedures), which were summed into a single score measuring the number of domains on which the respondent reported being trained (α= 0.90). A dichotomous item (yes/no) also measured perceived sufficiency of training. Knowledge was assessed using 5 multiple-choice items evaluating knowledge of best practices for suicide screening, risk assessment, safety planning, interventions, and monitoring, with responses averaged into a total knowledge score (α = 0.77). Self-efficacy was assessed with 7 items measuring confidence in one’s ability to deliver suicide-specific clinical practices on a Likert scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree), averaged into a total self-efficacy score (α = 0.90). Comfort working with suicidal patients was assessed with a single question (“You feel comfortable working with suicidal clients”) using a Likert scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). Clinical practices were assessed using 17 items measuring frequency of use of specific assessment and intervention practices on a Likert scale ranging from 1 (Never) to 5 (Always), averaged to derive scores for Assessment (7 items; α = 0.82) and Interventions (10 items; α = 0.88).

Procedure

To obtain a baseline measure of suicide training, knowledge, self-efficacy, and clinical practices, all clinicians working in participating clinics were asked to complete the ZSWS before they received any introduction to the statewide suicide prevention initiative or the onset of training and implementation activities. No incentives were offered; rather, project staff contacted clinicians directly with a link to the survey via personalized emails and followed up with supervisors to encourage participation. The survey was administered online via Qualtrics and took approximately 30 minutes to complete. Clinicians were informed that responses would be de-identified and aggregated. The Institutional Review Boards of the NYS Psychiatric Institute and Nathan Kline Institute reviewed the project and determined that the study did not meet the definition of human subjects research.

Data analysis was conducted in SPSS version 25. Descriptive results are reported as either proportions and 95% confidence intervals (CIs) or means and standard deviations. Logistic regression analyses were performed to determine if sociodemographic variables (gender, age, ethnicity, race, discipline, or years working in behavioral health) were associated with the experience of patient suicide. Separate mixed-effect linear regression models were fitted with a number of dependent variables. These variables included training (number of clinical practices in which the participant received formal training), knowledge (total suicide prevention knowledge score combined across domains), self-efficacy (total self-efficacy score combined across suicide prevention clinical practices), and utilization of evidence-based clinical practices (assessment and intervention scores). Separate mixed-effect linear regression models were also fitted with whether or not a clinician experienced a patient suicide as a fixed-effect predictor, demographic factors that differed between groups covaried, and agency-specific random intercepts to account for within-agency clustering (approximately 2.5% to 4% of the variance across variables). Hedge’s g was reported as an indicator of effect size appropriate for comparing different sized groups, and all analyses utilized a modified Bonferroni correction to adjust for multiple comparisons.

RESULTS

Characteristics of the total sample and of those who had and had not experienced a patient suicide death are reported in Table 1. A quarter of clinicians reported at least one patient death by suicide during or after treatment with them, including 8% reporting more than 1 patient death by suicide. Furthermore, 58% of clinicians reported at least 1 patient attempting suicide during treatment, with 37% reporting more than 1 patient. Nearly 70% reported being supported by their clinic through this experience, whereas 9% did not feel supported (21% neutral); 8% reported feeling blamed.

Logistic regression revealed significant sociodemographic effects between clinicians who had and had not experienced a patient suicide, X² (df = 15, N = 1837) = 256.45, P < 0.001, with sociodemographic variables explaining between 13.0% (Cox and Snell R square) and 19.1% (Nagelkerke R-squared) of the variance and correctly classifying 75.6% of cases. (Note that sociodemographic data were available for 1837 of the 2151 respondents who provided data on their experiences with patient suicide.) When controlling for all other factors in the model, the strongest predictors of experiencing a patient suicide were professional discipline and years working in behavioral health. Psychiatrists (57%; OR = 3.50) and nurses and nurse practitioners (36%; OR = 2.29) were more likely to have experienced a patient suicide than psychologists (30%), masters-level social workers and mental health counselors (20%), and other mental health professionals (19%). Those with fewer years of work experience in behavioral health (≤ 3 years: 7%, OR = 0.15; 4–10 years: 19%, OR = 0.44) were less likely to have experienced a patient suicide compared with those with more than 10 years of experience (37%). Black/African American clinicians (17%; OR = 0.48) were less likely to have experienced a patient suicide than white clinicians (27%). No significant ethnic, age, or gender differences were found (Table 1).

