|Year : 2020 | Volume
| Issue : 1 | Page : 31-42
Systematic Suicide Screening in a General Hospital Setting: Process and Initial Results
Andres J Pumariega1, Kolin Good2, Kelly Posner3, Udema Millsaps2, Barbara Romig2, Debra Stavarski2, Robert Rice2, Mary Jo Gehret2, Kevin Riley2, Thomas E Wasser4, Gayle Walsh2, Heather Yarger5
1 Department of Psychiatry, Division of Child and Adolescent Psychiatry, University of Florida, Gainesville, FL, USA
2 Department of Nursing, Psychiatry, and Information Management Systems, The Reading Hospital and Medical Center, Reading, PA, USA
3 Department of Psychiatry, Suicide Prevention Center, School of Medicine, Columbia University, New York, NY, USA
4 Complete Statistical Services, Macungie, PA, USA
5 Department of Psychology, University of Maryland, College Park MD, USA
|Date of Submission||18-Dec-2019|
|Date of Decision||31-Dec-2019|
|Date of Acceptance||16-Jan-2020|
|Date of Web Publication||21-Mar-2020|
Prof. Andres J Pumariega
Department of Psychiatry, Division of Child and Adolescent Psychiatry, University of Florida, Gainesville, FL
Source of Support: None, Conflict of Interest: None
Background: Suicide is one of the leading causes of death across all age groups globally and poses a significant public health burden. In response to the United States Joint Commission National Patient Safety Goals, a tertiary hospital in the Northeast U.S. developed a suicide risk assessment and response protocol, consisting of systematic screening of patients for suicidal ideation/behavior with a screening version of the Columbia Suicide Severity Rating Scale (C-SSRS) and a response algorithm based on risk levels derived from the screen. Methods: A total of 837 nurses were trained and 24,168 patients ages 12 and above were screened with the C-SSRS Screener. Results: Posttraining interrater reliability on the C-SSRS Screener definitions of ideation and behavior was high and independent of level of education or mental health experience. Of the patients screened, only 144 patients (0.93%) were in the highest risk category, and they were assigned patient safety monitors until a follow-up consultation. The highest risk levels from the C-SSRS Screener reasonably identified subsequent attempts at self-injurious behavior during hospitalization. Screening resulted in lower burden due to reduction in the rate of psychiatric consultations and one-to-one observation shifts. Conclusions: These findings suggest that a systematic screening and clinical response protocol using the C-SSRS Screener can potentially enhance the ability to identify suicide risk in the general hospital population and focus services on patients with the most need.
Keywords: General hospital, nursing practice, suicide prevention, suicide screening
|How to cite this article:|
Pumariega AJ, Good K, Posner K, Millsaps U, Romig B, Stavarski D, Rice R, Gehret MJ, Riley K, Wasser TE, Walsh G, Yarger H. Systematic Suicide Screening in a General Hospital Setting: Process and Initial Results. World Soc Psychiatry 2020;2:31-42
|How to cite this URL:|
Pumariega AJ, Good K, Posner K, Millsaps U, Romig B, Stavarski D, Rice R, Gehret MJ, Riley K, Wasser TE, Walsh G, Yarger H. Systematic Suicide Screening in a General Hospital Setting: Process and Initial Results. World Soc Psychiatry [serial online] 2020 [cited 2021 Apr 22];2:31-42. Available from: https://www.worldsocpsychiatry.org/text.asp?2020/2/1/31/281135
| Introduction|| |
The global suicide epidemic
Suicide is a serious global public health problem. Close to 800,000 people die due to suicide every year, which is one person every 40 seconds. Suicide is a global phenomenon and occurs throughout the lifespan.
For each adult who died by suicide, there may be more than twenty others attempting suicide. Suicide occurs throughout the lifespan and is the second leading cause of death among 15–29 year olds globally. The epidemic of suicide is a global phenomenon; in fact, 79% of suicides occurred in low- and middle-income countries in 2016. Suicide accounted for 1.4% of all deaths worldwide, making it the 18th leading cause of death in 2016.
Stigma, particularly surrounding mental disorders and suicide, means many people thinking of taking their own life or who have attempted suicide are not seeking help and are therefore not getting the help they need. The prevention of suicide has not been adequately addressed due to a lack of awareness of suicide as a major public health problem and the taboo in many societies to openly discuss it. To date, only a few countries have included suicide prevention among their health priorities and only 38 countries report having a national suicide prevention strategy.
Strikingly, 45% of individuals (across age groups) who died by suicide had contact with a primary care provider within the 30 days prior to their death and 77% had contact in the previous year (90% of adolescents). In addition, a large number of suicide attempters present to emergency departments for nonpsychiatric reasons and thus will be missed unless screening takes place. The problem in medical populations is serious in scope and severity, and medical professionals are in a key position to screen for early signs of suicidal ideation and behavior in order to enhance prevention of this serious adverse health outcome. Screening for suicide in health-care organizations was first designated by the United States Joint Commission for Healthcare Organizations as a National Patient Safety Goal, with the initial goal of screening all behavioral health patients and the eventual goal of screening all medical patients at all levels of care along with the development of assessment and safety procedures for identified patients. As of 2019, suicide in hospitals was the number six sentinel event. Most recently, the Joint Commission (TJC) updated their sentinel event alert to include all health-care settings. In addition, the WHO has recommended screening by nonspecialty health professionals as an important suicide prevention strategy.
