Article Text
Abstract
Background This is the first study of suicide attempts and completions in the emergency department (ED) in a large national medical system.
Methods All root cause analysis (RCA) reports completed of suicides and suicide attempts that occurred in ED in the Veterans Health Administration between 1 December 1999 and 31 December 2009 were reviewed. The method, location, anchor point for hanging and implement for cutting as well as the root causes were categorised.
Results Ten per cent of all RCA reports of suicides and suicide attempts that occur within the hospital occur in the ED. Hanging, cutting and strangulation were the most common methods. The most common anchor point for hanging was doors, and the most common implement for cutting was a razor blade. In eight of the 10 cases of cutting, the implement was brought into the ED. The most common root causes were problems communicating risk and being short-staffed.
Conclusions Based on these results the following recommendations are made for helping to reduce suicide attempts in the ED: (1) use a systematic protocol and checklist to review mental health holding areas periodically in the ED for suicidal hazards; (2) develop and implement specialised protocols for suicidal patients that include continuous observation when possible; (3) conduct thorough contraband searches with suicidal patients; (4) designate specialised holding areas, when practically possible, for suicidal patients that are free of anchor points for hanging, sharps and medications, and medical equipment; and are isolated from exits to reduce the risk of elopement.
- Mental health
- self-harm
Statistics from Altmetric.com
Suicide is the eleventh leading cause of death in the USA,1 resulting in the deaths of over 32 000 people each year.2 In 2003 the American Psychiatric Association reported that approximately 1500 suicides take place within hospital facilities in the USA each year, and one third of these take place while the patient is on 15-min checks.3 Dong et al4 found an inpatient suicide rate of 269 suicides per 100 000 psychiatric admissions in Hong Kong, while Shapiro and Waltzer5 reported rates of between five and 80 per 100 000 psychiatric admissions in the USA. A Joint Commission on Hospital Accreditation review of inpatient suicides in the USA found that 75% involved hanging and another 20% resulted from patients jumping from a roof or window.6 Other studies of inpatient suicides include patients who committed suicide while on pass or eloping from the hospital so it is difficult to discern the specific environment in which patients committed suicide7–9; however, all report hanging and jumping to be the most common methods. A recent study of inpatient suicides in Veterans Affairs (VA) hospitals also found that hanging was the most common method of inpatient suicide in VA hospitals (accounting for 43% of all inpatient suicides) and that doors and wardrobe cabinets accounted for 41% of the anchor points used when hanging was the method of self-harm.10
Suicide is of particular concern within the Department of Veterans Affairs. Lambert and Fowler11 have noted an increased prevalence among veterans of the common risk factors for suicide including male gender, age over 65 years, poor physical health, mental illness, poor social support and firearm availability. Although it is difficult to count completed suicides accurately in the USA,12 13 and methodologies differ for estimating the suicide rate among veterans,14 15 Kaplan et al15 reported that veterans are twice as likely as non-veterans to take their own lives.
In 2005, in the USA 372 722 people were treated in emergency departments (ED) for self-harm;2 however, there are few studies of suicides or suicide attempts that occur in ED. Mahal et al16 conducted a retrospective chart review of 145 patients involuntarily admitted as a ‘danger to self’. The authors concluded that many important risk factors for suicide are not documented as part of the initial ED assessment; for example, current suicidal ideation, plan, history of attempts and hopelessness were documented in less than 70% of all cases. Buzan and Weisburg17 report that all states have statutes permitting detention by physicians of self-destructive patients and recommend one-on-one observation, removal of potentially lethal objects such as scissors, and prevention from jumping to death or hanging once a careful assessment has revealed suicidal intent. Finally, the joint commission recently reported that 8.02% of all inpatient suicides reported to the sentinel event database in the USA occur in the ED.18
Whereas other studies have described the specific characteristics of patients who have committed suicide while in the hospital,19–22 or analysed environmental factors relevant to inpatient suicides or suicide attempts,3 6 10 this study describes suicide attempts in the ED and identifies environmental hazards associated with increased suicide risk in this setting. It is our perspective that studying specific types of adverse events can lead to systematic safety improvements in medicine.
Methods
Study design and theoretical model
This is a retrospective review of all root cause analysis (RCA) reports of suicide attempts or completions in the VA healthcare system between 1 December 1999 (when the RCA system started) and 31 December 2009.
