Abstract
Objective:
The authors aimed to assess clinicians’ attitudes toward suicide-related practices and their implementation, across roles and settings, before implementation of the Zero Suicide model in a health care system.
Methods:
Clinicians (N=5,559) were invited to complete a survey assessing demographic characteristics; confidence and self-reported suicide-related practice; leadership buy-in; and attitudes toward suicide prevention, safety planning, and continuous quality improvement (CQI).
Results:
Of 1,224 respondents, most felt confident conducting suicide screening but less confident performing other suicide-related care. Provider role and care setting were significantly associated with confidence (p<0.001, Kruskal-Wallis H test) and practice (p<0.001, Kruskal-Wallis H test) of providing suicide prevention care, with behavioral health providers and providers in the emergency department (ED) reporting the highest confidence. Attitudes toward safety planning were more positive among women (p<0.001, t test) and behavioral health providers (p<0.001, F test) than among their counterparts or peers. Positive attitudes toward CQI were significantly associated with male sex (p=0.01), non-White race (p=0.03), younger age (p=0.02), fewer years working in health care (p<0.001), administrative role (p<0.001), working in the ED (p<0.001), outpatient settings (p<0.02), and medical provider role (p<0.001).
Conclusions:
Behavioral health providers and those in the ED reported feeling prepared to deliver suicide-related care, with nurses feeling less confident and less supported. Initiatives to improve suicide-related care should account for clinical role and care setting during planning. CQI could help engage a broader range of clinicians in suicide-related care improvements.
Most people who die by suicide visit health services in the year before death (1, 2), often for nonpsychiatric reasons. Although most suicide decedents visit primary care shortly before death, acute care settings such as the emergency department (ED) and psychiatric inpatient units have a high number of patients with suicide risk (3). The Zero Suicide model is intended to detect and address suicide risk in health care settings (4). By engaging leadership, conducting training, and applying continuous quality improvement (CQI) strategies, health systems implementing Zero Suicide aim to better identify, engage, treat, and transition individuals at risk for suicide. The successful implementation of such a program depends in part on clinicians’ attitudes toward this public health issue, proposed interventions, and implementation strategies (5, 6).
Previous research has found varying levels of support for suicide prevention in health care settings. Several studies have reported better attitudes toward suicide prevention among women (7, 8), older individuals (8, 9), and those who had more training (9, 10). Professionals who have worked in psychiatric settings tend to have better suicide prevention attitudes than those working in nonpsychiatric settings (11–13). However, no previous study has sought to measure attitudes among all clinicians in the same health care system, regardless of setting or role.
The System of Safety Study (14) applied the Zero Suicide model to an entire health care system, comprising five hospitals and nine facility-based ambulatory clinics. The health care system provided both general medical and psychiatric services and served both pediatric and adult patients. This article describes the results of a baseline survey measuring clinicians’ attitudes toward suicide prevention, suicide-related practices, CQI, and leadership support across clinical roles and settings.
Method
Setting
The survey setting was a large nonprofit health care system in Massachusetts. At the time of this survey, the system had four hospitals, with six EDs (four general, one psychiatric, and one pediatric), 21 medical-surgical inpatient units, eight intensive care units, four inpatient psychiatric units, and eight facility-based primary care clinics with integrated behavioral health. Each year, the system has approximately 220,000 ED presentations, 45,000 inpatient discharges, and 1.5 million outpatient visits. All clinicians (N=5,559) working full-time across the health care system were invited to participate. A clinician was considered any health care professional responsible for suicide-related care processes within their clinical setting across a range of qualifications, including physicians, nurses, psychologists, social workers, residents, advanced practice providers, and medical assistants.
Procedure
After receiving institutional review board approval for the study from UMass Chan Medical School (H00011407), we sent each clinician a unique link to the baseline Web-based survey by e-mail, along with a fact sheet explaining its purpose. If clinicians did not respond after up to five attempts to reach them by e-mail, they were sent a paper copy of the survey, fact sheet, and a return envelope by mail. The survey took no longer than 15 minutes to complete, and most responses were returned between November 2016 and March 2017.
Measures
The baseline survey was administered before any implementation or training related to the Zero Suicide model across the health care system. Participants were asked to report their sex (male or female), age, ethnicity, race, clinical role, setting or settings (emergency medicine, inpatient services, or outpatient clinic) they worked in, whether they had an administrative role (yes or no), and how many years they had been working in health care (excluding time in training).
