View as “PDF version”

Patient Safety Tip of the Week

February 2, 2021

MGH Protocols Reduce Risk of Self-Harm in ED



Whether or not your hospital has a behavioral health unit, you will have patients come to your emergency department who have conditions which put them at risk for suicide or self-harm. You, therefore, need to plan for managing such patients safely. Massachusetts General Hospital recently published protocols and interventions it put in place to minimize the risk of self-harm (Donovan 2021).


They performed RCA’s (root cause analyses) of incidents in prior years and determined the areas for intervention: (1) safe bathrooms, (2) number and training of patient observers, (3) management of personal belongings, (4) clothing search or removal policies and training, and (5) additional protocols for high-risk patients.


Near the top of their list was one we have harped on for years. We sometimes, half in jest, bet hospital CEO’s that we will find at least 3 patient safety hazards or vulnerabilities within the first hour. When we do so, we head straight to the radiology suite or the ED, where we know we are likely to find bathrooms that patients could use for attempted suicide or self-harm. Bathrooms, of course, are common locations for in-hospital suicide attempts (see our August 29, 2017 Patient Safety Tip of the Week “Suicide in the Bathroom”). Well, the MGH protocols took care of at least the ED hazards in those bathrooms. Safety features they implemented included shatterproof fixtures and mirrors, paper wastebasket liners, minimal ligature risks, and mirrors and curtains allowing visibility while protecting privacy. The task force also included specific training for observers regarding ensuring adequate patient visualization in the bathroom with the least intrusion possible.


Our prior columns have focused on removing loopable fixtures or installing ligature-resistant fixtures in bathrooms. But the MGH took it a step further by considering that patients could use pieces of glass to harm themselves, hence the need to use shatterproof mirrors. Avoiding plastic waste basket lines also makes sense. They also focused on a point we’ve emphasized in our own prior columns – observation of patients in bathrooms has often been inadequate.


That universal observer training provided education about the critical role of constant observation to mitigate safety risks. It was a three-week orientation, with annual retraining, included training on key safety issues, including suicide risk, covert or surreptitious behavior, possession of dangerous items, ingestions, elopements, risk of harm to others, and responses for observation of concerning behavior.


They also created a checklist tool of safety concerns (for example, risk of harm to self, risk of harm to others, elopement risk), and goals for observation, including constant vigilance, using a safe bathroom, visualizing patient's hands and face continuously, and ensuring that no unsafe objects are in the environment. The observer responsible for that patient reviews the checklist with the patient's nurse, and both parties sign the tool.


Note that this is the sort of training we’ve recommended for “sitters” assigned to observe at-risk patients being temporarily housed on med/surg beds (see, for example, our several columns listed below on jumps from hospital windows).


The other key focus was on issues related to clothing and potential contraband that might be used for self-harm. They did decide to allow patients to have some personal belongings, such as books or cellphones, recognizing they might need some diversions if they were to spend long hours in the ED. However, such items would be kept in a secure locker and would be examined by staff (for potentially dangerous elements) before allowing patients to use these.


They encourage patients to change into safe clothing. But some patients will refuse such requests. They, therefore, require a risk assessment (performed by an ED physician) and only require “forced” disrobing for patients scoring at the highest risk levels. Note that they also considered “forced” disrobing to be the equivalent of using a “restraint”. Hence, such patients would also need to meet the criteria for restraint use and follow restraint protocols. (Note: we recommend you check the laws in your state regarding the disrobing and restraint issues.) For patients at lesser levels of risk for self-harm, interventions might include “reducing clothing to a single layer, turning pockets inside out, undergoing a pat-down, and removing high-risk items such as belts, drawstrings, and shoes.”


They also identified the “exceptionally high-risk patient”, such as one who had a history of numerous episodes of self-harm, often severe, in the ED, and created additional safety interventions for such patients. These interventions include 1:1 observation, additional or repeated searches of the patient or belongings (possibly including search of undergarments), and immediate psychiatry consultation to plan for patient safety. Such patients are flagged in the EHR (electronic health record) to alert providers about the exceptionally high risk and suggest these additional precautions.


The full MGH guidelines are also available in an appendix to the article.


In the 12 months prior to the protocol initiation, among 4,408 at-risk patients, there were 13 episodes of attempted self-harm (2.95 per 1,000 at-risk patients), and 6 that resulted in actual self-harm (1.36 per 1,000 at-risk patients). In the 12 months after the protocol was introduced, among the 4,523 at-risk patients, there were 6 episodes of attempted self-harm (1.33 per 1,000 at-risk patients) and only 1 that resulted in actual self-harm (0.22 per 1,000 at-risk patients). There were no deaths. Though these results did not meet criteria for statistical significance, they are nevertheless impressive.


The MGH team also audited compliance with the intervention and found only 42 breaches of the protocol out of 4,523 unique patient visits. 25 breaches were related to changing patients, 11 breaches related to patients having objects that could be used for self-harm, and 6 breaches related to observers (availability or performance). Significantly, though, in 3 of the 6 attempted self-harm events after protocol implementation, there were protocol breaches. Of those 3 breaches, 2 were related to observer performance (lack of observation in the bathroom). Many of the cases of suicide attempts we’ve previously discussed also occurred when staff failed to adequately observe patients in the bathroom. One suggestion we’ve made is that use of same-sex observers might minimize non-compliance with this aspect of observation.


The article goes on to discuss the costs of implementing such a safety program, including the costs of hiring and training observers, renovating bathrooms, providing storage areas for patient items, plus all the time and effort that went into planning and implementing.


This is really a useful study that almost every hospital can learn from. The legwork the MGH team has done should make it easier for you to begin such projects at your hospital. And, again, we recommend that you consider such a program not only for your ED, but also consider a nearly identical program for those med/surg floors that are, from time to time, required to care for patients at risk of suicide or self-harm. Lastly, don’t forget that these patients might also at some time require transport to Radiology or other area. You should make sure that your “Ticket to Ride” checklist and procedure includes all the important elements required to prevent self-harm during such transports.



Some of our prior columns on preventing hospital suicides:




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






Donovan AL, Aaronson EL, Black L, et al. Keeping Patients at Risk for Self-Harm Safe in the Emergency Department: A Protocolized Approach. Joint Commission Journal on Quality and Patient Safety 2021; 47(1): 23-30






Print “PDF version”
















Tip of the Week Archive


What’s New in the Patient Safety World Archive