View as “PDF version”

Patient Safety Tip of the Week

September 22, 2020

VA RCA’s: Suicide Risks Vary by Site



Much of the good work analyzing suicides and attempted suicides in hospitals has come from the VA hospital system. Several of our recent columns highlighted lessons learned in that system and almost all of our previous columns on suicide, listed below, have reference to something from the VA system.


Mills et al. (Mills 2020) recently published findings from 847 root cause analyses (RCA’s) of suicides and suicide attempts in the VA system. These included events on mental health units, inpatient medical units, emergency departments, outpatient clinics, community living centers, or on the hospital grounds.


The most striking finding is that the mode used for attempting suicide varied considerably by location, and that has implications for what you should be doing to prevent suicides.


Hanging accounted for 38.1% of the attempts and 70.9% of the deaths on mental health units, but also accounted for 25.9% of the attempts and 50.0% of the deaths on medical units. Hanging accounted for lesser percentages of attempts in emergency departments (19.6%). and community living centers (22.9%), though these did often result in death.


Cutting with a sharp object accounted for relatively equal percentages of attempts on mental health units (24.6), medical units (22.4%), and emergency departments (22.4%), but more so in community living centers (39.6%). However, cutting resulted in no deaths.


Overdose accounted for 54.5% of the deaths in residential units and all the deaths in emergency departments. Strangulation accounted for roughly equal percentages of attempts on mental health units (20.9%), medical units (15.3%), and emergency departments (29.0%), but accounted for very few deaths. Gunshot accounted for 59.1% of the deaths on hospital grounds (outside of inpatient or clinic area) and all the deaths in outpatient clinics. Jumping from high places was more common on hospital grounds.


And, as you’d expect, the root causes contributing to suicides and suicide attempts also varied by location. The paper has a nice table summarizing root causes by site, with some specific examples.


Mills et al., therefore, suggest strategies to prevent suicides should be tailored to the location. The authors note that the methods used in attempted suicides may be largely a function of what opportunities are available to the patients. For example, on mental health units and medical units, where environments are more controlled and observed, hanging and cutting are the most common means available.


On mental health units you need to eliminate all anchor points that can be used for hanging (and use above-door alarms as discussed in our August 11, 2020 Patient Safety Tip of the Week “Above-Door Alarms to Prevent Suicides”). All other lethal means, such as drugs, plastic bags (that can be used for asphyxiation) and sharps should be removed off the unit. Ligature-resistant hardware should be used in all places. You should be using the VA Mental Health Environment of Care Checklist (MHEOCC) that we’ve discussed in many of our columns. That checklist is available online on the VA Patient Safety website. Lights and electrical outlets should be modified so they cannot be used for self-harm. Doing regular environmental safety rounds is a good idea. And, given the root causes found in their RCA’s, attention to assessment and treatment are critical.


In the emergency department, overdoses were the only method of suicide resulting in death. These usually involved drugs or medication brought into the emergency department by the patient and used in the bathroom when not under supervision. So, in the emergency department, careful assessment of patients for suicide risk coupled with one-to-one observation of at-risk patients is most important. They stress that includes observation in the bathroom. And, though they note lack of an evidence base for efficacy, contraband searches make a lot of sense.


The authors noted that jumping was seen more on hospital grounds than on inpatient units and that most jumping events did not result in death. We’ve done several columns on jumping from windows and have noted some common patterns and themes. Many such incidents occur when patients are housed on non-psychiatric units or general medical floors. The typical patient is a young or middle-aged male, but occasionally elderly patients or females have also jumped through or out of windows. The patient is often admitted for an attempted suicide but, again, not always. Typically, he/she is confused or hallucinating. It’s not just patients with known psychiatric disorders or a history of suicide attempt that are at risk. Patients with brain injuries or delirium are at risk, particularly those who have demonstrated a tendency to wander or have verbalized their intent to “get out of here” or “go home”. And the incidents have commonly occurred while patients are already on 1:1 continuous observation and the observer is actually in the room. However, the “sitter” has often been inadequately trained in dealing with such patients. Hospital beds are often used as “launch pads”. Objects in the room, such as a chair or piece of medical equipment, are often used to break the glass in the window.