Suicide-related attitudes and practices
Participants were asked to rate, on a 5-point Likert scale (1, strongly disagree; 2, disagree; 3, uncertain; 4, agree; and 5, strongly agree), their confidence in their ability to screen patients for suicidality, assess suicide risk severity, provide brief counseling to suicidal patients, help patients at risk create a personalized safety plan, and find referral resources. Using a separate scale (1, none; 2, a few; 3, some; 4, most; and 5, all), participants also reported how many suicides they considered preventable; the proportion of all patients they personally screen for suicidal ideation; and the proportion of suicidal patients they personally assess for risk severity, create a written personalized safety plan for, briefly counsel, and provide referrals to outpatient or community resources. (We considered all patients eligible for suicide screening and suicidal patients eligible for assessment, safety planning, counseling, and referrals.) The items were summed to create a confidence scale and practice scale, respectively. These confidence and practice instruments have been used in previous research (15) and had high internal consistency in our sample (Cronbach’s alphas of 0.89 and 0.86, respectively). Participants were also asked to rate (from 1, strongly disagree, to 5, strongly agree) their views on barriers to universal screening for suicide and the degree to which they felt departmental leadership supported suicide prevention. These items did not display high internal consistency (Cronbach’s α=0.65) and are therefore presented separately in our analyses.
Attitudes towards safety planning intervention (SPI)
Participants were asked eight questions about the SPI on a 5-point Likert scale, ranging from 1, strongly disagree, to 5, strongly agree, assessing attitudes toward SPI itself (i.e., SPI results in patients at risk for suicide engaging in recommended treatment, and SPI is an important part of suicide prevention), barriers (i.e., SPI will take too much time to complete, SPI will result in a costly delay, and deciding which role will deliver SPI will be difficult [all reverse- coded]), facilitators (clinician will have the time necessary to implement it, and the current electronic health record [EHR] system has the capabilities to support it), and clinicians’ confidence in their own ability to administer it. The eight SPI items had adequate internal consistency (Cronbach’s α=0.74), and there were no items whose removal would have improved the reliability of the scale. Therefore, we summed the items to obtain an SPI attitudes score (range 8–40, mean±SD=25.8±4.1). About 20% of respondents skipped more than two items in the SPI section, possibly because they were not yet familiar with the intervention.
Attitude and Alignment on CQI
Participants were asked 13 questions to rate alignment on CQI within the whole organization (i.e., not specific to suicide-related improvements) on a 5-point Likert scale, ranging from 1, strongly disagree, to 5, strongly agree. Questions focused on the clinician’s own alignment with the organization, perceived commitment of departmental leadership to quality improvement, infrastructure and facilitators of quality improvement, and the organization’s competing priorities. The 13 CQI items had high internal consistency (Cronbach’s α=0.91), and there were no items whose removal would have improved the reliability of the scale. Therefore, we summed these items to generate a CQI alignment sum score (range 13–65, mean=45.5±8.2). Participants were also asked whether they had received formal performance improvement or CQI training (yes or no) and whether they were currently using CQI methods in their work (yes or no).
Analyses
We categorized clinical role into the following options: nurse, medical provider (e.g., attending physicians, resident physicians, and advanced practice providers), behavioral health clinician (e.g., psychiatric counselors, psychologists, or social workers, excluding psychiatrists), and patient care assistants (e.g., medical office assistants and observers). We also categorized age in years into approximate quartiles: <35 years, 35–44, 45–54, and ≥55 years.
For participants who were missing two or fewer items used in the scales, we imputed the missing items with the respondent’s mean response on that scale. The confidence and practice scores were not normally distributed, so we used Kruskal-Wallis H tests to examine differences in scores by role and setting. The scores on the SPI and CQI scales were normally distributed, so we conducted independent Student’s t tests and analyses of variance to examine differences in scores by participant characteristics. We used Bonferroni post hoc tests for multiple comparisons. All tests were two-sided, with a statistical significance level of α=0.05.
Results
Sample description
Of the 5,559 clinicians we invited, 1,224 completed the survey (22% response rate). Of the final sample, 668 (55%) were nurses, 333 (27%) medical providers, 90 (7%) behavioral health providers, 116 (10%) patient care assistants, and 17 (1%) other provider or missing provider information. In total, 256 (21%) worked in the ED, 795 (65%) in inpatient services, and 404 (33%) in an outpatient clinic (231 [19%] reported working in more than one setting). A total of 233 (19%) reported that they had an administrative role in their department. Overall, 986 (81%) participants were female, 209 (17%) were male, and 29 (2%) did not report their sex. Fifty participants (4%) were Hispanic, 1,053 (86%) White, 50 (4%) Black or African American, 36 (3%) Asian, and 6 (1%) American Indian or Alaska Native; 29 (2%) participants did not report their race. We obtained administrative data on the demographic composition of the 5,559 clinicians in the health care system to ascertain the representativeness of our sample and found our sample to be similar to the larger population (data not shown).
