Abstract
The legal standard of care for assessing and responding to suicide risk has historically been ambiguous, creating inconsistency in the testimony of forensic experts and uncertainty about clinical responsibilities among practitioners. In this article, I rigorously apply the legal concept of reasonable care to identify clinical activities that courts could collectively consider as evidence of reasonably careful suicide risk assessments. I derived six clinical activities, which I refer to as probable standards, from a review of legal scholarship in tort law, court cases involving suicidal behavior, and forensic papers on suicide risk assessment. I discuss the basis for each probable standard and offer commentary to aid in their interpretation. My intention is not to define the legal standard of care for suicide risk assessments (only courts can do so) or to create a clinical practice guideline, but to establish legally informed reference points to assist forensic experts in providing objective, consistent, and compelling testimony.
In civil cases, judges and juries must decide whether care rendered to suicidal patients was adequate in the eyes of the law. For this, they rely partially on the testimony of forensic experts. However, because there is no specific legal standard for conducting suicide risk assessments, judges and juries must sort out the legal truth from the competing testimonies of experts on both sides. To clarify matters, I reviewed negligence law, court cases, and forensic scholarship to identify a set of clinical activities that courts are likely to accept as evidence that the legal standard of care was met. I refer to these activities as “probable standards” for suicide risk assessment to convey their hypothetical nature; legal standards of care are defined by courts not by practitioners.1 When courts settle on a standard of care, they typically do so in a way that narrowly applies to the facts of one case. My intent, however, was to formulate standards that can be flexibly applied to any malpractice case involving suicide while avoiding excessively prescriptive actions (e.g., “clinicians shall …”). A probable set of standards may benefit the field by helping forensic experts assess clinical care in malpractice cases comprehensively, ground their testimony in legal concepts, and insulate their testimony from personal biases. Probable standards may also inform current suicide risk assessment practices and training curriculums for prelicensure candidates.
What is Reasonable?
The doctrine of negligence expects people to exercise “reasonable care” when their actions (or lack of action) pose a risk of injury to others. Failing to do so constitutes negligence. In most cases, jurors are entrusted with determining what constitutes reasonable care in a particular circumstance. However, they need guidance to understand what reasonable care means in situations requiring specialized knowledge, such as the treatment of a patient at risk of suicide. The law provides guidance by giving jurors specific instructions. For example, in California, jurors are informed that the standard of care (SOC) to use is “… the level of skill, knowledge, and care in diagnosis and treatment that other reasonably careful [practitioners] would use in the same or similar circumstances” (Ref. 2, p 382). Jurors in California are then instructed to rely on expert testimony to understand what “reasonably careful” practitioners do. Although the standard of medical custom (i.e., the “locality rule”) is still in use, I focus exclusively on the reasonable-person standard because only a handful of states still observe medical custom,3 and courts have historically not deferred to medical custom when, in their view, disregarding a precaution, no matter how routine, creates an unacceptable risk of harm.4,5
According to legal scholars, the reasonably careful person possesses several qualities: she is attentive, gathers information to arrive at a full appreciation of the nature and degree of risk, anticipates the likelihood and severity of harm, weighs the pros and cons of different actions, takes precautions that reduce or eliminate risk, and monitors changes in the risk picture.6 Naturally, practitioners should evidence these same qualities when completing a suicide risk assessment (SRA).
Because the reasonable-person standard is a general one, courts ask forensic experts to establish standards of medical care, relevant to a case, against which to judge practitioners. For SRAs, forensic experts have expressed a continuum of opinions about what the medical standards are. The opinions range from “there can be no standard” to “there are many specific standards.” Representing the former are Simon and Shuman7 who persuasively argue that SOCs for suicide risk assessment are “elusive” because states define the SOC differently, the facts and circumstances of a case shape the SOCs, and experts commonly disagree on what constitutes the SOC.