Suicide screening in health-care settings
There are many challenges in implementing suicide screening. These have included the lack of low burden, valid, reliable, and universally accepted tools addressing suicidal ideation and behavior based on clear operational definitions; the need for training on suicide risk screening; and time pressures around the administration of screening. McPherson pointed out the relative paucity of research on suicide scales. In addition, existing suicide screening instruments are limited in assessing suicidal outcomes in their lack of standard nomenclature, lengthy administration time, lack of adequate psychometric evaluation, and unknown reliability and validity. There is also divergence over the most effective method to inquire about suicidality. Kaplan et al. found a generally high level of agreement between in person and self-report suicide assessments, except for recent suicidal ideation, where patients tended to disclose more on a self-report form. McAuliffe and Perry suggested that assessment should be embedded in a caring conversation, maximizing the use of a therapeutic alliance for facilitation, but other studies have found acceptance of straightforward inquiry.
Most research on suicide assessment tools have been conducted within psychiatric units, after the patient was identified as being at risk.,, Despite the increasing risk of suicidality in older populations and medically ill patients, the majority of studies focused on youth, particularly through school-based screening. These programs have demonstrated significant results in improving early identification and reducing suicide risk.,, Some studies have focused on other high-risk populations, such as prisoners, and high-risk ethnic/racial populations.,,,, Studies on screening have included primary care settings,,, emergency departments,,, and other medical settings (e.g., nursing homes).,, One of the most promising primary care studies is that by Simon et al., which retrospectively examined the predictive validity of Item 9 of the Patient Health Questionnaire-9 item scale (PHQ-9) for suicide attempts using data from over 84,000 outpatients in an electronic health record. They found that those patients who reported thoughts of death or self-harm on a daily basis had a ten times greater risk of a suicide attempt over the subsequent year than those who responded, “not at all.” It also proved a moderate predictor of suicide death over the subsequent year.
Utilizing MEDLINE and PsychInfo searches, up until very recently, there were no reports of systematic hospital-wide suicide screening in medical surgical hospitals. However, an important development in this realm, the Columbia Suicide Severity Rating Scale (C-SSRS), is a low-burden standardized screening tool for suicidal ideation and behavior that contains clear definitions now adopted by the U.S. Centers for Disease Control (CDC) and supported by robust psychometrics. The C-SSRS was developed within the multi-site NIMH-sponsored Treatment of Adolescent Suicide Attempters study to evaluate interventions to prevent repeat suicide attempts among adolescents, in response to the lack of an adequate assessment measure to identify high risk and severity of suicidal ideation and behavior. The C-SSRS has been used in a multitude of settings with adolescents and adults to evaluate safety and efficacy in research, assessment of suicide risk in communities, and quantification for surveillance efforts. It has been translated into over 110 languages and has been adopted, mandated, or recommended by many international regulatory agencies and institutions including the US Food and Drug Administration, the CDC, and the US Army. Many states and nations are conducting top-down policy efforts to mandate the use of the C-SSRS.
Posner et al. described the initial validity and internal consistency of the C-SSRS across three populations including adolescents and adults, with predictive, convergent, and divergent validity with other multi-informant suicidal ideation and behavior scales. It had a 99.4% specificity and a 100% sensitivity relative to the Columbia Suicide History Form in correctly identifying aborted attempts and 100% sensitivity and specificity for both interrupted and actual attempts. It also had a 100% sensitivity and 96% specificity for combined actual and interrupted attempts. Both the ideation and behavior subscales were sensitive to change over time. The intensity of ideation subscale demonstrated moderate-to-strong internal consistency. In the adolescent suicide attempters study, worst-point lifetime suicidal ideation on the C-SSRS predicted suicidal behavior during the study, whereas the Scale for Suicide Ideation did not. Participants who were rated at the two highest levels of suicidal ideation severity (intent or intent with plan) at baseline had higher odds for attempting suicide during the study. In another recent study of approximately 35,000 administrations among nonsuicidal patients across medical diagnoses, suicidal ideation and behavior reported on the C-SSRS predicted short-term subsequent suicidal behavior with two highest levels of ideation and behavior being nine times more predictive than without. One of the reasons for creation of the C-SSRS was to assess the full range of suicide behaviors with evidence that interrupted attempts, aborted attempts, and preparatory behaviors are equally predictive to a suicide attempt and constitute the majority of suicidal behaviors that individuals engage in. Recent findings from hospital settings using the C-SSRS demonstrated lower rate of false positives and lower rate of false negatives: the C-SSRS had 6.2% positive screens relative to 23.8% to that on the PHQ-9 and found cases that were missed by the PHQ-9. In the Emergency Department Safety Assessment and Follow-up Evaluation study, telephone follow-up assessments with the C-SSRS detected three times the number of suicide attempts when compared to chart review.