Descriptive statistics are presented in Table 2. Controlling for demographic/professional factors that differed between groups, clinicians who experienced a patient suicide death were more likely to report that their training in suicide prevention was insufficient (F_1,1915 = 19.87, P < 0.001, Hedge’s g = 0.24); however, they also reported training in a significantly greater number of suicide prevention techniques than clinicians who had not experienced a patient suicide (F_1,1923 = 19.63, P < 0.001, Hedge’s g = 0.23). Clinicians who experienced a patient suicide death also reported significantly greater suicide prevention knowledge (F_1,1925 = 18.81, P < 0.001, Hedge’s g = 0.23) and comfort working with suicidal patients (F_1,1925 = 12.48, P < 0.001, Hedge’s g = 0.20) than clinicians who had never experienced patient suicide. There were no group differences in self-efficacy for delivering suicide prevention clinical practices (F_1,1726 = 1.72, P = 0.19, Hedge’s g = 0.08) or utilization of evidence-based assessment (F_1,1729 = 3.18, P = 0.08, Hedge’s g = 0.09) or intervention practices (F_1,1722 = 0.19, P = 0.67, Hedge’s g = 0.03).

Table 2.

Mean self-reported suicide prevention training, knowledge, comfort treating suicidal patients, self-efficacy, and utilization of evidence-based clinical practices, grouped by those who had and had not experienced a patient suicide death.

	Min / Max	Total Mean (SD) or % (95% CI)	Experienced Patient Suicide Death Mean (SD) or % (95% CI)		F
			Yes	No
Suicide prevention training
Number of domains trained	0 / 7	5.35 (2.21)	5.72 (1.98)	5.22 (2.27)	19.631***
Sufficient training to assist suicidal patients	-	48.7% (46.6–50.8%)	40.7% (36.5–44.9%)	51.7% (49.3–54.1%)	19.87***
Suicide prevention knowledge	0 / 100%	74.8% (73.0–76.6%)	77.4% (73.9–80.9%)	73.8% (71.8–75.9%)	18.81***
Comfort treating suicidal patients	1 / 5	3.88 (0.90)	4.01 (0.89)	3.83 (0.90)	12.48***
Self-efficacy	1 / 5	4.24 (0.66)	4.27 (0.63)	4.22 (0.67)	1.72
Utilization of evidence-based clinical practices
Assessment	1 / 5	4.21 (0.79)	4.26 (0.77)	4.19 (0.80)	3.18
Interventions	1 / 5	4.04 (0.69)	4.03 (0.68)	4.05 (0.69)	0.19

^***P < 0.001; statistically significant after modified Bonferroni correction.

DISCUSSION

The current study found that approximately 25% of participating clinicians in outpatient behavioral health settings had lost patients to suicide. Of these 25% of clinicians affected by patient suicide, 9% did not feel supported by their clinic and 8% reported feeling blamed. Despite feeling that their suicide prevention training was insufficient, clinicians who had experienced a patient suicide had formal training in a greater number of suicide prevention techniques than clinicians who had never experienced a patient suicide. Clinicians who reported losing patients by suicide displayed greater suicide prevention knowledge and were more comfortable working with suicidal patients; however, they did not differ in self-efficacy or utilization of evidence-based assessment and intervention practices. Relative to other mental health disciplines, psychiatrists, nurses, and nurse practitioners disproportionately experienced patient suicide. These higher rates may be attributed to the fact that these professions may be more likely to work in a hospital setting, have larger caseloads, or treat high-risk patients with more severe psychopathology; in addition, suicidal patients may be more likely to be treated with psychotropic medications. In addition, those with more than 10 years of experience working in behavioral health disproportionately experienced higher numbers of patient suicide, which is likely attributable to a higher probability of losing a patient over time.

A possible explanatory theory for higher levels of training, knowledge, and comfort reported by clinicians who lost patients to suicide is that of “posttraumatic growth.” Introduced by Tedeschi & Calhoun,¹⁷ posttraumatic growth refers to the idea that a traumatic experience can present an opportunity for critical self-reflection and self-improvement. Gutin et al⁵ drew parallels between the trajectory of bereavement due to patient suicide and the process leading to posttraumatic growth, noting that various models detailing the sequelae of patient suicide among clinicians^18–20 suggest that resolution is often achieved through self-reflection, acknowledgment of individual responsibility, and acceptance of loss. As in posttraumatic growth, such outcomes constitute meaningful and constructive personal and professional development. Preliminary empirical support for the concept of posttraumatic growth is provided by studies where clinician-survivors who continue to treat suicidal individuals self-report enhanced knowledge of interventions and ability to assess risk and protective factors following a patient suicide.⁵