The VA National Center for Patient Safety RCA programme
The VA healthcare system provides comprehensive healthcare services to over 6 million veterans across the USA through 154 VA medical centres. Patient safety, the development of safety culture and the investigation of adverse events is coordinated by the National Center for Patient Safety (NCPS) along with patient safety officers at the Veteran Integrated Service Networks and patient safety managers at each VA hospital. The NCPS has instituted a RCA programme to analyse adverse events individually and collectively.23 24
RCA is a powerful method for examining the underlying causes of an adverse event such as a hospital-related death, surgical error or suicide. The focus of a RCA is on the systemic and organisational factors that may have contributed to an adverse event, including environmental factors, breakdowns in communication of critical information from one clinician to another, non-standardised processes for assessing or treating patients, training and fatigue.23 24 The NCPS broadly defines adverse events as ‘untoward incidents, therapeutic misadventures, iatrogenic injuries or other adverse occurrences directly associated with care or services provided within the jurisdiction of a medical centre, outpatient clinic or other facility’.25 Adverse events may result from acts of commission or omission; examples include the administration of the wrong medication, failure to make a timely diagnosis or institute the appropriate therapeutic intervention, adverse reactions or negative outcomes of treatment. Hospital-level patient safety managers are responsible for reporting adverse events to the NCPS. The RCA process within VA hospitals is conducted by multidisciplinary facility teams organised by the facility's patient safety manager where the adverse event occurred. The RCA team is composed of clinicians and others who are knowledgeable about the systems and process under review in the area but are not directly involved in the case. The RCA is guided by specific tools provided by the NCPS. In general, an RCA describes what happened, how it happened and what should be done to avoid the same event happening again. The focus is on systemic changes that can be made at the unit, clinic or facility level to reduce the chances of the same type of event happening again.26 Details about the patients or clinicians involved are purposely left out of RCA reports in order to keep the process away from blaming clinicians or patients; all statements of cause must follow the following rules of causality: (1) root cause statements must clearly show the ‘cause and effect’ relationship; (2) negative descriptions should not be used; (3) each human error must have a preceding cause; (4) violations of procedure are not root causes; they must have a preceding cause; (5) failure to act is only causal when there was a pre-existing duty to act.27
Because of the focus on the system, rather than possible errors made by clinicians, the information contained in the RCA reports does not generally have demographic or epidemiological data about the patients involved in the events. RCA reports that are submitted to the NCPS include narrative descriptions of the event, all contributing factors, a final understanding of the event, and a specific action plan for addressing underlying causes of the event. RCA reports must be completed by the facility with 45 days of the event.
All adverse events that are reported within VA hospitals are rated by the hospital's patient safety staff against two criteria: what harm occurred to the patient (from catastrophic to minor) and what is the probability or likelihood that such an event would occur again (from frequent to remote). Each event is coded both for the actual harm that occurred and the potential harm that could have occurred. Harm and probability are combined to produce a safety assessment code score from 1 to 3,23 24 and those events rated at a level of 3, the most significant risk, are required to undergo a RCA. All completed suicides in the ED require a RCA. RCA reports come into the VA NCPS via a secure computerised reporting system where they are categorised and available for review by NCPS staff.
Analysis of RCA reports
Our goal was to identify suicide attempts and completed suicides that occurred in the emergency room or ED. We conducted a search of all RCA reports received between 1 December 1999 and 31 December 2009 to identify any events that involved suicides or suicide attempts that occurred while the patient was being treated in an ED in a VA hospital. The search was completed through the use of both event code for suicide or suicide attempt entered in the RCA and the use of natural language processing software to identify the terms related to suicide or suicide attempt anywhere in the RCA text (PolyAnalyst, Megaputer, Bloomington, Indiana).
Once the initial search was complete, a list of RCA reports of interest was created by excluding the following: those not involving suicides or suicide attempts and those occurring in any area except the ED or other emergency care setting, for example, urgent care centre. The RCA reports that were the subject of our more detailed review included suicide attempt and suicide in the ED or other emergency care setting. For this review, ‘suicide attempt’ was defined to be an incomplete suicide in which action had been taken (eg, neck in noose, cut wrists) verses a threat or gesture.
Data processing
Each RCA report was coded for the location of the event and the method of suicide or suicide attempt. For cases of hanging, the type of anchor point and lanyard material were coded; for cases of cutting the type of cutting implement and whether the implement was brought into the ED or procured in the ED was coded. For each RCA report that identified the ED as the location, the root causes identified were listed. The coding system was developed in previous studies of RCA reports of suicide (coefficient κ=0.96 (28)). In addition, the remainder of the text of the RCA reports was reviewed to identify demographic information about the adverse event. Whereas much of this information was incomplete (due to the design of the RCA process) we were able to compare the age and gender of the ED cases with those cases of inpatient suicides and suicide attempts in the psychiatry unit.