Suicide screening and triage response program at a Northeast US hospital
In response to the TJC National Patient Safety Goal on Suicide Risk, the Department of Nursing at a Northeastern US hospital (NEUSH) began a process of identifying an optimal screening tool to be used hospital wide. An interdisciplinary team developed a comprehensive suicide screening and response protocol in collaboration with the Department of Psychiatry and identified the C-SSRS as an evidence-based best practice that should be implemented in the institution. In consultation with Dr. Kelly Posner and her team at Columbia University, a C-SSRS Screener was developed that had a lower time burden than the full C-SSRS (2 to 3 min. The protocol received approval from the Patient Care Committee, Patient Safety Committee, and Medical Executive Committee, with a start date of December 8, 2009. A significant infrastructure was developed to implement the suicide screening process at NEUSH, including updated policies and procedures, for maintenance of patient safety, a clinical response protocol for patient screening positive at different levels of risk, an electronic version of the C-SSRS Screener inserted in the “Clin-Doc” electronic nursing record for administration by the primary nurse (embedded in the initial nursing assessment for all admissions), and a systematic education process for all nurses and consultation psychiatrists (C-SSRS Screener, full C-SSRS, and the clinical protocol). The clinical response protocol that was developed in association with the implementation of systematic suicide screening was consistent with recommendations by Dr. Posner and supported by the robust predictive findings regarding ideation severity of 4 and 5 and suicidal behavior predicting short-term risk of suicidal behavior, (See [Appendix 1], [Figure 1] for the scale and response protocol).
Triage responses with corresponding next steps to be taken were developed for the five levels of risk and presence of suicidal behaviors. When individuals endorse ideation or behavior assigned to two different triage responses, the highest numerical (1–4) triage response was taken. (1) No action taken – individuals who report no recent (past week, 1 week – 1 year) or lifetime (over a year ago) suicidal behavior and no suicidal ideation (risk level 0); (2) Mental Health Self-Referral – individuals at risk levels 1 or 2 (positive for questions 1 and/or 2, and/or having any lifetime suicidal behavior over 1 year prior [suicide attempt, interrupted suicide attempt, aborted suicide attempt, or preparatory behavior]) are recommended to self-refer to a mental health professional, with the recommendation entered in their discharge planning instructions and into a progress note notifying their attending physician, provided a list of mental health resources and a crisis hotline number; and (3) Care Team Consult – individuals with an ideation consistent with item 3, and/or positive for any recent (between a week and a year ago only) suicidal behavior on item 6 received consultation from a psychiatric nurse, termed a care team consult. This nurse reviewed the patient's history and observed behavior, briefly interviewed the patient, re-rated the C-SSRS Screener, and triaged the response either up to risk level 4 (psychiatric consultation) or down to risk level 2 (outpatient self-referral); (4) Psychiatric Consultation – individuals with an ideation severity positive for questions 4 and/or 5, and/or who had any recent (within the past week only) suicidal behavior on item 6 were automatically recommended for a psychiatric consult. The attending physician was notified both on a progress note and by a page from nurses (risk level 4/5).
Patients rated at risk levels 3 through 5 were placed on an environmental safety precaution protocol. This protocol includes patient placement in a private room with a discrete identifier for suicide risk monitoring for anyone entering the room, receiving a “safe” dietary tray (either plastic utensils or finger foods), and a patient safety monitor assigned to sit within arm's length of the patient at all times. The room is safety proofed by removing all potentially hazardous objects, including unnecessary furniture, tubing, cords, bags, and any liquids potentially harmful if ingested.
The “Clin-Doc” system automatically scored the C-SSRS Screener and followed the algorithm outlined in the Panel to generate an appropriate recommendation. If a psychiatric consult is ordered or a psychiatric nursing care team consult is pursued, the consultant probes into lifetime suicidal ideation and other indicators of suicidal risk using the full C-SSRS to decide patient disposition.
The C-SSRS Screener could be deferred until the patient is medically stable and cognitively able to answer questions. Conditions necessitating deferral include confusion/delirium, sedation/confusion from medication overdose or anesthesia/recovery, and temporary communication barriers (unavailable interpreters, aphasia post- transient ischemic attack (TIA), or stroke, etc.) If a patient demonstrated irreversible conditions, the screening is waived. These conditions include obtunded or comatose patients, patients with significant intellectual challenges (estimated intelligent quotient < 50), and patients with permanent communication barriers (permanent effect from a stroke, or severe dementia).
