Voltaire observed that "Doubt is not a pleasant condition, but certainty is an absurd one." In their Open Forum, Brown and his colleagues (1) remind us that lore, tradition, and authority influence our thinking about suicidal patients. But that is not their main message. It is that suicide assessment must be continuous throughout the course of a patient's illness, even as the patient improves.

Should we trust a suicidal patient's sudden improvement, especially if the patient is judged to be at high risk of suicide? No evidence-based psychiatry can fully inform the clinician's judgment in this challenging clinical situation.

Psychiatric inpatients interact with clinicians along a dynamic continuum of truthfulness and deception, making it difficult for clinicians to discern real from feigned improvement at any given time and circumstance. Clear and unambiguous examples of real versus feigned improvement among high-risk suicidal patients are found only at the poles of the continuum.

A patient admitted to a hospital for treatment is expected to improve; that is the norm. But sudden clinical improvement among patients assessed at high risk of suicide poses a dilemma: is this a valid though rapid result of treatment, a decision to die that has brought relief, or a deception aimed at rapid discharge? Such a development challenges the psychiatrist's clinical acumen. The psychiatrist and the treatment team must determine not merely whether the suicidal patient's improvement is real or feigned but also whether any improvement is durable enough to sustain the patient's safety after discharge.

Complicating any risk assessment is the fact that the length of stay on most psychiatric inpatient units is brief, usually because of pressure from utilization reviewers. Some inpatients have stays of only five days or less. Inevitably, many are discharged to outpatient follow-up when they are still at moderate risk of suicide. Because the clinical focus is on returning psychiatric inpatients to the community as soon as possible, the high-risk suicidal patient who evinces sudden improvement fits in all too well with this treatment model and is easily discharged prematurely. Thus the extent of the persisting high risk can be easily overlooked. However, all improvements, not just sudden ones, require careful assessment to determine readiness for discharge.

Many psychiatric inpatients improve rapidly shortly after admission. The structured milieu, initiation of treatment, effects of medications, safety measures provided, and peer interactions can promote rapid improvement. Antianxiety, antipsychotic, and sleep medications can be effective within minutes or hours; a good night's sleep can result in rapid improvement in the patient's clinical condition. Apart from medications, psychosocial interventions—especially group therapy—decrease anxiety, reduce isolation, improve reality testing, provide needed hope and support, and shorten the hospital length of stay (2). Providing detoxifying services to patients who have been abusing substances can rapidly reduce high suicide risk, often dramatically; the effect is lasting when abstinence can be sustained. The genuinely improving patient, whose improvement is confirmed by multiple observers, usually gives the staff permission to speak to others who know the patient, adheres to treatment recommendations, conforms to unit policies, attends and participates in group meetings, socializes with other patients, and is visible on the unit.

Most high-risk suicidal patients show gradual improvement. Real improvement can be rapid but is not often sudden and unexpected. Basic indicators of improvement occur in sleep, appetite, symptom reduction, treatment adherence, and socialization. Real improvement is a process, even when rapid.

Feigned improvement is usually sudden and unexpected. It is not a gradual process but an event that can occur at any point during inpatient hospitalization. It is this group of suicidal patients that is not distinguished in Brown and colleagues' review (1).

Outwardly, the patient takes medications as prescribed, social isolation ceases, and the patient attends and participates in group therapy and other unit activities. However, the patient's desire to die and escape psychic pain remains unchanged. The clinician is obliged to return to basics in the assessment; core symptoms of psychiatric disorders usually remain unchanged. For example, the vegetative symptoms of depression usually persist. And lurking behind the sham improvement of psychotic patients are thought disorder, hallucinations, and delusions. The patient restricts access to collateral sources of information. Lethal implements may be secreted in the patient's room or elsewhere on the unit. Because many inpatients improve rapidly and are discharged, the high-risk suicidal patient's specious improvement can be very difficult—and, on occasion, impossible—to distinguish from real improvement.

The high-risk suicidal patient who feigns improvement by denying suicide ideation, intent, or plan often displays behaviors and attitudes that belie these negations (3). The true intent is to obtain release from the hospital as soon as possible or to wait out a short length of stay, while planning to complete suicide shortly after discharge. Some other reasons for sudden improvement include the desire to just get out of the hospital; to complete substance detoxification or, conversely, to resume substance abuse as soon as possible; and "flight into health," a quick improvement in symptoms to deny mental illness and to avoid treatment.

