Renegotiating the 'Contract for Safety' in the ER

May 1, 2002

There is a range of meaning that underlies patients' violent acts against themselves. The usefulness of so-called "safety contracts" to actually predict suicidal behavior is questioned.

"Will the patient contract for safety?" a clinician or insurance representative at the other end of the line wants to know. "I don't know," I answer. Often I will add, "And I don't care." I then explain why I believe that an emergency room (ER) "contract for safety" is clinically unsound and can, for some patients, lead to disaster.

Interpreting a 'Cry for Help'

"I want out," a despondent "Julia" told her primary care physician in his office. Along with hypertension, fluid retention, glaucoma and asthma, this 36-year-old woman had been treated for depression during the last year. Concerned about her safety after hearing this apparent cry from the depths, the doctor drove her to the ER. "I want out," Julia told the triage nurse during the initial medical evaluation.

When I entered her room, Julia was lying quietly on a gurney. Her mother and father sat in chairs, looking grave. I needed to talk to Julia alone and suggested to the parents that they go to the waiting room.

Julia was not married, did not have children and lived with her parents. She worked as a nurse-technician at a nearby hospital. Julia had been feeling well until a year before, when her 89-year-old grandfather died. After his death, Julia was overwhelmed with grief and became depressed. Her primary care physician started her on fluoxetine (Prozac), which was titrated to 40 mg. The drug worked at first by calming her down. Julia's doctor told her that she had a "disease." She did not see a psychotherapist.

One month before coming to the ER, Julia's 89-year-old grandmother died. With this second death of a close family member in less than a year, Julia's depression became more severe. She began to lose sense of who she was. Her antidepressant was increased to 60 mg without apparent effect. In some way Julia could not specify, she began to feel "unsafe."

Julia continued working but found it hard to concentrate. She felt under constant stress, had interrupted sleep and was tired during the day. She "ate everything in sight," gained weight, had little interest in doing many of the things she normally did and associated only with family members and co-workers.

Fifteen years earlier, Julia had been depressed and saw a psychotherapist for six months. She could not recall what was going on in her life at that time that might have predisposed her to depression. Julia did not drink alcohol or use illicit drugs. She was taking medication for asthma and hypertension; both were well controlled.

After bringing Julia to the ER, her doctor, alarmed by Julia's worsening condition, contacted a psychiatrist at the hospital and arranged an appointment for the following week. Hearing about Julia's veiled threat of self-harm, the psychiatrist had called the ER before I arrived to do the evaluation and left a message saying he felt that the patient should be admitted to a psychiatric unit.

Julia was immediately and fully responsive to the questions I asked. Though occasionally tearful, she seemed happy to be talking to someone. She was obviously frightened about her situation and was trying hard to rein in her emotions. Julia was not sure what was happening to her, and it soon became evident that the anxiety associated with this confusion was the acute component of her yearlong mood disturbance.

I was surprised when Julia let it be known, almost incidentally, that she felt guilty about her grandfather's death. She was the last person to see him before he died, and a cousin was miffed that she had not been contacted so she, too, could have been present. That Julia could be made to feel guilty for such an "infraction," particularly considering how much time she spent with each grandparent during their final illnesses, shows how psychologically vulnerable this woman was. The guilt her cousin put on her was clearly a factor in this case of complicated bereavement.

I reminded Julia that living to age 89 is a privilege given to few. As I said this, a part of her burden seemed to lift. Her affect brightened through the interview, and she was smiling near the end. She told me she was grateful that I was talking to her about her grandparents and how she felt about them. I explained that fluoxetine might calm her down (and/or give her a "boost"), but it could not resolve the issues created by her loss or neutralize the guilt she was feeling.

I still had to find out what Julia meant when she said, "I want out," the statement that got her to the ER and set the stage for our interview. At first, Julia said she did not know what these words meant. She acknowledged being confused and overwhelmed by her feelings. I asked if, during the last month, as her depression deepened, she had ever thought of doing anything specific that would threaten her life. No, she said, emphatically. She denied ever trying to harm herself in the past. So what did Julia mean by wanting "out?" What did she want "out" of? Not life itself, apparently.

Like so many patients I have evaluated in the ER who said "I want out" and "I don't want to be here anymore," I believe Julia wanted "out" of the intense emotional pain that depression and guilt were causing her. Julia was confused and hurting, but she was not hopeless and despairing. We came to a shared understanding that she was not likely to harm herself.

Before discharging her from the ER, I needed to know how Julia saw herself getting through the next few days. She was anticipating a restless night or two, with some floor-walking to offset her nervous energy. Julia wanted a week off from work, away from the stress of her job. "I need someone to talk to," she said with conviction. "You have been very helpful." I took her words to mean that she had come to implicitly recognize a psychodynamic reason for her pain and the need for psychotherapy.

Just to get her reaction, I asked if Julia felt a hospitalization would help her. No, she said, citing among other reasons her reluctance to be confined to small spaces, like a hospital room. I was convinced that she would not benefit from being admitted and most likely would have her recovery slowed by this kind of confinement. Her parents felt it was safe for her to go home. Her father agreed to take a few days off from work to be with her. Her doctor agreed that she could be monitored as an outpatient.