Suicide prevention confidence, practice, and attitudes
Most of the surveyed clinicians (N=799, 65%) reported that they were confident that they had the skills needed to screen patients for suicidality, but fewer reported confidence in performing other suicide-related practices, namely further assessment (N=612, 50%), brief counseling (N=463, 38%), safety planning (N=361, 30%), and providing referrals or resources (N=527, 43%). Few clinicians reported personally delivering suicide-related practices to all or most of their eligible patients: 406 (33%) for screening, 445 (36%) for further assessment, 287 (23%) for brief counseling, 128 (11%) for safety planning, and 285 (23%) for providing referrals or resources. (We considered all patients eligible for suicide screening and suicidal patients eligible for assessment, safety planning, counseling, and referrals.) Confidence and self-reported practice differed substantially by role (Table 1) and by setting (Table 2), with behavioral health providers and ED clinicians tending to have higher confidence in conducting and a higher level of suicide-related practice. In total, 675 (55%) participants endorsed the belief that most or all suicides were preventable, 234 (19%) believed that universal screening for suicide would result in increased psychiatric evaluations, and 597 (49%) agreed that universal screening for suicide would slow clinical care. Regarding leadership, 606 (50%) agreed that departmental leadership supported improvement in interventions for suicidal patients, and 546 (45%) felt that treatment of suicidal patients was a top priority in their department. Again, behavioral health providers and those working in the ED tended to have the most favorable attitudes toward suicide prevention and views on leadership support (Table 1 and Table 2).
TABLE 1.
Scores on confidence, self-reported practice, and attitudes toward suicide prevention, by clinical role
| Item | Nurse (N=664)a | Medical provider (N=333) | Behavioral health provider (N=90) | Patient care assistant (N=115)a | Kruskal-Wallis H statisticb | p | ||||
|---|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
|
|||||||
| Median | IQR | Median | IQR | Median | IQR | Median | IQR | |||
|
| ||||||||||
| Confidence in performing suicide carec | ||||||||||
| Screening | 4.0 | 3.0–4.0 | 4.0 | 4.0–4.0 | 4.0 | 4.0–5.0 | 4.0 | 3.0–4.0 | ||
| Assessment | 3.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | 4.0 | 4.0–5.0 | 3.0 | 3.0–4.0 | ||
| Brief counseling | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 4.0 | 4.0–5.0 | 3.0 | 2.0–4.0 | ||
| Safety planning | 3.0 | 2.0–3.0 | 3.0 | 2.0–4.0 | 4.0 | 4.0–5.0 | 3.0 | 2.0–4.0 | ||
| Making a referral | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 4.0 | 4.0–5.0 | 3.0 | 2.0–4.0 | ||
| Overall confidence score | 3.1 | 2.6–3.6 | 3.4 | 2.8–3.4 | 4.0 | 3.8–4.8 | 3.0 | 2.4–4.0 | 101.1 | <.001d |
| Suicide prevention practicee | ||||||||||
| Screening | 3.0 | 1.0–5.0 | 3.0 | 2.0–4.0 | 3.0 | 2.0–5.0 | 1.0 | 1.0–1.0 | ||
| Assessment | 2.0 | 1.0–5.0 | 3.0 | 1.0–5.0 | 4.0 | 1.0–5.0 | 1.0 | 1.0–1.0 | ||
| Brief counseling | 1.0 | 1.0–2.0 | 2.0 | 1.0–5.0 | 4.0 | 2.0–5.0 | 1.0 | 1.0–1.0 | ||
| Safety planning | 1.0 | 1.0–1.0 | 1.0 | 1.0–2.0 | 2.0 | 1.0–4.0 | 1.0 | 1.0–1.0 | ||
| Making a referral | 1.0 | 1.0–2.0 | 3.0 | 1.0–5.0 | 4.0 | 1.0–5.0 | 1.0 | 1.0–1.0 | ||
| Overall practice score | 1.8 | 1.0–2.8 | 2.6 | 1.8–3.6 | 3.5 | 1.6–4.4 | 1.0 | 1.0–1.4 | 154.4 | <.001f |
| Attitudes toward universal screening and departmental leadershipg | ||||||||||
| Universal screening for suicide results in increased psychiatric evaluations. | 3.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 7.4 | .06 |
| Universal screening for suicide slows down clinical care. | 2.0 | 2.0–3.0 | 3.0 | 2.0–4.0 | 2.0 | 2.0–3.0 | 3.0 | 2.0–4.0 | 25.1 | <.001h |
| Our departmental leadership supports improvement in interventions for suicidal patients. | 3.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | 33.4 | <.001i |
| In providing clinical care, treatment of suicidal patients is a top priority in our department. | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 4.0 | 3.0–5.0 | 3.0 | 3.0–4.0 | 36.2 | <.001j |
| Our departmental leadership has made sure we receive the tools and training to manage suicidal patients. | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 4.0 | 3.0–4.0 | 3.0 | 2.0–4.0 | 56.9 | <.001k |
A small number of respondents in this group had skipped the questions about practice.
df=3.
Scores range from 1 to 5, with higher scores indicating greater confidence in ability.
Behavioral health provider (BH) > medical provider (MP) (p<0.001), nurse (p<0.001), patient care assistant (PCA) (p<0.001); MP > nurse (p<0.001). Post hoc results from H test of the overall confidence score.