Bathrooms are a frequent location for suicide attempts (see our August 29, 2017 Patient Safety Tip of the Week “Suicide in the Bathroom”). While you can make structural changes to all bathrooms on your mental health units, it is not feasible to do so in all other bathrooms. Nevertheless, we recommend you look at modifications in “high-risk” bathrooms (for example, those in the ED or radiology suite or in medical unit rooms in which you most often house suicidal patients). In several columns we’ve also noted that the suicide occurred while the observer was waiting outside the bathroom or shower. We pondered whether the “modesty” factor or “gender” factor may have played a role and suggest that use of same-sex observers might make sense when suicidal patients need to be observed in those locations.


The authors acknowledge that on medical units it is not possible to eliminate all dangerous items because many are need for provision of medical care. They note the importance of one-to-one observation and that the observer needs to be both aware of the suicide risk and properly trained to deal with such patients. They again stress the importance of continuing observation in the bathroom.


In several of our own columns we have recommended hospitals which often have a need to house such patients on medical units make one or two rooms safer in that regard. For example, such “dedicated” rooms might have the type of windows installed on behavioral health units that are not breakable or subject to manipulation. They might also have doors with anti-ligature features and above-the-door alarms. On a general medical unit, a suicidal patient also has more easy access not only to medications but also to things like cleaning agents that might be in unlocked locations. And we echo Mills’ point about the importance of having observers that have been appropriately trained in handling such patients.


And, of course, don’t forget our frequent warnings about the dangers during intrahospital transports. We’ve noted cases where patients are taken to the Radiology suite for an imaging study and go into a bathroom there and hang themselves. We recommend one of the items on your “Ticket to Ride” intrahospital transport checklist should be related to suicide risk.


We’ll also add that you always need to be aware of dangers just outside your units as well. There have been cases where patients have used fire alarms to sneak off locked units, then jump off rooftops in adjacent areas. Therefore, it is important that you check all your stairwells and make sure doors to areas like rooftops are locked.


Ironically, the GAO (Government Accounting Office) recently issued a report criticizing the VA for its experience with on-campus suicides (GAO 2020). The report acknowledges that the VA has made suicide prevention a priority and noted the several interventions it has implemented. However, it found that identification of all suicides was incomplete and inaccurate. Specifically, the GAO identified four cases of undercounting (deaths that should have been reported as an on-campus veteran suicide but were not) and 10 cases of overcounting (deaths that were reported as on-campus veteran suicides but should not have been). It also found that no RCA’s were done on several on-campus suicides which the VA responded did not meet the written criteria for RCA’s because the individuals had not received recent psychiatric care within the VA system. We don’t think the GAO report should detract from the good programs the VA has implemented and the useful lessons learned it has disseminated outside the VA system. The report does have some nice photos of the above-door alarms that we discussed both today and in our August 11, 2020 Patient Safety Tip of the Week “Above-Door Alarms to Prevent Suicides” and photos of some of the barriers they have installed to prevent jumping from high places.



Some of our prior columns on preventing hospital suicides:




Some of our past columns on issues related to behavioral health:







Mills PD, Soncrant C, Gunnar W. Retrospective analysis of reported suicide deaths and attempts on veterans health administration campuses and inpatient units. BMJ Quality & Safety 2020; Published Online First: 20 August 2020



Mental Health Environment of Care Checklist (VA)




GAO (Government Accounting Office). Veteran Suicide: VA Needs Accurate Data and Comprehensive Analyses to Better Understand On-Campus Suicides. GAO-20-664: GAO (Government Accounting Office) 2020; September 9, 2020






Print “PDF version”














Tip of the Week Archive


What’s New in the Patient Safety World Archive