Dr. Simon is clinical professor of psychiatry and director of the Program in Psychiatry and Law at Georgetown University School of Medicine in Washington, DC.

Acknowledgments: The author reports no financial, academic, or other support of this work.


 

Abstract

The standard of care requires that psychiatrists adequately assess patients at suicidal risk. An exception to the standard of care for performing adequate suicide risk assessments does not exist for the clinician practicing in managed care settings. However, it is often more difficult for the clinician to perform systematic suicide risk assessments due to a high volume of patients and limited time and treatment resources. Data-based adequacy of the clinician’s evaluation of suicide risk in the managed care setting depends on a number of factors. These include completing a formal suicide risk assessment, weighing suicide risk factors, obtaining information from collateral sources, assessing the working alliance between the patient and the clinician, and close collaboration with the treatment team in inpatient settings.

 

Introduction

In managed care settings, only the sickest patients are admitted to inpatient psychiatric units. These patients usually have a primary diagnosis of schizophrenia, major affective disorder, or a severe personality disorder with prior hospital admissions. Many of these patients have comorbid substance-related disorders. Most of the patients are at heightened risk of suicide. In outpatient settings, patients at suicide risk may be seen by clinicians in split treatment situations or for a few brief, authorized visits. In the managed care era, all but those patients at highest risk of suicide are treated as outpatients.

Predicting which patients will commit suicide is an impossible task.1 No professional standard of care exists for the prediction of who will commit suicide. Nevertheless, the clinician has a professional and legal duty to perform an adequate assessment of the patient at risk for suicide.2 A variety of suicide risk assessment methods are available.3-7 No model of suicide risk assessment has been empirically tested for reliability and validity.8 Clinicians are free to construct and use their own comprehensive risk assessment methods based on their training, clinical experience, and the pertinent psychiatric literature.

Clinicians’ suicide risk assessments range on a continuum from merely asking patients if they are suicidal to performing formal, systematic suicide risk assessments. Especially in managed care settings, the former is much more common because of the large volume of patients treated. Nonetheless, clinicians continue to have a professional, ethical, and legal duty to provide adequate care to their patients regardless of managed care restrictions and protocols.9 In managed care settings, clinicians need to gather information about patients quickly and accurately. By accessing multiple informational sources, the clinician can be confident that suicide risk assessment rests upon a firm factual foundation.

 

The Adequacy of Suicide Risk Assessment

In managed care settings, clinicians often practice under conditions of too little time and too little information. The liability risk of the clinician is high if insufficient information is gathered and the patient subsequently attempts or commits suicide. Errors of fact, rather than errors of judgment, greatly increase a clinician’s liability exposure. The clinician cannot make sound judgments in the absence of essential facts. If the clinician gathers sufficient information upon which to base an adequate suicide risk assessment but makes an error in judgment that harms the patient, it is less likely that a lawsuit will succeed against the clinician. Errors alone do not constitute malpractice. The clinician is not held legally responsible for mistakes provided due care was used in making clinical decisions.10

Since the standard of care is the assessment of suicide risk rather than prediction, the clinician’s confidence in the adequacy of the suicide risk assessment is paramount. Systematic suicide risk assessment is ultimately a data-informed clinical judgment. There is a direct correlation between the database upon which the risk assessment is conducted and the level of adequacy to which it can be ascribed. A convergence of confirmatory data from a number of sources acts as an adequacy indicator. The adequacy level of the clinician’s suicide risk assessment is dependent upon whether sufficient, reliable information has been gathered to support a systematic analysis of suicide risk.

Suicide risk assessment has been analogized to weather forecasting.11,12 Weather forecasters make probability predictions based on available data and computer models. For example, a weather forecast may predict “a chance of rain tomorrow.” Forecasters try to achieve additional accuracy by expressing the chance of rain as a percentage, such as 30% or 70%. Forecasters may add a numerical confidence factor to their predictions expressed on a scale of 1–10, 10 being the most confident. The confidence in any weather forecast is diminished by time. For example, the forecaster’s confidence is much greater for 24-hour predictions than for 3–5 day forecasts.