Clinicians rated the proportion of patients for whom they personally performed the type of practice (1, none; 2, a few; 3, some; 4, most; 5, all).
BH > nurse (p<0.001), PCA (p<0.001); MP > nurse (p<0.001), PCA (p<0.001); nurse > PCA (p<0.001). Post hoc results from H test of the overall practice score.
Scores range from 1, strongly agree, to 5, strongly disagree.
MP > BH (p=0.04), nurse (p<0.001); post hoc results from H test.
Nurse < MP (p<0.001), BH (p<0.001); post hoc results from H test.
BH > MP (p<0.001), nurse (p<0.001), PCA (p=0.02); post hoc results from H test.
BH > MP (p<0.001), nurse (p<0.001), PCA (p=0.02); MP > nurse (p=0.001); PCA > nurse (0.004). Post hoc results from H test.
TABLE 2.
Scores on confidence, self-reported practice, and attitudes toward suicide prevention, by practice setting
| Item | Emergency department only (N=134)a | Inpatient setting only (N = 577)a | Outpatient setting only (N=216)a | More than one setting (N=231) | Kruskal-Wallis H statisticb | p | ||||
|---|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
|
|||||||
| Median | IQR | Median | IQR | Median | IQR | Median | IQR | |||
|
| ||||||||||
| Confidence in suicide carec | ||||||||||
| Screening | 4.0 | 4.0–5.0 | 4.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | 4.0 | 3.0–5.0 | ||
| Assessment | 4.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | ||
| Brief counseling | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 3.0 | 3.0–4.0 | 3.0 | 2.0–4.0 | ||
| Safety planning | 3.0 | 2.0–4.0 | 3.0 | 2.0–3.0 | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | ||
| Making a referral | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 4.0 | 3.0–4.0 | 3.0 | 2.0–4.0 | ||
| Overall confidence score | 3.4 | 2.8–4.0 | 3.2 | 2.4–3.6 | 3.4 | 2.8–4.0 | 3.4 | 2.8–4.0 | 35.7 | <.001d |
| Suicide prevention practicee | ||||||||||
| Screening | 4.0 | 3.0–5.0 | 2.0 | 1.0–4.0 | 2.0 | 1.0–4.0 | 3.0 | 1.0–4.0 | ||
| Assessment | 4.0 | 2.0–5.0 | 2.0 | 1.0–4.0 | 2.0 | 1.0–4.0 | 3.0 | 1.0–5.0 | ||
| Brief counseling | 2.0 | 1.0–4.0 | 1.0 | 1.0–3.0 | 1.0 | 1.0–4.0 | 2.0 | 1.0–4.5 | ||
| Safety planning | 1.0 | 1.0–2.0 | 1.0 | 1.0–1.0 | 1.0 | 1.0–1.8 | 1.0 | 1.0–2.0 | ||
| Making a referral | 2.0 | 1.0–4.0 | 1.0 | 1.0–2.0 | 1.0 | 1.0–4.0 | 2.0 | 1.0–5.0 | ||
| Overall practice score | 2.8 | 1.8–3.5 | 1.8 | 1.0–2.8 | 1.8 | 1.0–3.4 | 2.2 | 1.2–3.5 | 42.2 | <.001f |
| Attitudes toward universal screening and departmental leadershipg | ||||||||||
| Universal screening for suicide results in increased psychiatric evaluations. | 4.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 12.3 | .006h |
| Universal screening for suicide slows down clinical care. | 2.0 | 2.0–3.0 | 2.0 | 2.0–3.0 | 3.0 | 2.0–3.0 | 3.0 | 2.0–3.0 | 12.6 | .006i |
| Our departmental leadership supports improvement in interventions for suicidal patients. | 4.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 4.0 | 3.0–4.0 | 3.0 | 3.0–4.0 | 40.5 | <.001j |
| In providing clinical care, treatment of suicidal patients is a top priority in our department. | 4.0 | 3.0–4.0 | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 35.2 | <.001k |
| Our departmental leadership has made sure we receive the tools and training to manage suicidal patients. | 4.0 | 3.0–4.0 | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 3.0 | 2.0–4.0 | 49.7 | <.001l |
Some respondents in this group had skipped the questions about confidence, self-reported practice, and attitudes toward suicide prevention.
df=3.
Scores range from 1 to 5, with higher scores indicating greater confidence in ability.
Inpatient setting (IP) < emergency department (ED) (p= 0.002), outpatient setting (OP) (p<0.001), more than one setting (MO) (p<0.001); post hoc results from H test of the overall confidence score.
Clinicians rated the proportion of patients for whom they personally performed the type of practice (1, none; 2, a few; 3, some; 4, most; 5, all).
ED > IP (p<0.001), OP (p<0.001); MO > IP (p<0.001). Post hoc results from H test of the overall practice score.
Scores range from 1, strongly agree, to 5, strongly disagree.