The Meaning of Self-Injury

I have helped hundreds of ER patients acknowledge the reason for their self-injurious gestures. Typically, someone who has just had a dustup with a parent, a significant other or a spouse will take a number of pills. These pills may be the patient's prescribed medication, someone else's prescribed medication, an over-the-counter medication or a mixture of these drugs. Some patients call 911 themselves. Others tell someone else about what they have done, and that person calls 911 or brings the person to the hospital. Once a patient is in the ER, the poison control center is contacted and given the specifics of the overdose. Experts there on pharmacology make recommendations for treatment based on protocols. After the patient has been medically cleared, someone from the psychiatry service is called in to do an evaluation.

In all but a handful of overdose cases, I have been able to work with the patient to understand what the overdose meant. Most patients reveal their intention quickly and gladly. "I wanted attention" is the most common reason given. Others admit they were trying to punish a family member or significant other who they felt had wronged them. Typically someone, who was not getting their way and was not willing to take an instrumental next step to resolve the underlying conflict, decided to dramatize the point. Countless mothers (and not a few fathers) have told me, weighing in on the situation, that their adolescent son or daughter had just "thrown a pity party." What some patients try to say with pills others express with superficial cuts, usually to the ventral wrist.

Patients love to talk about why they make these gestures. It is as if they have finally been listened to and understood. They are now getting the "attention" they have craved for so long. This success justifies to them the need for having spoken in the code of gestural self-harm. Many feel that the discomfort of the nasogastric tube and the charcoal lavage, or the sight of the bloody wrist was worth it (although most also say they will not try it again). They leave the ER believing they have accomplished something and, in a sense, they have. I try to get them to acknowledge that there are better ways to be heard and to get the kind of attention they should be seeking. When I see patients absorbing this point, I feel that my time with them was well spent.

Some patients come to the ER down on their luck because of their bad behavior. This often involves substance abuse and the betrayal of everyone they know because of it. They have no money and no place to stay. These patients often make a point of not contracting for safety. "I'm suicidal," they will tell the triage nurse. Some add, "I'm also homicidal," presumably hoping to cover all the bases in their bid for a hospital bed. There is that old ER saying: "The patient knows what to say," i.e., what to tell the ER staff to get admitted and receive the proverbial "three hots and a cot."

Clinicians who evaluate psychiatric patients in the ER should be able to determine what the patients who "know what to say" really mean. After interviewing between 200 and 300 ER patients, many of them substance abusers, I began to feel what I can only describe as a new degree of confidence in making a determination about a patient's "safety" (Muller, 2000). Surgeons call this blending of discipline and instinct unconscious competence (Lahr, 2001). Clinicians just honing their skills will wish, as the saying goes, to err on the side of caution. I have had patients who were broke and homeless, whom I knew to be no threat to themselves or to others, say "I'm so suicidal I'll never make it home." One patient told me, "I'm so suicidal I'll never make it out of the ER." We need to learn how not to be manipulated by those who would use mental health care facilities to reduce the consequences of their bad behavior (Muller, 1998).

At times, I do not know in my head and in my gut why a patient took an overdose or chose some other act of self-harm. Sometimes, I sense that a patient may not have exhausted the need to act out a conflict in a harmful way. I hospitalize these patients, regardless of what they "contract" for. Ultimately, I try to determine whether a patient can imagine a future, in spite of the difficulties, and has the will and the capacity to deal with what is ahead. True despair, the kind that drives the real suicidal act, is the lack of this will and capacity.

I did not ask Julia to contract for safety as a condition for discharge from the ER. I have come to see these contracts, made in the heat of crisis with a clinician whom the patient does not know, as intrinsically unreliable and essentially different from the agreement often made in outpatient therapy. Outpatient contracts, struck with a clinician when a therapeutic alliance is already established, have proved effective in containing the self-destructive impulses of suicidal patients.

No 'Safety Contract' Is Ironclad

Like those patients who have no other place to go and "know what to say" to be admitted, patients who have already made up their minds to do serious harm to themselves or to others also "know what to say" to be discharged from the ER. One patient's ER contract for safety, apparently accepted at face value, led to a tragedy. I know the story only second-hand, but the source is a reliable one.

A woman in her 30s was brought to the ER of another hospital by her brother. Two facts dominated the case: 1) at home, the depressed woman had threatened to kill herself and 2) the brother's gun was missing, and he was sure that the patient had taken it. A social worker was called in to do an evaluation and make a determination. The patient vehemently denied having any intention or plan to harm herself or that she had taken her brother's gun. She wanted to be discharged. A verbal contract for safety was made with the social worker.

The brother felt strongly that the patient's threat earlier that day to harm herself was real, and he demanded that she be certified. Believing the patient's claim in the ER that she was not suicidal, the social worker discharged her. Several days later, the patient drove to a neighboring state, checked into a motel and shot herself to death. The family sued the hospital.

I do not wish to second-guess the social worker's clinical evaluation, although I wonder if the brother's missing gun would have triggered my suspicion of the patient's suicidal intention and plan. I do not know all that went into the decision to discharge this woman from the ER, but I suspect that her contract for safety figured significantly.




Lahr J (2001), The alchemist. The New Yorker 77(31):88-96.


Muller RJ (1998), Malingerers and manipulators in the ER. Psychiatric Times 15(3):23-25.


Muller RJ (2000), The narrative in psychiatric diagnosis. Psychiatric Times 17(2):14-15.