Though lacking probability and computer models, the clinician’s confidence in suicide risk assessment similarly depends upon the timeliness, sufficiency, and reliability of the data obtained and evaluated. If the clinician only asks a patient if he or she has suicidal ideation, intent, or a plan, the confidence in such an assessment must be necessarily low. When the clinician performs a systematic suicide risk assessment that accesses multiple data sources, confidence should be greater that the assessment identifies the patient’s level of risk. Like weather forecasts, frequent updates keep the suicide risk assessment current and maintain confidence in the assessment.

 

Suicide Risk Assessment—A Medical Necessity

Formal Suicide Risk Assessment

The purpose of suicide risk assessment is to identify modifiable suicide risk factors and to inform the clinician about the patient’s safety and treatment needs. Especially in managed care settings, the rapid identification of a patient’s treatable suicide risk factors and safety requirements is of critical importance. Suicide risk assessment is first an essential clinical tool that informs treatment, and only secondarily a risk management technique.

As noted previously, suicide risk assessments range from the clinically naive to the clinically sophisticated. On one end of the spectrum, just getting a perfunctory “yes” or “no” answer to the question, “Are you suicidal?” is an inadequate assessment. A layman can just as easily ask that question. On the other end, a systematic suicide risk assessment provides a higher degree of confidence that important sources of clinical information are not overlooked in the overall evaluation of the patient. The formal suicide risk assessment can be thought of as a data-gathering engine. Both risk and protective factors are considered in the overall determination of suicide risk. An adequate suicide risk assessment reliably informs treatment and management.

In a formal suicide risk assessment, clinical, interpersonal, situational, statistical, and individual risk and protective factors are assessed.8 Careful weighing of risk and protective factors as low, moderate, or high according to the patient’s clinical presentation provides additional confidence in the adequacy of the suicide risk assessment. Evaluating suicide risk factors without also considering protective factors diminishes clinical confidence, since that is only half of the suicide risk assessment. Further, differentiating overall suicide risk as low, moderate, or high provides for more clinical precision in determining the patient’s safety needs (eg, suicide precautions: 5-, 10-, or 15-minute check-ups or one-on-one supervision). In addition, a formal suicide risk assessment permits the clinician and utilization reviewer to provide managed care organizations (MCOs) with credible data needed for the decision to admit a patient and to determine the length of stay and appropriateness of discharge.

Risk factors can also be evaluated using a variety of data parameters that enhance suicide risk assessment.13 As noted, the patient at suicide risk can be assessed on a facilitating-protective axis. A patient who is assessed at moderate risk when considering only suicide risk factors may have an overall low suicide risk assessment when protective factors such as positive interpersonal relationships and work satisfaction are present.

Other parameters that enhance suicide risk assessment include necessary and sufficient (individual and situational) and acute and chronic risk factors. Suicides generally occur when both necessary (depression) and sufficient (perturbation) factors are present.14,15 For example, a patient with major depression who also experiences an important personal or occupational loss has both necessary and sufficient suicide risk factors. Short-term and long-term risk factors are frequently observed in patients with comorbid Axis I and Axis II disorders. Discerning acute from chronic suicide risk factors focuses clinical attention on treating and modifying acute suicide risk factors, especially with limited outpatient visits or brief hospital stays.

A major resistance to performing formal suicide risk assessment is the perception that it is an onerous and time-consuming task—it is neither. A formal suicide risk assessment is an integral part of the mental status examination performed at admission or initial intake in outpatient settings. Documentation of the suicide risk assessment takes just a few minutes. The initial suicide risk assessment needs to be systematic. Subsequent suicide risk assessments can focus primarily on the patient’s acute risk factors that have required hospitalization or outpatient treatment. Current protective factors are also assessed. Chronic suicide risk factors such as prior attempts, family history of suicide, and demographic factors are identified during the initial suicide risk assessment. Given brief lengths of stay and limited outpatient visits, the ongoing systematic assessment of acute suicide risk factors and the patient’s response to treatment and management become the center of clinical attention.