ED > IP (p=0.006), OP (p=0.01); post hoc results from H test.
IP < MO (p=0.005); post hoc results from H test.
ED > IP (p<0.001), OP (p=0.02), MO (p<0.001); OP > IP (p= 0.007). Post hoc results from H test.
ED > IP (p<0.001), OP (p=0.004), MO (p=0.005); post hoc results from H test.
ED > IP (p<0.001), OP (p<0.001), MO (p=0.02); MO > IP (pcO.001). Post hoc results from H test.
Attitudes towards the Safety Planning Intervention (SPI)
Respondents had generally favorable attitudes toward the SPI, with most participants agreeing that it would result in patients at risk for suicide engaging in recommended treatment (N=670 of 1,112, 60%) and is an important part of suicide prevention (N=873 of 1,140, 77%). Fewer participants believed that the SPI would take too much time (N=219 of 1,106, 20%), would cause costly delays (N=149 of 1,099, 14%), or would be difficult to assign to a specific role (N=281 of 1,105, 25%). Only 23% (N=258 of 1,113) of respondents reported that they were confident they would have the time necessary to implement the SPI, 22% (N=235 of 1,064) agreed that the current EHR system had the capabilities to support it, and 32% (N=353 of 1,116) reported confidence in their own ability to administer the SPI. On the summed SPI attitude score, both sex and clinical role were associated with attitudinal differences, with women and behavioral health providers reporting significantly more favorable attitudes, compared with men and other types of clinician, respectively (Table 3). Post hoc tests indicated that behavioral health providers scored significantly higher on SPI attitudes than all other groups, and nurses and patient care assistants both scored significantly higher on SPI attitudes compared with medical providers. Of all clinician groups, male physicians had the lowest SPI attitude scores (23.9±3.8), and female behavioral health providers had the highest (29.2±4.3). We found no association between SPI attitude score and other demographic characteristics, clinical setting, having an administrative role, or years working in health care.
TABLE 3.
Association between participant characteristics (N=979) and SPI attitude scoresa
| Characteristic | Attitude score (M±SD) | Test statistic | df | p |
|---|---|---|---|---|
| Sex | t=3.5 | 957 | <.001 | |
| Male | 24.8±4.5 | |||
| Female | 26.1±4.0 | |||
| Ethnicity | t=1.3 | 938 | .19 | |
| Non-Hispanic | 25.9±4.1 | |||
| Hispanic | 26.7±3.1 | |||
| Race | t=.6 | 977 | .55 | |
| Non-White | 26.0±4.5 | |||
| White | 25.8±4.5 | |||
| Age group (years) | F=.5 | 3, 975 | .71 | |
| 18–34 | 26.0±4.5 | |||
| 35–44 | 25.6±3.9 | |||
| 45–54 | 25.9±3.9 | |||
| ≥55 | 25.8±4.1 | |||
| Administrative role | t=.6 | 962 | .57 | |
| No | 25.8±4.1 | |||
| Yes | 25.6±4.4 | |||
| Years in health care | F=.7 | 2, 942 | .48 | |
| <10 | 26.0±4.4 | |||
| 10–24 | 25.9±3.9 | |||
| ≥25 | 25.7±4.0 | |||
| Works in emergency department | t=.4 | 977 | .69 | |
| No | 25.9±3.9 | |||
| Yes | 25.7±4.9 | |||
| Works in inpatient setting | t=1.0 | 977 | .32 | |
| No | 26.0±4.5 | |||
| Yes | 25.7±3.9 | |||
| Works in outpatient setting | t=.2 | 977 | .98 | |
| No | 25.8±4.1 | |||
| Yes | 25.9±4.2 | |||
| Role | F=20.6 | 3, 975 | <.001b | |
| Nurse | 25.7±3.8 | |||
| Medical provider | 24.9±4.2 | |||
| Behavioral health provider or social work | 28.9±4.4 | |||
| Patient care assistant or technician | 26.9±4.2 |
Total safety planning intervention (SPI) attitude scores range from 8 to 40, with higher scores indicating more favorable attitudes.
Behavioral health provider > medical provider (MP) (p<0.0001), nurse (p<0.001), patient care assistant (PCA) (p=0.01); MP < nurse (p=0.03), PCA (p<0.001). Results of post hoc tests.
Attitudes towards Continuous Quality Improvement (CQI)
More than half of the respondents (N=647, 53%) reported that they had received formal performance improvement or CQI training, and 65% (N=800) reported that they were currently using CQI methods in their work. Respondents held favorable attitudes toward CQI, especially related to leadership engagement. On the summed CQI score, several participant characteristics–namely, non-White race, male sex, younger age, having an administrative role, fewer years working in health care, working in an outpatient care setting, and medical provider role–were associated with more favorable attitudes (Table 4). Post hoc tests revealed that respondents ages 18–34 years scored significantly higher on CQI than those ages ≥55, those working in health care for <10 years scored significantly higher than those working ≥10 years, and medical providers scored significantly higher than nurses. We detected no significant association between the CQI score and ethnicity.