A near-lethal suicide attempt usually precedes inpatient admission; once admitted, the patient displays a spectrum of signs, from subtle to overt, indicating an unchanged condition with persisting suicidal intent, despite denials. Typical behavioral signs include a lack of change in disturbed eating and sleeping patterns; averted gaze; poor personal hygiene; disheveled appearance; and isolated, seclusive behavior, with the patient spending most of the time in his or her room. Only minimal or superficial contact is maintained with the staff and other patients. Also, the patient attends group therapy sporadically or participates minimally or not at all, and the patient refuses to permit the staff to speak with family members. Approximately 25% of patients at risk of suicide deny having suicidal ideation to the clinician but do admit it to their families (4).

In addition, the patient does not form a therapeutic alliance, an important protective factor against suicide, with the psychiatrist and other members of the treatment team (5,6,7). Adherence to the medication regimen and to unit rules is grudging or confrontational.

Patients depressed and hospitalized for the first time, who are high-functioning achievers and unable to work or unable to work productively, are often at high risk of suicide (3). The reality of having a mental illness is mortifying. It is viewed as a personal failing and experienced as a devastating narcissistic injury. Before committing suicide, many of these previously high-functioning patients gradually withdraw from important relationships, much like patients who are terminally ill and are preparing to die. These patients may press staff for an immediate discharge, with statements such as, "I don't belong here with these crazy people" or "I will lose my job and family." The patient will likely challenge involuntary commitment based on apparently intact functioning, no prior suicide attempts, and agreement to participate in outpatient care. Consultation may help discern real from feigned improvement. The clinician should "never worry alone" (personal communication, Gutheil TG, 2006).

When patients feign improvement, fundamental changes do not usually occur in their clinical condition, behavior, or attitude. For example, despite occasionally laughing while playing a game on the unit, such patients may be only momentarily distracted from their depression and suicidal intent. In the absence of other clinical indicators of improvement, this levity should not be construed as improvement.

The patient who is determined to commit suicide considers the psychiatrist and clinical staff the enemy, because they stand in the path of the patient's intent (8). Tragic consequences can ensue from assuming simplistically that every suicidal patient wants help and will cooperate with treatment. The clinician may trust what the patient says but must verify that information by performing systematic suicide risk assessments. The clinician must be able to trust the adequacy of his or her risk assessment, if not fully trust the patient.

When conducting suicide risk assessments, clinicians should include behavioral risk factors, because those associated with suicide often speak louder than the deceptive patient's words (9). The behavioral manifestations of most psychiatric disorders (important risk factors for suicide) are usually observable.

Gathering information from collateral sources is basic to competent suicide risk assessment of high-risk patients who suddenly improve. Family members are primary sources of information. When the patient refuses to authorize discussion with family members, the clinician can still solicit family data and just listen without revealing confidential information. Prior treatment records, especially hospital discharge summaries, usually cannot be obtained quickly, but calls or e-mails to clinicians who have previously treated the patient may provide critical clinical information. Prior treaters may be willing to reveal confidential information in an emergency. In an emergency situation, such as treating a high-risk suicidal patient, an exception to maintaining confidentiality exists both ethically and legally (10).

The treatment team that meets daily is an essential source of patient information that is obtained 24 hours a day, seven days a week. The multidisciplinary team has "a thousand eyes." The patient's current psychiatric record must be read carefully: it contains vital behavioral and other information regarding the patient's risk of suicide.

Clinicians must beware of iatrogenic sudden improvement (miraculous insurance cures), which is fraught with liability risk. Inflating suicide risk may ensure insurance coverage on admission, but a lapse in coverage can promote rapid and possibly premature discharge.

In the usual inpatient setting, the psychiatrist typically is at the top of a pyramid of staff; this organizational structure implies both a pyramid of decision making (final decisions) and of observational data (the psychiatrist as final common pathway for information). Regrettably, psychiatrists do not always review the entire chart or attend sufficiently to input from clinical staff. Even more regrettable, some psychiatrists assume that only what the patient tells the psychiatrist is true or important, discounting data from other team members. Both authors have been involved as experts in malpractice in cases where the psychiatrists cheerfully admitted in deposition that they did not read the nurses' notes or other entries.

Most clinicians are familiar with patients who put on a "smiley face" when speaking to the doctor but let their guard down with other team members and admit their underlying despair. This kind of selective candor also resembles how family members may be taken into the patient's confidence and may not hear the persisting depression or suicidal intent. This familiar situation requires the team leader to pay close attention to observations of front-line staff, who usually have seen the patient not for an hour, but for an entire shift.

Sudden, unexpected clinical improvement of patients who are assessed at high risk of suicide challenges the psychiatrist's clinical acumen. Behavioral indicators can help the clinician distinguish between real and feigned sudden improvement. Although genuine clinical improvement is a process, sudden patient improvement is a suspect event. Extensive data gathering from front-line staff is critical.

The authors acknowledge their indebtedness to members of the psychiatry and law programs at both institutions for helpful suggestions.

Dr. Simon had a one-time consulation with Novartis. Dr. Gutheil reports no competing interests.