Results
Of the inpatient suicides (n=59) and suicide attempts (n=291) initially identified in RCA reports, 37 (10.57%) occurred in the ED: 87.1% of the patients were men with an average age of 52.1 years. For comparison, over the same time period there were 177 RCA reports of inpatient suicides or suicide attempts on psychiatry units, of these, 80.5% were men and the average age was 45.7 years. Analysis of variance comparison between these two groups reveals a significant difference in age (F=5.44, p=0.021) but not gender (F=0.75, p=0.389). There was one RCA report of a completed suicide in the ED during this time period. Figure 1 displays the methods for these 37 events. Note that hanging accounts for 32.4% of these events, followed by cutting (27.0%), strangulation (13.5%), stabbing (10.8%) and overdose (10.8%). The single RCA report of a completed suicide in the ED involved an overdose. Figure 2 displays the anchor points used for the 12 cases of hanging. The primary materials used for lanyards were clothing, belts and bedding (three cases each), while the call cord, window blind cord and medical equipment were used once each. The implements used in the 10 cases of cutting include: razor blade (four), knife (three), scalpel (one), glass (one) and scissors (one). In eight of the 10 cases of cutting the implement was brought into the ED by the patient. Finally, we found that 12 of the 37 attempts (32.4%) took place while the patient was using the toilet room or being afforded extra privacy because they were attempting to use the urinal while in their room. Of these, six were cutting or stabbing, five were hanging or strangulation and one was an overdose.
The patient safety managers conducting the RCA identified 96 contributing factors/root causes in the 37 RCA reports (figure 3). Most of the RCA reports reviewed had multiple contributing factors or root causes. Note that ‘problems with communication of risk’ make up 14.6% (14 of 96) of the root causes, while being ‘short-staffed’, problems with the contraband search and problems with the physical layout of the ED each make up 9.4%.
Discussion
This study examines the specific environmental factors involved in suicide attempts and completions in an ED in a large, nationally represented hospital sample (VA). In this healthcare setting the ED had the second highest number of reported completed suicides and attempts, second only to inpatient psychiatric units (medical units, domiciliaries and nursing home care units were third, fourth and fifth, respectively). Therefore, understanding more about the method of suicide and suicide attempt, as well as the factors contributing to these events is an important initial step in eliminating these preventable adverse events.
As with other studies of RCA,28 29 we found that ‘problems communicating risk’ was the most common identified root cause for suicide attempts and completions in the ED. ED handoffs, in which specific information, responsibility and authority for patient care are exchanged, are vulnerable occurrences30 31 that have been associated with adverse events.32 Specific information related to suicide risk may represent important information that is miscommunicated between ED providers and nursing staff, particularly during a handoff in patient care. When critical information about a patient's level of suicide risk is lost appropriate monitoring of the patient and other safeguards may not be established during the ED visit.
Another important contributing factor identified was not enough staff to provide immediate one-to-one observation for suicidal patients (labelled ‘short-staffed’ in figure 3). Whereas one-to-one supervision of suicidal patients has been recommended in other literature,17 consistently providing this level of staffing may be challenging due to competing demands on ED personnel (eg, clinical care, patient education, point-of-care testing and patient transport), ED overcrowding and acuity level of patients. Whereas departmental policy may outline one-to-one supervision of suicidal patients, interdepartmental assistance or resources may be needed to implement this action successfully at all times.
Problems with the structural and operational aspects (eg, contraband search, system for managing suicidal patients) of the ED were also identified as contributing to suicide attempts and completions. We found that 32.4% of the suicide attempts in the ED were hangings and patients used a variety of anchor points and lanyards. As with our study of overall inpatient suicides,10 doors are the most common anchor point for hanging in the ED. Because of this hazard, it is important to hold suicidal patients in rooms that have no unobserved interior doors, including locker or cabinet doors. Closing a knotted sheet in the top or side of a door creates an anchor point as do door handles and hinges. Beyond doors, any protrusion from the wall or ceiling can serve as an anchor point, and all must be removed from areas in which suicidal patients may be held without being under direct observation by staff. It is important to note that in one review of studies of hanging, 50% of the lanyards were suspended from points that were below the patient's head,33 so it is critical to identify any anchor points above the floor. Anchor points near the floor can be used by attaching a lanyard and then looping it over a higher object such as the back of a chair. As almost any type of clothing or bedding (including belts, shoelaces and sheets) can and will be used as a lanyard, the most effect way to reduce the environmental hazards is to make efforts to eliminate the anchor points.