TABLE 4.
Association between participant characteristics and CQI attitude scoresa
| Characteristic | Attitude score (M±SD) | Test statistic | df | p |
|---|---|---|---|---|
| Sex | t=2.7 | 1,174 | .01 | |
| Male | 46.9±8.0 | |||
| Female | 45.3±8.1 | |||
| Ethnicity | t=1.7 | 1,152 | .10 | |
| Non-Hispanic | 45.5±8.1 | |||
| Hispanic | 47.5±8.0 | |||
| Race | t=2.2 | 1,198 | .03 | |
| Non-White | 46.8±8.7 | |||
| White | 45.3±8.1 | |||
| Age group (years) | F=.5 | 3, 1,196 | .02b | |
| 18–34 | 46.9±8.1 | |||
| 35–44 | 45.5±7.8 | |||
| 45–54 | 45.4±7.5 | |||
| ≥55 | 44.5±8.8 | |||
| Administrative role | t=5.0 | 1,182 | <.001 | |
| No | 44.8±8.4 | |||
| Yes | 47.8±6.8 | |||
| Years in health care | F=8.9 | 2, 1,159 | <.001c | |
| <10 | 46.8±7.9 | |||
| 10–24 | 44.5±8.6 | |||
| ≥25 | 45.0±7.9 | |||
| Works in emergency department | t=3.3 | 1,198 | <.001 | |
| No | 45.1±8.3 | |||
| Yes | 47.0±7.4 | |||
| Works in inpatient setting | t=4.5 | 1,198 | <.001 | |
| No | 46.9±7.7 | |||
| Yes | 44.7±8.3 | |||
| Works in outpatient setting | t=2.4 | 1,198 | .02 | |
| No | 45.1±8.3 | |||
| Yes | 46.3±7.8 | |||
| Role | F=10.0 | 3, 1,196 | <.001d | |
| Nurse | 44.4±8.2 | |||
| Medical provider | 47.3±7.3 | |||
| Behavioral health provider or social work | 45.7±7.6 | |||
| Patient care assistant or technician | 46.4±10.0 |
Total continuous quality improvement (CQI) attitude scores range from 13 to 65, with higher scores indicating greater organizational alignment with CQI.
18–34 years > 55 years and older (p<0.001); results of post hoc tests.
Less than 10 years > 10–24 years (p=0.005), 25 years or more (p<0.001); results of post hoc tests.
Medical provider > nurse (p<0.001); results of post hoc tests.
Discussion
We believe this is the first survey to examine clinicians’ attitudes towards suicide prevention across an entire health system, including a wide range of clinical roles and settings. We found that most clinicians felt confident about screening for suicide but less confident about performing more complex practices such as risk assessment and safety planning. They also reported delivering these practices to a minority of eligible patients. Previous research has also shown higher clinician confidence in screening compared with other suicide- related care processes (10, 15–17), especially safety planning (18). Among the clinical roles examined, behavioral health providers had the highest confidence and performance across types of practice and were the most likely to believe that suicide was preventable and that suicide prevention was a key priority of leadership. This finding echoes those of previous research suggesting that individuals working in behavioral health roles have more positive attitudes toward suicide prevention and suicidal patients (11–13).
In our study, participants who worked in the ED tended to have the highest confidence in, self-reported delivery frequency of, and positive attitudes toward suicide prevention, followed by those who worked in outpatient settings. Participants working in the ED had concerns about increased psychiatric evaluations related to universal suicide screening, and medical providers were concerned that universal screening would slow clinical care, two common attitudinal barriers to implementation of universal suicide screening (19). Few studies have explored the influence of care setting on attitudes toward universal screening, but Kishi et al. (12) found worse attitudes among clinicians in the ED than in inpatient psychiatric units.
Most respondents had a positive attitude toward the SPI but expressed concerns about having the time and capability to deliver it, echoing concerns about gaps in safety planning training identified in previous research (18). Furthermore, many clinicians lacked the confidence to provide brief counseling to suicidal patients. This finding reflects a broader pattern in the literature, where patients at risk for suicide do not receive effective interventions in acute care services (20). Those who die by suicide often visit acute care before their death (1–3), creating increasing pressure for health care systems to go beyond assessment and ensure that suicidal patients receive effective interventions in a timely manner.
Our survey revealed a strong culture of CQI, especially related to leadership buy-in. Scores assessing CQI alignment were higher among younger and more recently trained clinicians, medical providers, those in administrative roles, and those working in the ED than among their peers. CQI is an important implementation strategy for the Zero Suicide model because it provides a framework to support, monitor, and guide improvement of the intervention (4, 21). A lack of CQI infrastructure can impede the ability to effectively implement and evaluate quality improvement in suicide prevention (20). Frontline engagement, leadership support, and a supportive culture help to ensure that CQI leads to long- term improvements in patient care and outcomes (22). We note that in our sample, attitudes about CQI were more favorable among men and medical providers, the same respondents who tended to have less favorable attitudes toward the SPI. It is possible that leveraging CQI strategies to implement suicide-related interventions might help to engage subgroups of clinicians who might otherwise resist such efforts.