In managed care settings, much reliance is placed on no-harm or “suicide prevention” contracts because of the volume and rapid turnover of patients. A no-harm contract is inherently unreliable with a new and unknown patient. Even with known patients, no-harm contracts should not be allowed to create a false sense of confidence. There is no evidence that no-harm contracts diminish the risk of suicide; they merely create an illusion of safety. No-harm contracts often diminish the perception of suicide risk without changing the reality of risk. No-harm contracts are actually most useful when patients reject them; the clinician is not deceived by disingenuous assent. No-harm contracts that take the place of performing a systematic suicide risk assessment prematurely cut off the essential task of data gathering.16

The usefulness of the suicide risk assessment is further strengthened when it is conducted as a process rather than as an event. Suicide risk assessment is a “here and now” assessment that time rapidly attenuates.8 Suicide risk assessment as a process continues to incorporate changing clinical data about the patient as well as additional information obtained from the patient’s family or other sources. Suicide risk assessments should always be documented.

MCOs, of necessity, require that suicide risk assessment be a process. In inpatient settings, the hospital utilization reviewer is queried by MCOs regularly about a patient’s level of suicide risk. Hospital length-of-stay determinations are based on the clinician’s assessment of the patient’s suicide risk. Some MCOs contact the utilization reviewer daily. The absence of continuing suicide risk assessments leaves the utilization reviewer unable to respond to the MCO’s questions about a patient’s level of suicide risk. Premature termination of the patient’s insurance benefits may result.

 

Suicide Risk Factors

Suicide risk factors can be divided into three groups: long-term, short- term, and individual. Long-term suicide risk factors are derived from community-based psychological autopsies and the retrospective study of completed suicide by psychiatric patients.17 Long-term suicide risk factors for patients with major affective disorder are significantly associated with completed suicides 2–10 years following assessment.18

Short-term suicide risk factors are derived from prospective studies of completed suicides in patients with major affective disorders.18 Short-term suicide risk factors are statistically significant within 1 year of assessment. They are predominantly anxiety based, usually severe, and often treatable.19 Short-term risk factors include panic attacks, psychic anxiety, loss of pleasure and interest, alcohol abuse, depressive turmoil, diminished concentration, and global insomnia.18 Identification of modifiable risk factors that can quickly lower suicide risk provides important data for patient care in all managed care settings. Many patients admitted to psychiatric units experience acute, severe anxiety that drives suicidal impulses. Recognition and treatment of anxiety symptoms often diminishes suicide risk and promotes rapid stabilization of the patient.

Patient availability and accessibility to guns significantly increases near-term suicide risk. The clinician must determine if guns are accessible to the patient. The patient may deny that availability and access to guns exists. Family members or others who live with the patient should be asked if guns are present in the home. Locking up or hiding guns is insufficient. Patients who are intent on killing themselves manage to find ways of gaining access to guns that are hidden or locked away. Guns must be removed from the home and stored in a safe place. The recent purchase of a handgun, especially by a female patient, should alert the clinician that the patient is at very high risk for suicide.

General suicide risk factors derived from retrospective studies have a high base rate of occurrence in depressed patients who do not commit suicide (high sensitivity, low specificity).20 Suicide risk factors unique to the individual are often the most reliable in suicide risk assessment. Clinicians who have treated patients over an extended period of time usually are able to recognize the recurrence of suicidal patterns. Patients vary greatly in the way that they manifest their suicidal patterns. Some patients display unique, even idiosyncratic symptoms not described in the professional literature, such as sleepwalking, stuttering, or compulsive whistling that presage suicidal behavior. Other patients may experience more conventional patterns, such as suicidal ideation, within a few hours or days following the onset of early morning awakening.

In the managed care era, due to split-treatments, brief outpatient and partial hospitalization visits, short hospital lengths of stay, and rapid patient turnovers, the emergence of unique, individual suicide risk factors are not usually recognized by clinicians. Collateral sources of information from family or friends may provide this vital data.