The second most common report of attempted suicide in VA ED was cutting with a sharp object. Razor blades and pocket knives were the most common type of cutting implement, but any sharp object can be used. Our data point towards two important strategies for managing patients who may cut themselves: First, as 80% of the cutting implements were carried into the ED from the outside, it is important to conduct a thorough contraband search with all potentially suicidal patients. This issue is also identified as one of the top root causes for suicide attempts in the ED. Second, make sure the holding room does not have any sharp objects with which the patient can cut themselves. Again, access to equipment that can be used for self-harm was also an identified root cause.
In our study of the use of a checklist to identify suicidal hazards on psychiatry units in VA hospitals,34 we found that the most hazardous room on the unit was the bathroom. We believe this is because patients have the most unobserved time in the bathroom and so can more easily harm themselves without intervention from others. In the current study we also found that 32.4% of the suicide attempts took place either in the toilet, or while the patient was afforded extra privacy because they were attempting to urinate. If the ED does not have a specific protocol for observing suicidal patients using the toilet room, it is critical to have a toilet room for patient use that is free of anchor points and sharps. We developed a specific checklist to review mental health areas, including toilet rooms and shower rooms, for environmental suicide hazards. We recommend that consideration be given to using the appropriate sections of the checklist also to review mental health areas of the ED.34 35
One issue that our data do not capture is the number of suicidal patients who were able to elope or leave the ED and attempt suicide once outside. It can be easy for suicidal patients to leave the area if they are not closely monitored. It is recommended that once patients are determined to be potentially suicidal they are monitored at all times until further evaluation can be conducted to establish that they are either safe to leave or are transferred to a more secure unit. That being said, it is easy in a busy ED to lose track of patients unless there is a specific protocol to move suicidal patients to a secure holding area that is free of anchor points for hanging, sharps, medications, and physically removed from exits to reduce the chances of elopement. These problems were identified as important root causes (short-staffed, inadequate holding area for suicidal patients and problems with the physical layout) for the suicide attempts and completions identified in these RCA. ‘Physical layout problems’ came up because staff had difficulty monitoring patients because of corners or other blocks to viewing patients. ‘Inadequate holding area’ came up when the RCA team found that there was not a safe place to hold suicidal patients in their ED.
Limitations
This study has several limitations: first, our data only contain suicide attempts and completions that were reported through our patient safety system, thus there are likely to be some events that we do not know about. It is probably the case that some suicide attempts that cause little or no harm were not reported. Second, the RCA reports focus on the systemic vulnerabilities in the ED and the hospital that may have contributed to the adverse event rather than the specific characteristics of the patients involved, so we do not have good information about these individual patients. There is, however, a sizeable literature on the characteristics of patients who are at risk of suicide3 where this information can be found. Third, this paper focused more on the environmental hazards for suicide in the ED rather than on the possible effectiveness of staff protocols for reducing suicide attempts. Finally, we are only reporting on events that took place in the VA hospital system in the USA, so the results may not generalise well to other hospital systems or other countries.
Conclusions and recommendations
These limitations notwithstanding, this is the first study to present data on suicide attempts and completions in the ED in a large national medical system. Based on these results, and building upon the recommendations of Buzan and Weisburg,17 we make the following recommendations for helping to reduce suicide attempts in the ED:
Use a systematic protocol and checklist to review mental health holding areas periodically in the ED for suicidal hazards (see http://www.patientsafety.gov/SafetyTopics.html#mheocc for an example of a checklist developed for units treating suicidal patients).
Develop and implement specialised protocols for suicidal patients that include continuous observation when possible.
When possible, conduct thorough contraband searches with suicidal patients. Note that this recommendation may be limited by the local laws governing physical searches.
Designate specialised holding areas, when practically possible, for suicidal patients that are free of anchor points for hanging, sharps and medications, and medical equipment, and are removed from exits to reduce the risk of elopement.
References
Footnotes
The views expressed in this article do not necessarily represent the views of the Department of Veterans Affairs or of the US government.
Funding This material is the result of work supported with resources and the use of facilities at the Department of Veterans Affairs National Center for Patient Safety at Ann Arbor, Michigan, and the Veterans Affairs Medical Centers, White River Junction, Vermont.
Competing interests None.
Ethics approval This study was conducted with the approval of the Research and Development Committee, VAMC White River Junction VT, USA, and the Committee for the Protection of Human Subjects, Dartmouth College, considered this project exempt.
Provenance and peer review Not commissioned; externally peer reviewed.