The goal of the evaluation of this program was to determine the feasibility of the C-SSRS Screener and response protocol in a hospital setting. The primary aims were as follows:
- To evaluate the interrater agreement on the C-SSRS Screener in a large heterogeneous sample of trained nurse raters
- To determine the demographic and suicide-related clinical characteristics of the RH patient population (as characteristic of a typical medical-surgical hospital)
- To evaluate the rates of false-positive or false-negative screens of ratings on the C-SSRS Screener definitions of ideation and behavioral reports
- To examine the level of care outcomes for patients downgraded or upgraded to a different suicide risk level during psychiatric consultation
- To determine the impact of implementing the C-SSRS Screener on patient safety monitor utilization and the number of suicide-related psychiatric consults.
| Methods|| |
A total of 837 nurses and student nurses (93% female) were educated on the C-SSRS Screener. The education level ranged from a high school diploma to a doctorate. Staff members' years of nursing experience ranged from less than 1 year to 43 years, with a mean of 11 and a median of 6 years. One hundred and twenty-eight participants had some mental health specialty training, while 161 had previously received training to evaluate suicidal risk. Additional participants in C-SSRS Screener training included all consultation psychiatrists and psychiatric nursing care team consultants.
A total of 42,251 patients 12 years of age or older admitted at NEUSH between December 8, 2009, and June 30, 2011, were screened. Data used in the present analysis were collected from July 1, 2010, to June 30, 2011, on a total of 26,747 patients 12 years and older.
Missing data from this analysis
The C-SSRS Screener was administered from December 8, 2009, to June 30, 2011. The time period chosen for data collection for this study was July 1, 2010, to June 30, 2011. Data collected during the first 7 months were not used due to several process changes to refine the screening procedure for deferrals, such as incorporating reassessments and a “hard stop” on the Clin-Doc medical record system that required the completion or deferral of the C-SSRS Screener on all patients.
Suicidal Ideation and Behavior – The C-SSRS Screener included the Suicide Ideation subscale from the original C-SSRS and one composite Suicide Behavior item [see previous description and [Appendix 1]. The C-SSRS Screener was administered at admission or deferred until the patient was able to respond. The standard form of the C-SSRS (which in addition to the severity of ideation, assesses suicidal behaviors individually, lethality of behaviors, and intensity of ideation) was administered by the care team psychiatric nurse and/or the consultation-liaison psychiatrist when a patient was positive on the C-SSRS Screener. For the purposes of this study, intensity and lethality were not included in the analyses.
A set of 14 hypothetical brief case vignettes with multiple-choice questions were used as a training posttest on raters' learning of the C-SSRS definitions (suicide attempt, aborted suicide attempt, interrupted suicide attempt, preparatory suicidal behaviors, nonsuicidal self-injurious behavior, wish to be dead, and ideation with plan and intent), as well as to evaluate interrater agreement on the C-SSRS Screener. This tool also included information on gender, level of education, and level of experience of each staff. Scores on vignettes of suicide scenarios (rated by nurses during education on the C-SSRS Screener) were used to evaluate training effectiveness.
Other data sources about care delivery included (1) number of psychiatry consults for suicidality during the time period of July 1, 2010, to June 30, 2011, and 6 months before implementation; (2) patient safety monitor utilization and incident reports involving suicidal or nonsuicidal self-harm behaviors on all patients, 6 months before and after implementation of the protocol; and (3) information on patients not screened due to deferrals or waivers.
Data management and analysis
Data were downloaded on a monthly basis from the NEUSH information system and used to evaluate:
- Demographic characteristics (age, gender, and race/ethnicity) across the patients assigned to the five risk levels of the C-SSRS Screener, using Chi-square analyses
- False positives (cases where minimal or low suicide risk was found by psychiatric consult) and false negatives (from incident reports on episodes of suicidal behavior) compared to the results of the C-SSRS Screener
- Changes in the rate of patient safety monitor shift utilization between 6 months prior to C-SSRS Screener and during screening protocol implementation
- Changes in the rate of psychiatric consults and appropriateness between prior year baseline and during C-SSRS Screener protocol implementation.
In addition to the above analyses, intraclass correlations (ICCs) were used to test posttraining interrater reliability of suicidal behavior and ideation definitions on the C-SSRS using hypothetical brief clinical case vignettes. Chi-square analysis was also performed to test correlations between the response accuracy to each of the 14 hypothetical training case vignettes and years of experience and education level of trainees. In order to maintain group sizes, work experience was consolidated into three grouping (unknown, 0–10, and 11 or more), and education level was consolidated into three groups (high school diploma and associate degree, bachelor's degree, and master's or doctoral level). Due to the exploratory nature of these tests, no multiple comparison corrections were calculated.
Human subjects' review
The evaluation of this clinical protocol was approved by the Institutional Review Board of the NEUSH, which granted a waiver of written informed consent as well as a waiver of authorization under the Health Information Portability and Accountability Act. All data collected from medical records were de-identified and aggregated to prevent the identification of any individual patients.
| Results|| |
Interrater reliability during training
A total of 814 staff responded to the 14-item post-test during the initial round of training on the C-SSRS Screener. The interrater reliability (ICC 3, 14; average measure) was 0.999 [Cronbach's alpha = 0.999; [Table 1] in [Appendix 2]. Interrater reliability on the Screener posttest did not differ by the prior educational level (0.999 below bachelor's degree, 0.998 for bachelor's degree, and 0.987 above bachelor's degree) [Table 2] in [Appendix 2]. Interrater reliability of raters did not differ across years of professional experience (0.997–0.999; Cronbach's alpha scores in the 0.998–0.999 range) [Table 2] in [Appendix 2]. Nurse raters who had previous mental health training or were trained on suicide risk assessment had slightly higher agreement than those without that background [Table 2] in [Appendix 2]. There was a small difference in interrater reliability based on rater gender [Table 2] in [Appendix 2], likely attributable to sample size (number of females far exceeded the number of males).