 

Collateral Sources of Information

Given the short length of hospital stays for patients (often 3–5 days), the psychiatrist may have insufficient patient information to conduct an initial, systematic suicide risk assessment. Information from other sources must be obtained quickly and reliably. Prior hospitalization records should be obtained and reviewed. A faxed discharge summary should be obtained or a phone call made to prior treaters after proper authorization by the patient or substitute decision-maker. Sending an authorization by mail and waiting for the records to arrive is futile. The patient will have been discharged long before the records arrive.

Prior records are usually very informative. For example, the psychiatrist may learn that during prior hospitalizations the patient secreted potentially lethal instruments or cheeked medication. The psychiatrist may discover that the patient accepted voluntary hospitalization after being confronted with the prospect of involuntary hospitalization, only to sign out against medical advice shortly thereafter. This kind of information is especially important when the patient is being treated by the clinician for the first time.

Family members are often a very important source of information for suicide risk assessment. Fawcett and colleagues18 discovered that approximately 25% of patients at suicide risk do not admit to being suicidal. However, in most cases the patient had communicated suicidal ideation or intent to family members. Some patients will admit suicidal thoughts or intentions in the emergency room and be certified by an MCO for admission but deny suicidal ideation once in the psychiatric unit. Patients who deny suicide risk usually do not meet managed care criteria for hospital admission. Regrettably, clinicians may exaggerate suicide risk to ensure admission.

The confidence that the clinician places on information from family members is contingent. Information from family members or friends may be unreliable due to denial, shame, or ambivalence toward the patient, or due to mental illness in family informants. Social workers are valuable members of the treatment team who can provide in-depth family evaluations and access collateral sources of important information about the patient.

A patient’s family or friends may provide vital information that the patient will not or cannot give. For example, the patient may deny or minimize their abuse of alcohol or street drugs. However, the family may reveal that the patient has an extensive history of abuse related to depression and suicide attempts. Obtaining information about the patient’s truthfulness and impulsivity provides additional important data. However, family or friends may not be available immediately or at all. When suicide risk assessment is a process rather than an event, new information that becomes available both clinically and from other sources can be incorporated into ongoing suicide risk assessments. Although the patient may not consent, the clinician may need to talk with the patient’s family members if a suicide crisis develops.2

 

The Therapeutic Alliance

The therapeutic alliance between the clinician and the patient affirms the willingness of each party to work together for the benefit of the patient’s treatment.21 If a solid working alliance exists, the patient usually cooperates with the clinician in disclosing suicide risk. However, the therapeutic alliance may vary considerably between outpatient visits or during inpatient treatment sessions because of instability and volatility of the patient’s condition. There are patients who have maintained a solid therapeutic alliance only to kill themselves between outpatient visits or inpatient sessions. A sustaining therapeutic alliance was present during the last session but was overcome by a downturn in the patient’s clinical condition.

In managed care settings, the time spent by the clinician with the patient is often insufficient to develop and sustain a therapeutic alliance. The frustration and distraction felt by both the therapist and the patient over managed care limitations on treatment and administrative “red tape” can disrupt the working alliance. The clinician’s confidence that the patient has provided reliable information about suicide risk is uncertain. On the inpatient unit, the psychiatrist can obtain information about patients at suicide risk by working closely with the treatment team. The treatment team has “a thousand eyes.” Members of the team are continually observing and working with the patient. The psychiatrist’s visits with the patient are often relatively brief. Daily careful review of the patient’s chart invariably provides a treasure trove of clinical information recorded by the treatment team.

 

Miraculous Cures and Other Fictions

Certain anomalous clinical practices arise when clinicians attempt to “game” managed care policies and procedures. For example, additional hospital days are authorized by MCOs based on the clinician’s assessment of a continuing high level of patient suicide risk. Nonetheless, if the MCO determines that the patient can be treated in a partial hospitalization program or as an outpatient, authorization for additional hospital stay is denied. Faced with denial of patient insurance coverage, the patient may be declared “miraculously cured” by the clinician and discharged.