The substantial number of mental health providers affected by patient suicide suggests that this experience is not uncommon in outpatient behavioral health settings. The current findings also suggest that patient suicides do not only occur among less knowledgeable or novice clinicians. In fact, clinicians who lost patients to suicide were no different in their self-efficacy or utilization of clinical best practices, and they demonstrated superior knowledge, training, and comfort working with suicidal patients compared with peers who had never experienced a patient suicide. In light of the high percentage of clinicians who experience patient suicide, sufficient training in suicide prevention approaches must be available to clinicians and trainees. In addition, clinics, institutions, and training programs should have postvention policies in place to support and guide clinicians of all disciplines and training levels if a patient suicide should occur. Unfortunately, few protocols exist on how best to help clinician survivors, and clinics and training programs generally lack evidence-based postvention policies.^21–23

Guidelines are available for clinics interested in implementing postvention protocols,^23–25 and they typically recommend that psychological first aid (ie, support, reassurance, information, and time) should precede exhaustive case reviews even if there are concerns about liability and/or clinical conduct.²⁶ Prematurely focusing on administrative and legal issues may adversely affect clinicians, and the complex emotional reactions associated with a recent patient suicide loss can impair clinicians’ abilities to carry out critical case reviews.^23,27 Guidelines also recommend making more intensive services available to clinicians (ie, consultation, education, support, and resources offered by the American Association of Suicidology Clinician Survivor Task Force²⁴), but not presuming that they will be required.²⁵ Only after sufficient time for bereavement has passed should efforts be made to enhance clinical training and patient services,²⁸ including provision of suicide prevention training and considering what may have “gone wrong” via psychological autopsy or critical case review. Appropriate postvention helps clinicians not only to effectively implement future suicide prevention activities, but also to endure loss, reduce self-blame, and maintain a strong sense of professional self-efficacy.²⁰

Study Limitations

It is important to note that this sample was limited to clinicians who remained in the field and continued practicing after patient suicide loss. Given that significant numbers of clinicians transition away from clinical practice following a patient suicide death,^3,5,19 the current data may be skewed toward more positive findings. Likewise, it is possible that the composition of the sample (ie, currently licensed, practicing mental health clinicians) may have influenced results relative to other populations (eg, trainees, non-psychiatric physicians). In addition, the setting in which clinicians completed the survey may have influenced results. More specifically, completing the survey as part of a work initiative may have led some clinicians to provide higher ratings than they might have otherwise. However, it is unclear if this tendency toward higher ratings would differ between groups, and the likelihood of bias was somewhat mitigated as respondents were informed that data would be de-identified and only shared in aggregate form, hopefully encouraging them to be honest in their responses.

Moreover, while analyses controlled for pre-existing demographic and professional factors, the survey data are cross-sectional, meaning there is no way to determine if higher levels of knowledge, training, and confidence preceded or were the result of experiencing a patient suicide loss. It is possible that clinicians with more training and knowledge may receive more referrals or be more likely to treat high risk patients, thereby increasing the likelihood of a patient death by suicide. Alternatively, clinicians who lose a patient to suicide may seek out additional training, translating into greater knowledge and comfort working with suicidal patients and perhaps helping them to overcome loss. Despite experiencing the death of a patient by suicide, clinicians remained in practice and reported self-efficacy and utilization of evidence-based suicide prevention clinical practices at comparable levels to their peers who never lost a patient, demonstrating that the loss of a patient does not necessarily diminish clinicians’ willingness or ability to treat suicidal patients.

CONCLUSIONS AND FUTURE DIRECTIONS

The literature on this topic is limited, and much of the existing research is qualitative and/or dated. The findings presented here from a large sample of behavioral health clinicians suggest that psychiatrists, nurses, and nurse practitioners disproportionately experience patient suicide, and that risk for experiencing a patient suicide increases with more years of experience in the field. When controlling for discipline and years of experience, those who have experienced a patient suicide and continue to provide clinical services are more comfortable treating suicidal patients and have more suicide prevention training and knowledge than clinicians who have not experienced a patient death by suicide. Longitudinal studies of mental health providers are needed to clarify the temporal relationship between patient suicide and the acquisition of suicide prevention training, knowledge, and increased comfort working with suicidal patients. Ultimately, it is critical that sufficient training be available to clinicians, not only to reduce patient deaths, but also to help support clinicians of all disciplines and training levels through suicide loss, and to increase their comfort, knowledge, and skill for addressing patient suicide risk.