Examination of the relationship between performance for each training vignettes and clinical experience level indicated a statistical association with years of experience on 4 of the 14 vignettes (29%), with higher experience levels associated with more incorrect responses. This would indicate that nurses can be trained to reliability regardless of years of experience. It may also suggest that more senior clinicians make more incorrect inferences than less experienced staff. There were no significant associations between education level and accuracy of responses in any of the 14 case vignettes. This result would also indicate that education level had no impact on the ability to be trained to reliability on the screening tool.
Demographics and rating results for screened patients (July 1, 2010, to June 30, 2011)
A total of 26,747 patients were screened. Of that number, the patients who were deferred because they could not answer the questions at that time totaled 1428 and the patients who were deemed unable to answer the questions ever totaled 1151. Therefore, a total of 24,168 patients 12 years of age or over were evaluated from July 1, 2010, to June 30, 2011, and received a C-SSRS Level of 0–5. The gender breakdown for the evaluated group was roughly equal at 12,277 females (50.8%) and 11,891 males (49.2%). The race/ethnicity breakdown of the evaluated group was as follows: 20,306 non-Hispanic Whites; 1057 non-Hispanic African-Americans; 228 non-Hispanic other; 1358 Hispanic Whites; 685 Afro-Hispanics; 320 Hispanic other; 6 American Indian/Eskimo; 24 Asian-Pacific Islanders; and 184 unknown.
Of all patients evaluated, 89 patients were rated at Level 3; 17 at Level 4, and 144 at Level 5 using the C-SSRS Screener. Therefore, the C-SSRS Screener indicated only 1.03% (250/24,168) of admissions as needing any type of consult. Another 1083 admissions or 4.48% of admissions (1083/24,168) were rated at Levels 1 or 2 and received the county crisis number and mental health referral information. Most patients where a consult was recommended had a positive mental health history per se lf-report on the psychiatric consultation. Of those rated at Level 3 who received a consultation, about 85.5% (71/83) self-reported a previous history of mental health treatment, whereas of those rated at Levels 4 or 5 receiving a psychiatric consultation, 73.3% (118/161) self-reported a previous history of mental health treatment. Gender (female; χ2 = 36.857, P < 0.001), race/ethnicity (non-White; χ2 = 107.611, P < 0.001), and Latino status (χ2 = 33.818, P < 0.001) were all significantly correlated to C-SSRS Screener level scores. African-Americans did not have a significantly higher C-SSRS Screener score, and the number of other racial/ethnic groups was too small to analyze. These findings are consistent with the emerging data on suicidality in diverse populations.,,
Of the 1428 patients with a C-SSRS Screener deferred at admission, 377 were assessed when the patient became able to answer questions. The rest were permanently unable to be assessed due to their medical condition. Of these 377 patients, psychiatric consults were recommended by the C-SSRS Screener for 7 deferred patients (Levels 3, 4, or 5; 1.9%), with 12 patients rated at Levels 1 or 2 (3.2%). Only five patients out of the total screening sample refused to answer the screener questions (entered on the electronic medical record), demonstrating a very high level of acceptability (24,163/24,168 or 99.98%). As per the NEUSH protocol, those five patients received a full psychiatric consultation evaluation and were all found to have a psychiatric disorder. Four of the five were initially sedated or comatose from overdose attempts. Four of the five were admitted psychiatrically, and one was referred to outpatient care.
Screening false negatives
There were 26 incidents of patients attempting to harm themselves in the hospital in 2010. Of these, 23 were on the psychiatric unit (where the C-SSRS Screener was not in use at that time). Of the remaining three, one patient who was identified at a Level 0 attempted to harm him/herself but was not suicidal. A second patient, rated at Level 4, was already identified and on a 1–1 observation, attempted self-harm using a small object with a patient safety monitor within reach. The third patient was rated at Level 0 and attempted to harm himself during alcohol withdrawal. There was only one patient who attempted suicide 2 weeks after leaving the hospital who was not rated at risk by the C-SSRS Screener (risk level 0); the patient had consistently denied suicidal ideation during the hospital stay. Lack of integration of collateral reports, which is an important aspect of the C-SSRS, may have contributed to not identifying this patient. Aside from this unfortunate post-discharge event, out of 24,747 patients screened, there were no reports of suicide attempts or completions by patients screened using the C-SSRS Screener from July 1, 2010, to June 30, 2011.