This study had several limitations. Although every clinician in the health system was invited to participate in the survey, less than one-quarter responded, leading to response bias, such that the observed attitudes, self-efficacy, and practices may have been positively inflated or overestimated. It is possible that this minority of respondents possessed more positive attitudes, higher self-efficacy, or higher levels of suicide-related practice or prioritized suicide prevention more than did clinicians who did not respond. Still, the observed response rate is typical of clinician surveys and was likely optimized by our multimodality approach and multiple reminders (23). Furthermore, the respondents’ demographic profile was similar to that of the clinicians within the health care system as a whole, supporting the representativeness of the sample. We found that patient care assistants and those not in administrative roles were more likely to start but not complete the survey, and about one-fifth of respondents skipped the SPI questions, perhaps because they were not yet familiar with the intervention. We could not differentiate between respondents working in psychiatric inpatient units versus medical inpatient units, and we did not collect data on which facility a clinician worked at most often, which would have allowed us to explore facility-related differences. Finally, the measures used have yet to be psychometrically validated, although our study data indicated acceptable internal consistency of both the safety attitudes instrument and the CQI attitudes instrument. Additionally, the confidence instrument and attitudinal instrument have been used in previous studies assessing clinician attitudes (15). The strengths of the study include the relatively large sample, compared with previous studies, the range of items assessed, and the representation of multiple clinical roles and settings across the same health care system.
Overall, the findings of our study suggest that those seeking to implement a systemwide suicide prevention program should consider a range of potential influences while designing the program. Our findings not only support the need for further training focused on attitudes and skills (ideally including live training and hands-on demonstration sessions) but also indicate the strong potential for CQI-type implementation strategies, such as small cyclical tests of change and audit-and-feedback strategies. Implementers should consider the complexity of and familiarity with the care practices proposed as well as contextual facilitators (such as existing CQI infrastructure and perceived leadership support) and potential barriers (time and cost), which may vary significantly across different clinical roles and settings.
Conclusions
We observed variability in clinicians’ self-reported confidence in suicide-related care, with most clinicians reporting higher confidence in conducting screening for suicide risk and lower confidence in performing more complex assessment and intervention practices. Behavioral health clinicians and those working in the ED reported the highest confidence and highest level of suicide-related practice. Attitudes toward safety planning were more favorable among women and behavioral health providers, whereas attitudes about CQI were more positive among men and physicians, indicating that CQI could be a helpful means of engaging a broader range of clinicians in suicide prevention efforts.
Highlights.
Although clinicians’ self-reported confidence in suicide-related care varied, most reported higher confidence in conducting screening for suicide risk and lower confidence in more complex assessment and intervention practices.
Behavioral health clinicians and those working in the emergency department reported the highest confidence in and the highest level of suicide-related practice.
Scores assessing attitudes toward safety planning were high among women and behavioral health providers.
Scores assessing attitudes toward continuous quality improvement were high among men and medical providers (such as physicians), indicating that the approach could be helpful for engaging a broader range of clinicians in suicide prevention efforts.
AUTHOR AND ARTICLE INFORMATION
This research was supported by NIMH (award R01 MH-112138).
The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Footnotes
The authors report no financial relationships with commercial interests.
Contributor Information
Celine Larkin, Department of Emergency Medicine, University of Massachusetts Chan Medical School, Worcester; Department of Psychiatry, University of Massachusetts Chan Medical School, Worcester.
Catarina I. Kiefe, Department of Population and Quantitative Health Sciences, University of Massachusetts Chan Medical School, Worcester.
Alexandra L. Morena, Department of Psychology, University of Massachusetts Lowell, Lowell.
Mhd B. Rahmoun, Department of Emergency Medicine, University of Massachusetts Chan Medical School, Worcester.
Peter Lazar, Department of Population and Quantitative Health Sciences, University of Massachusetts Chan Medical School, Worcester.
Ana Vallejo Sefair, Department of Emergency Medicine, University of Massachusetts Chan Medical School, Worcester.
Edwin D. Boudreaux, Department of Emergency Medicine, University of Massachusetts Chan Medical School, Worcester; Department of Psychiatry, University of Massachusetts Chan Medical School, Worcester.