In the managed care era, a heightened risk of patient suicide arises from premature discharge.8 Uncertainty about the number of additional outpatient treatment or hospital days that the MCO may approve often destabilizes patients by disrupting a therapeutic termination and discharge process. Suicide risk assessment may help prevent a premature discharge that is likely doomed to fail. Formal suicide risk assessment provides a solid basis for appealing denial of benefits when patients remain at high suicide risk. If the appeal fails, the patient may still require additional hospital treatment. Although the quality of care may be negatively influenced by the MCO’s restriction of services, clinicians continue to have a professional and legal duty to provide competent care to their patients while still in crisis.13

Formal suicide risk assessment gives the clinician confidence that a realistic evaluation of a patient’s suicide risk has been performed. “Gaming” the MCO becomes unnecessary. Exaggerating the severity of the patient’s condition to qualify for MCO approval exposes the clinician to increased liability risk when the “miraculously cured” patient is prematurely discharged and then attempts or commits suicide.

A related conundrum arises if a psychiatrist issues a therapeutic pass to assess the patient’s ability to function outside the hospital.22 Writing an order for any kind of pass raises a red flag for MCOs. MCOs have greatly restricted the use of this important assessment and treatment intervention. Once a therapeutic pass is issued, the MCO may deny further hospital benefits based on the presumption that the patient is ready for discharge. Careful suicide risk assessment before issuing a pass and upon the patient’s return from the pass is essential in preventing a premature discharge. The patient may return from a pass in worse clinical condition than when he or she left the hospital.

By the time the patient appears ready for a pass, the clinician usually knows more about him or her. The clinician has had time to collect sufficient information from a variety of sources upon which to conduct an adequate suicide risk assessment. Unfortunately, psychiatrists are very reluctant to issue patient passes that may lead to the MCO denying further benefits. If the pass does not go well, a formal suicide risk assessment can support both the clinician and utilization reviewer’s attempts to obtain authorization for additional hospital days.

Finally, upon inpatient admission patients at suicide risk usually are placed on safety precautions. Most of these patients want to smoke. In some psychiatric units patients must leave the unit to smoke, which requires that they be taken off suicide precautions in order to receive a staff-accompanied off-unit pass. They may refuse nicotine gum, patches, or inhalers. In order to smoke, patients frequently deny suicidal ideation even though they remain at heightened risk. Patients can mount considerable pressure on the clinical staff to be allowed to smoke, even threatening to leave if their requests are denied. If not in the unit, the psychiatrist is contacted, informed of the patient’s demand to smoke, and told that the patient denies suicidal thoughts. Based on that phone call, the patient may be allowed to leave the unit to smoke. A suicide risk assessment should be performed by the clinician before such changes are made in a patient’s safety requirements. When safety precautions are prematurely discontinued, the MCO may determine that the patient is not at suicide risk and deny further benefits.

Patients who manipulatively declare that they are “suicidal” can test a clinician’s equanimity. This situation frequently arises in emergency rooms with individuals who want to gain admission to the hospital but are not felt to be at substantial suicide risk. Similarly, at the time of discharge, patients may claim that they are suicidal in order to avoid discharge. Some manipulative individuals have learned that the magic words, “I am suicidal” can lead to hospital admission or a forestalling of discharge. Clinicians frequently feel helpless, manipulated, and frustrated by this declaration, but nevertheless may passively acquiesce. Confronted with this situation, some clinicians feel that they cannot confidently assess a patient’s potential for self-harm, and the patient is admitted. Also, the ever-present fear of being sued if the patient is not admitted and commits suicide induces feelings of helplessness that can make the clinician yield to the patient’s manipulation.

Clinicians need not be held hostage by questionable threats of suicide. A data-based formal suicide risk assessment is the best antidote. Suicide risk assessment can provide the clinical confidence that replaces paralyzing doubt. Clinicians will not be held legally responsible for adverse outcomes such as patient suicide if reasonable suicide risk assessments are conducted and documented. However, psychiatrists will be held liable for malpractice if they breach the duty to exercise that degree of skill and care ordinarily employed in similar circumstances by other psychiatrists.23 Failure to adequately assess patients at risk of suicide is a deviation from the standard of care.