Of the 89 Level 3 consults recommended by the C-SSRS Screener, 69 consults were undertaken by the Care Team and 14 by a psychiatrist. Most patients rated at Level 3 (53/83; 63.9%) were downgraded to a Level 1 or 2, and the rest (30/83; 36.1%) were upgraded to a Level 4, whereas 6 patients at Level 3 received neither consult. Of the 161 patients rated at Level 4 or 5 (where psychiatric consults were recommended by the screening), 145 had psychiatric consults ordered, whereas 16 patients did not receive a psychiatric consultation based on the judgment of their attending physician. Just over half of patients who received a psychiatric consult either from a Level 4 or 5 or an upgraded Level 3 rating were admitted to a psychiatric unit (50.3%; 73/175). Thirty-one patients (21.4%; 31/175) received a direct outpatient mental health referral and were given the county crisis service number. The rest had a variety of other dispositions resulting from their medical condition and clinical judgment of the consulting psychiatrist. [Table 3] [Appendix 3] outlines the full range of dispositions resulting from consults indicated by the C-SSRS Screener screen.
Brief versus full Columbia Suicide Severity Rating Scale ratings
We were only able to collect concurrent full C-SSRS ratings on 66 patients (of 227 or 29.1%) seen on consultation by either the psychiatric nurse or the psychiatrist. The C-SSRS Screeners were rated at a mean of 0.64 risk levels higher than that the full C-SSRS on those patients (one-tailed t-test of −3.701 [P< 0.001]). The median risk levels were identical for both (3.0). This finding may have been different if the full C-SSRS was performed on every consult. However, results are consistent with the C-SSRS Screener operating as a screen (i.e., having a lower threshold for identification by a less experienced rater than a more detailed follow-up assessment by a more experienced mental health professional).
Impact on care delivery and service utilization
Rate of psychiatric consults over time
During the time period of July 1, 2010, to June 30, 2011, when the C-SSRS Screener was implemented, the psychiatry consult service saw 2495 consults. A total of 1736 of these were initial consults. By comparison, during July 1, 2009, through June 30, 2010, the psychiatry consultation service saw 2486 consults, of which 1166 were initial consults. Of the 1736 initial consults for the July 1, 2010, to June 30, 2011 period, 353 were for suicidal ideation or suspected suicide attempt, with 39.9% (141/353) being for a C-SSRS rating of 4/5. Another eight were consults requested for a Level 3 and done by our psychiatric nursing care team. The remaining 204 of the 353 consults were requested for suicidal ideation or suspected suicide attempts without the C-SSRS Screener. Of those, 82 were deferred at admission, as the patients were not able to answer, 50 had a C-SSRS rating of 0, 58 had a C-SSRS level of 1 to 4, and 14 were in areas where the C-SSRS Screener was not implemented. Of the 50 with a C-SSRS Screener rating of 0, 30 reported suicidal ideation later in their hospital stay and received a consult. The remaining 20 denied suicidal ideation at admission but had admitting injuries such as a stab wound or a medication overdose that the admitting physician suspected as a suicide attempt which triggered a consult.
Rate of utilization of patient safety monitors
After initiation of the suicide screening protocol in the first quarter of 2010, use of Patient Safety Monitors began to decrease compared to the baseline during the fourth quarter of 2009 (508 staff shifts compared to 585). However, since the initial decrease, use of Patient Safety Monitors increased to 586 staff shifts [Appendix 4] [Figure 1]. In tandem, the number of consults for overdoses and suicide attempts decreased during the first quarter of 2010 (43 in the 4th quarter of 2009; 21 in the 1st quarter of 2010) and subsequently increased (41) in the second quarter of 2010 [Appendix 4] [Figure 2]. The latter increase may be partially explained by the economic downturn in the area NEUSH is located and its impact on mental health during this timeframe.
| Discussion and Conclusions|| |
This study is one of the first reports of a systematic suicide screening program in a general medical-surgical hospital. It meets the latest standard recommended by the TJC in its recommendations for addressing the Patient Safety Goal around suicide in all health-care populations, both in its use of hospital-wide systematic screening and the development of a clinical protocol and related policies and procedure. The protocol developed by the NEUSH was also consistent with many empirically based recommendations for managing suicide risk in primary care.
This study reports on the feasibility and initial outcomes of systematic screening. Inherent limitations, given its real-world setting, include lack of concurrent measures for external validity; limited numbers of consultation patients not rated with the full C-SSRS; and the typical challenges (particularly clinician variability) around adherence to a systematic practice in a clinical setting. The use of the electronic health record as the screening platform facilitates administration and scoring of the C-SSRS Screener, as well as the rapid implementation of a clinical response algorithm. The latter is essential to ensure consistent response to patients with high suicidal risk.
The C-SSRS Screener makes systematic hospital-based suicide screening feasible, addressing barriers identified by Posner et al., and other authors concerning time burden for administration (2–3 min on average). The low refusal rate by patients validates applicability in a general medical surgical setting. The total number of consults did not rise, but there was increase in new consults as a result of the constant number of patients admitted with actual suicide attempts in addition to those identified by systematic screening as a result of ideation or past suicidal behavior. However, screening algorithm assisted in the shifting of psychiatric staff to address patients with higher levels of risk and need. The increase in workload is worthwhile given the low numbers of self-harm or suicidal attempts in the hospital or shortly thereafter.