References
- 1.Stene-Larsen K, Reneflot A. Contact with primary and mental health care prior to suicide: a systematic review of the literature from 2000 to 2017. Scand J Public Health. 2019. Feb;47(1):9–17. [DOI] [PubMed] [Google Scholar]
- 2.Ahmedani BK, Simon GE, Stewart C, et al. : Health care contacts in the year before suicide death. J Gen Intern Med 2014; 29: 870–877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Roaten K, Horowitz LM, Bridge JA, et al. : Universal pediatric suicide risk screening in a health care system: 90,000 patient encounters. J Acad Consult Liaison Psychiatry 2021; 62:421–429. [DOI] [PubMed] [Google Scholar]
- 4.Labouliere CD, Vasan P, Kramer A, et al. : “Zero Suicide”–a model for reducing suicide in United States behavioral healthcare. Suicidologi 2018; 23:22–30. [PMC free article] [PubMed] [Google Scholar]
- 5.Turner K, Sveticic J, Almeida-Crasto A, et al. Implementing a systems approach to suicide prevention in a mental health service using the Zero Suicide Framework. Aust N Z J Psychiatry. 2021. Mar;55(3):241–53. [DOI] [PubMed] [Google Scholar]
- 6.Damschroder LJ, Aron DC, Keith RE, et al. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implement Sci. 2009. Dec;4(1):1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Grimholt TK, Haavet OR, Jacobsen D, et al. Perceived competence and attitudes towards patients with suicidal behaviour: a survey of general practitioners, psychiatrists and internists. BMC Health Serv Res. 2014. Dec;14(1):1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Suominen K, Suokas J, Lönnqvist J. Attitudes of general hospital emergency room personnel towards attempted suicide patients. Nord J Psychiatry. 2007. Jan 1;61(5):387–92. [DOI] [PubMed] [Google Scholar]
- 9.Neville K, Roan NM. Suicide in hospitalized medical-surgical patients: Exploring nurses’ attitudes. J Psychosoc Nurs Ment Health Serv. 2013. Jan 1;51(1):35–43. [DOI] [PubMed] [Google Scholar]
- 10.Wakai S, Schilling EA, Aseltine RH Jr, et al. Suicide prevention skills, confidence and training: Results from the Zero Suicide Workforce Survey of behavioral health care professionals. SAGE Open Med. 2020. Jun;8:2050312120933152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Giacchero Vedana KG, Magrini DF, Zanetti AC, et al. Attitudes towards suicidal behaviour and associated factors among nursing professionals: A quantitative study. J Psychiatr Ment Health Nurs. 2017. Nov;24(9–10):651–9. [DOI] [PubMed] [Google Scholar]
- 12.Kishi Y, Kurosawa H, Morimura H, et al. Attitudes of Japanese nursing personnel toward patients who have attempted suicide. Gen Hosp Psychiatry. 2011. Jul 1;33(4):393–7. [DOI] [PubMed] [Google Scholar]
- 13.Draper B, Krysinska K, De Leo D, et al. Health care professional attitudes toward interventions to prevent suicide. Suicidol Online. 2014. Jan 1:24. [Google Scholar]
- 14.Boudreaux ED, Larkin C, Sefair AV, et al. Studying the implementation of Zero Suicide in a large health system: Challenges, adaptations, and lessons learned. Contemp Clin Trials Commun. 2022. Dec 1;30:100999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Betz ME, Sullivan AF, Manton AP, et al. : Knowledge, attitudes, and practices of emergency department providers in the care of suicidal patients. Depress Anxiety 2013; 30:1005–1012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Maina R, Bukusi D, Kumar M. Suicide prevention by emergency nurses: perceived self-efficacy in assessment, management and referral at Kenyatta National Hospital in Kenya. Ann Gen Psychiatry 2019; 18:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Fernandes J, Scheuermeyer FX, Chakraborty AT, et al. : What are Canadian emergency physicians’ attitudes and self-perceived competence toward patients who present with suicidal ideation?. CJEM 2021; 23:668–672 [DOI] [PubMed] [Google Scholar]
- 18.Reyes-Portillo JA, McGlinchey EL, Toso-Salman J, et al. Clinician experience and attitudes toward safety planning with adolescents at risk for suicide. Arch Suicide Res 2019; 23:222–233 [DOI] [PubMed] [Google Scholar]
- 19.Goldstein Grumet J, Boudreaux ED. Universal suicide screening is feasible and necessary to reduce suicide. Psychiatr Serv 2023; 74:81–83 [DOI] [PubMed] [Google Scholar]
- 20.Bridge JA, Olfson M, Caterino JM, et al. : Emergency department management of deliberate self-harm: a national survey. JAMA Psychiatry 2019; 76:652–654 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bowden CF, Worsley D, Esposito JM, et al. : Pediatric emergency departments’ readiness for change toward improving suicide prevention: a mixed-methods study with US leaders. J Am Coll Emerg Physicians Open 2022; 3:e12839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Coles E, Anderson J, Maxwell M, et al. : The influence of contextual factors on healthcare quality improvement initiatives: a realist review. Syst Rev 2020; 9:94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cho YI, Johnson TP, VanGeest JB. Enhancing surveys of health care professionals: a meta-analysis of techniques to improve response. Eval Health Prof 2013; 36:382–407 [DOI] [PubMed] [Google Scholar]