 

Avoiding Inadequate Assessment

As with weather forecasting, the clinician can determine the confidence he or she should place on the adequacy of the suicide risk assessment based on the data-gathering process that forms the foundation for the patient evaluation. Determining the clinician’s level of confidence in the patient data is essential for the treatment and management of those at suicide risk. The accompanying Table contains a suicide risk assessment data checklist that can be used by clinicians. The standard of care requires that the clinician gather sufficient information upon which to base an adequate suicide risk assessment. The checklist can alert the clinician to deficiencies in the data collection.

 

A formal assessment itself is the impetus to gather information about a patient’s level of suicide risk. The therapeutic alliance, though extremely important as a conduit of information about the patient’s mental state, can be absent, limited, or variable. The formal checklist reminds the clinician to consider multiple data sources that create the foundational basis for the overall evaluation of the patient’s suicide risk and the continuing suicide risk assessment process.

 

Clinical Vignette

A 42-year-old single man is admitted to a psychiatric unit from the hospital emergency room after threatening to jump off a building. The patient is extremely depressed and suspicious. He seems to be responding to auditory hallucinations. The patient provides minimal history. He is a first-time admission who is unknown to the psychiatrist and the hospital staff. Next of kin live in another state. The patient says he has not seen his parents and sister for over 5 years. The psychiatrist makes a preliminary diagnosis of major depression with psychotic features. The patient denies suicidal ideation, intent, plan, or prior attempts. A formal suicide risk assessment is not performed. He is placed on 15-minute suicide precautions. The treatment team notes that the patient is seclusive. He is found positioning himself by the exit door. Staff considers the patient to be an elopement risk. Agitation manifested by constant pacing is observed. The patient is prescribed a benzodiazepine with marked decrease in pacing.

The vignette is illustrative of a typical first-time admission from the emergency room of a general hospital to a psychiatric inpatient unit. Although a rudimentary suicide risk assessment is performed, considerable data are available for a more thorough suicide risk assessment, such as diagnosis and behavioral information including agitation, seclusiveness, auditory hallucinations, and an absent therapeutic alliance with staff.

A quick glance at the suicide risk assessment data checklist can remind the staff that much more information is needed. In the vignette, little confidence should be placed on the initial suicide risk assessment. As a result, suicide risk precautions must remain high. The first 24 hours following admission are critical for evaluation of suicide risk and safety management. As hospitalization continues, the treatment team learns more about the patient. The social worker usually contacts collateral information sources. More clinical data about the patient become available for a formal suicide risk assessment, especially when considering discharge.

In the vignette, the initial denial of suicidal ideation, intent, or plan by a seriously depressed, suspicious patient provides an inadequate and misleading suicide risk assessment. In the managed care era, the suicide risk assessment data checklist may help clinicians counteract this data deficiency.

 

Conclusion

Professional standards do not exist for the prediction of suicide. The standard of care does require that psychiatrists adequately assess patients at risk. An exception to the standard of care does not exist for the clinician practicing in a managed care setting. However, it is often more difficult for the clinician to perform systematic suicide risk assessments due to limited time and treatment resources. Also, with shortened hospital lengths of stay, the responsibility for treating an acutely ill patient shifts to partial hospitalization programs and other outpatient settings, where adequate suicide risk assessment continues to be just as important.