The challenge around screening results is around the problem of patients who deny or conceal suicidal ideation or intent, many times in the face of documented serious self-injurious acts. The NEUSH protocol already calls for psychiatric consultation for patients who refuse to cooperate with the C-SSRS Screener, with significant results in identifying the underlying intent and psychopathology. The other lessons learned from the implementation of this protocol include (a) obtaining and integrating family or significant other collateral information in rating the C-SSRS Screener and (b) rating the C-SSRS Screener using the self-injurious act itself (an automatic Level 4/5) if suicide intent is suspected. Both are important fail-safes in implementing such screening protocols. In fact, the newest version of the C-SSRS Screener has extended the cutoff for the highest risk rating from suicidal behavior to the last 3 month versus the 1 week cutoff we tested.
The initial reduction in patient care monitor utilization was reversed in the latter part of the evaluation period, likely due to increases in medical admissions for suicide attempts that coincided with the economic crisis that hit the area that NEUSH is located in, with a historically high (9%) unemployment rate in July 2010, in comparison with a 5.5% rate in 2008. This necessitated higher numbers of patient care monitor shifts, especially as psychiatric inpatient beds were extremely full.
The protocol developed by NEUSH using the C-SSRS Screener and response algorithm is a potential model for addressing suicide risk through health-care settings. It can be readily replicated in medical-surgical settings across the globe even in low- and middle-income nations (LMIC) with minimal burden, adapting the response algorithm to the local mental health resources available in the community. This protocol would make a significant contribution not only to the prevention of suicide in such settings, but also to the reduction of stigma around suicide, legitimizing it as a health concern worthy of attention in general medical settings.
Our results suggest that the C-SSRS Screener demonstrates similar properties in a real-world setting as the full C-SSRS in both experimental and real-world settings. It also demonstrated good interrater reliability of the C-SSRS definitions among a large and diverse group of nurse raters, including those with no mental health background or training. Ratings using the C-SSRS Screener by less experienced individuals are significantly higher than ratings by mental health professionals (psychiatrists and psychiatric nurses), consistent with the operation of the C-SSRS Screener as a screen (lower thresholds for less experienced raters). These characteristics are important in assuring that the tool can be administered by individuals with a wide range of training and experience. In fact, the Columbia Lighthouse Project, which now promotes the use of the C-SSRS both clinically and in prevention, recommends the use of the C-SSRS Screener for lay people in multiple settings, such as schools, workplaces, the military, and among concerned relations of those possibly affected by suicidal ideation. This suggests additional utility for LMIC nations that rely on para-professionials and lay health workers.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Luoma JB, Martin CE, Pearson JL. Contact with mental health and primary care providers before suicide: A review of the evidence. Am J Psychiatry 2002;159:909-16.
King CA, O'Mara RM, Hayward CN, Cunningham RM. Adolescent suicide risk screening in the emergency department. Acad Emerg Med 2009;16:1234-41.
McPherson A. An overview of the assessment tools available to mental health professionals to help determine patients at risk of suicide. Int J Psychiatr Nurs Res 2005;10:1129-42.
Hermes B, Deakin K, Lee K, Robinson S. Suicide risk assessment: 6 steps to a better instrument. J Psychosoc Nurs Ment Health Serv 2009;47:44-9.
Kaplan ML, Asnis GM, Sanderson WC, Keswani L, De Lecuona JM, Joseph S. Suicide assessment: Clinical interview vs. self-report. J Clin Psychol 1994;50:294-8.
McAuliffe N, Perry L. Making it safer: A health centre's strategy for suicide prevention. Psychiatr Q 2007;78:295-307.
Ballard ED, Bosk A, Snyder D, Pao M, Bridge JA, Wharff EA, et al
. Patients' opinions about suicide screening in a pediatric emergency department. Pediatr Emerg Care 2012;28:34-8.
Lung FW, Lee MB. The five-item Brief-Symptom Rating Scale as a suicide ideation screening instrument for psychiatric inpatients and community residents. BMC Psychiatry 2008;8:53.
Lang M, Uttaro T, Caine E, Carpinello S, Felton C. Implementing routine suicide risk screening for psychiatric outpatients with serious mental disorders: II. Quantitative results. Arch Suicide Res 2009;13:169-77.
Mazza J. School-based suicide prevention programs: Are they effective? Sch Psychol Rev 1997;3:382-97.
Shaffer D, Craft L. Methods of adolescent suicide prevention. J Clin Psychiatry 1999;60 Suppl 2:70-4.
Hallfors D, Brodish PH, Khatapoush S, Sanchez V, Cho H, Steckler A. Feasibility of screening adolescents for suicide risk in “real-world” high school settings. Am J Public Health 2006;96:282-7.
Mills J, Kroner D. Screening for suicide risk factors in prison inmates: Evaluating the efficiency of the Depression, Hopelessness and Suicide Screening Form (DHS). Legal and Criminological Psychology 2005;10:1-12.
Konrad N, Daigle MS, Daniel AE, Dear GE, Frottier P, Hayes LM, et al
. Preventing suicide in prisons, part I. Recommendations from the international association for suicide prevention task force on suicide in prisons. Crisis 2007;28:113-21.
Gary FA, Baker M, Grandbois DM. Perspectives on suicide prevention among American Indian and Alaska native children and adolescents: A call for help. Online J Issues Nurs 2005;10:6.
Lafromboise TD, Lewis HA. The Zuni life skills development program: A school/community-based suicide prevention intervention. Suicide Life Threat Behav 2008;38:343-53.
Niven JA. Screening for depression and thoughts of suicide: A tool for use in Alaska's village clinics. Am Indian Alsk Native Ment Health Res 2007;14:16-28.
Mullany B, Barlow A, Goklish N, Larzelere-Hinton F, Cwik M, Craig M, et al
. Toward understanding suicide among youths: Results from the white mountain apache tribally mandated suicide surveillance system, 2001-2006. Am J Public Health 2009;99:1840-8.
Brown M, Grumet J. School-based suicide prevention with African American youth in an Urban setting. Prof Psychol Res Pract 2009;2:111-7.
Wintersteen MB. Standardized screening for suicidal adolescents in primary care. Pediatrics 2010;125:938-44.
Uebelacker LA, German NM, Gaudiano BA, Miller IW. Patient health questionnaire depression scale as a suicide screening instrument in depressed primary care patients: A cross-sectional study. Prim Care Companion CNS Disord 2011;13. pii: PCC.10m01027.
Simon GE, Rutter CM, Peterson D, Oliver M, Whiteside U, Operskalski B, et al
. Does response on the PHQ-9 depression questionnaire predict subsequent suicide attempt or suicide death? Psychiatr Serv 2013;64:1195-202.
Horowitz LM, Wang PS, Koocher GP, Burr BH, Smith MF, Klavon S, et al
. Detecting suicide risk in a pediatric emergency department: Development of a brief screening tool. Pediatrics 2001;107:1133-7.
Mitchell AM, Garand L, Dean D, Panzak G, Taylor M. Suicide assessment in hospital emergency departments: Implications for patient satisfaction and compliance. Top Emerg Med 2005;27:302-12.
Folse VN, Hahn RL. Suicide risk screening in an emergency department: Engaging staff nurses in continued testing of a brief instrument. Clin Nurs Res 2009;18:253-71.
Reiss NS, Tishler CL. Suicidality in nursing home residents: Part I. Prevalence, risk factors, methods, assessment, and management. Prof Psychol Res Pract 2008;3:264-70.
Reiss NS, Tishler CL. Suicidality in nursing home residents: Part II. Special issues. Prevalence, risk factors, methods, assessment, and management. Prof Psychol Res Pract 2008;39:271-75.
Garand L, Mitchell AM, Dietrick A, Hijjawi SP, Pan D. Suicide in older adults: Nursing assessment of suicide risk. Issues Ment Health Nurs 2006;27:355-70.
Posner K, Brown GK, Stanley B, Brent DA, Yershova KV, Oquendo MA, et al
. The Columbia-suicide severity rating scale: Initial validity and internal consistency findings from three multisite studies with adolescents and adults. Am J Psychiatry 2011;168:1266-77.
Mundt JC, Greist JH, Jefferson JW, Federico M, Mann JJ, Posner K. Prediction of suicidal behavior in clinical research by lifetime suicidal ideation and behavior ascertained by the electronic Columbia-suicide severity rating scale. J Clin Psychiatry 2013;74:887-93.
Viguera AC, Milano N, Laurel R, Thompson NR, Griffith SD, Baldessarini RJ, et al
. Comparison of electronic screening for suicidal risk with the patient health questionnaire item 9 and the Columbia suicide severity rating scale in an outpatient psychiatric clinic. Psychosomatics 2015;56:460-9.
Arias SA, Miller I, Camargo CA Jr., Sullivan AF, Goldstein AB, Allen MH, et al
. Factors associated with suicide outcomes 12 months after screening positive for suicide risk in the emergency department. Psychiatr Serv 2016;67:206-13.
Shrout PE, Fleiss JL. Intraclass correlations: Uses in assessing rater reliability. Psychol Bull 1979;86:420-8.
Pumariega AJ, Rothe EM, Swanson J, Holzer CE, Linskey AO, Quintero-Salinas R. Suicidality and acculturation in Hispanic adolescents. In: Sher L, Vilens A, editors. Immigration and Mental Health. New York, NY: Nova Science Publishers; 2010.
Crosby AE, Han B, Ortega LA, Parks SE, Gfroerer J, Centers for Disease Control and Prevention (CDC). Suicidal thoughts and behaviors among adults aged≥18 years–United States, 2008-2009. MMWR Surveill Summ 2011;60:1-22.
Kann L, McManus T, Harris WA, Shanklin SL, Flint KH, Hawkins J, et al
. Youth risk behavior surveillance – United States, 2015. MMWR Surveill Summ 2016;65:1-74.
Bryan C, Corso K, Neal-Walden T, Rudd M. Managing suicide risk in primary care: Practice recommendations for behavioral health consultants. Prof Psychol Res Pract 2009;2:148-55.