The adequate assessment of the patient at suicide risk requires the presence of reliable, timely information. A suicide risk assessment is only as useful as the information it is based upon. Data-based adequacy of the clinician’s evaluation in the managed care setting depends on a number of factors. These include weighing suicide risk factors, obtaining information from collateral sources, assessing the working alliance between the patient and the clinician, and close collaboration with the treatment team in inpatient settings. Data-based suicide risk assessment increases the clinician’s confidence in both treatment interventions and the safety management of patients at risk.   PP

 

References

1.    Pokorny AD. Prediction of suicide in psychiatric patients: report of a prospective study. Arch Gen Psychiatry. 1983;40:249-257.
2.    Simon RI. The suicidal patient: accepting patient risk and therapist anxiety. In: Lifson LE, Simo RI, eds. The Mental Health Practitioner and the Law: A Comprehensive Handbook. Cambridge, MA: Harvard University Press; 1998.
3.    Simon RI. Concise Guide to Psychiatry and Law for Clinicians. 3rd ed. Washington, DC: American Psychiatric Press; 2001.
4.    Jacobs DG, Brewer M, Klein-Benheim M. Suicide assessment: an overview and recommended protocol. In: Jacobs DG, ed. Guide to Suicide Assessment and Intervention. San Francisco, CA: Jossey-Bass; 1999.
5.    Maris RW, Berman AL, Maltsberger JT, et al. Assessment and Prediction of Suicide. New York, NY: Guilford Press; 1992.
6.    Chiles JH, Strohsall K. The Suicidal Patient: Principles of Assessment, Treatment and Case Management. Washington, DC: American Psychiatric Association; 1995.
7.    Clark DC, Fawcett J. An empirically-based model of suicide risk assessment of patients with affective disorders. In: Jacobs DG, ed. Suicide and Clinical Practice. Washington, DC: American Psychiatric Association; 1993.
8.    Simon RI. Psychiatrists awake! Suicide risk assessments are all about a good night’s sleep. Psychiatr Ann. 1998;28:479-485.
9.    Simon RI. Psychiatrists’ duties in discharging sicker and potentially violent inpatients in the managed care era. Psychiatr Serv. 1998;49:62-67.
10.    Macbeth JE, Wheeler AM, Sither JW, et al. Legal and Risk Management Issues in the Practice of Psychiatry. Washington, DC: Psychiatrist’s Publishing Group; 1994.
11.    Simon RI. Clinical Psychiatry and the Law. 2nd ed. Washington, DC: American Psychiatric Press; 1992.
12.    Monahan J, Steadman HJ. Violent storms and violent people: how meteorology can inform risk communication in mental health law. Am J Psychol. 1996;51:931-938.
13.    Simon RI: Taking the “sue” out of suicide: a forensic psychiatrist’s perspective. Psychiatr Ann. 2000;30:399-407.
14.    Schneidman ES. Definition of Suicide. New York, NY: Wiley; 1985.
15.    Goodwin FK, Runck BL. Suicide intervention: integration of psychosocial, clinical, and biomedical traditions. In: Jacobs DG, ed. Suicide and Clinical Practice. Washington, DC: American Psychiatric Association; 1992.
16.    Simon RI. The suicide prevention contract: clinical, legal and risk management issues. J Am Acad Psychiatry Law. 1999;27:445-450.
17.    Fawcett J, Clark DC, Busch K. Assessing and treating the patient at risk for suicide. Psychiatr Ann. 1993;23:244-255.
18.    Fawcett J, Scheftner WA, Fogg L, et al. Time-related predictors of suicide in major affective disorder. Am J Psychiatry. 1990;147:1189-1194.
19.    Fawcett J. Profiles of completed suicides. In: Jacobs DG, ed. Guide to Suicide Assessment and Intervention. San Francisco, CA: Jossey-Bass; 1999.
20.    Murphy GE. On suicide prediction and prevention. Arch Gen Psychiatry. 1983;40:343-344.
21.    Karasu TB. Psychoanalysis and psychoanalytic psychotherapy. In: Kaplan HI, Sadock BJ, eds. Comprehensive Textbook of Psychiatry. Vol 2. Baltimore, MD: Williams & Wilkins; 1989:1442-1461.
22.    Simon RI. Discharging sicker, potentially violent psychiatric inpatients in the managed care era: standard of care and risk management. Psychiatr Ann. 1997;27:726-733.
23.  Stepakoff v Kantar, 473 NE2d 1131, 1134 (Mass 1985).