There are several drawbacks to this view of medical standards for SRAs. First, the absence of standards can undermine objective analysis of malpractice cases by expert witnesses.8 Second, standards that are created solely by the particulars of one case leaves practitioners rudderless: none can know what to do, but all are nevertheless expected to do it (whatever “it” is). Finally, leaving the SOC undefined does not advance the national agenda to address the profound dearth of graduate training in suicide risk assessment.9,10
For other experts, only detailed medical standards for SRAs and prescribed courses of action suffice.11,–,13 For example, Rudd and Joiner13 give a checklist of three domains and 23 areas they believe constitute the SOC for suicide risk assessment. However, the more exhaustive the medical standards are, the more likely they are to reflect ideal rather than reasonable care and thus to be more akin to clinical practice guidelines. Another concern is that a detailed checklist may lead jurors to believe that missing one or two areas is negligent. Last, overly detailed standards lack flexibility; they are too rigid to apply across settings (i.e., emergency room versus inpatient unit) and do not readily allow for developments in the science of suicide risk assessment.
Staking out a middle ground are experts who recommend certain actions to minimize liability exposure. The recommendations are typically prescriptive and are intended to avoid negligent treatment of at-risk patients rather than to identify reasonable clinical actions. For example, Baerger14 offered five recommendations: conduct an examination inclusive of current and historical suicidal thoughts and behavior; consider hospitalization or a safety plan when risk is high; reassess risk during stressful periods; maintain records that justify treatment decisions; and take special precautions for at-risk patients. Other experts offer lists of treatment failures when conducting SRAs.15
Probable Standards of Care for Suicide Risk Assessment
My focus is clarifying what the law expects of reasonable clinicians completing an SRA, as opposed to what the field of psychiatry asserts is the medical standard based on accepted practice and science. So, rather than review medical evidence to identify clinical activities that could meet the reasonable care standard, I relied on legal scholarship in tort law, court cases involving suicidal behavior, and forensic papers on SRAs. In general, I identified clinical activities that met all of the following criteria: the activities were consistent with legal scholarship on negligence, endorsed in court decisions as evidence of reasonable care, and singled out by forensic experts as likely to meet the standard of reasonable care. Next, I considered how the clinical activities should be qualified to acknowledge the realities of day-to-day treatment. For example, gathering past records is a clinical activity that often depends on the patient's degree of cooperation. Finally, I formulated probable standards. By probable standards, I mean likely expectations of clinical care, as they pertain to SRAs, based on the legal concept of the reasonably careful person. My intent is to communicate legally informed expectations for reasonable care that forensic experts can flexibly apply to diverse clinical situations. Each of the following sections begins with the probable standard followed by commentary that elaborates on the standard's basis and supports its interpretation.
Probable Standard 1: Gathering Information from the Patient
To the extent that the patient is cooperative and the treatment context permits, the clinician inquires about current suicidal thinking, surveys current and historical suicide risk factors, and assesses mental status.
The law expects the reasonable clinician to diagnose and treat patients competently. Courts have concluded that making a diagnosis means “ascertaining a patient's medical condition through examination and testing” (Ref. 16, p 7) and have frequently held that diagnosis includes assessing suicide risk,17 which involves components such as assessing mental status,18 taking an adequate psychosocial and self-harm history,19 and thorough follow-up questioning regarding psychiatric symptoms.20 Thus, omitting an SRA as part of the diagnostic progress is akin to failing to appreciate risk6 and is evidence of inattentive care.
Courts also have accepted that suicide risk is established by evidence of a variety of risk factors21 and that demographic factors alone are probably inadequate.22 Thus, Standard 1 requires clinicians to appreciate suicide risk fully by covering the major areas of inquiry: current suicidal thinking and behavior, current and historical risk factors, and mental status. Clinicians should also personally complete an SRA, an expectation shared by the courts.23,24
Many forensic experts believe that the activities in Standard 1 constitute minimum thresholds for reasonable care. For example, Berman and colleagues assert: “The assessment of risk involves, at a minimum, attention to the possibility of suicidal behavior through the asking of questions about suicidal thoughts, plans, intent, and actions, in addition to known risk factors” (Ref. 25, p 260). Similarly, Beckson and Penn state that “A good faith psychiatric interview and examination of the patient is required to meet the standard of care. Questioning the patient about suicidal ideation, intent, and plan is required” (Ref. 26, p 17). However, only inquiring about suicidal ideation, plans, and means is “grossly inadequate for defending against allegations of negligence” and “A suicide assessment that focuses solely on the here and now is very likely to fall below the standard of care” (Ref. 27, p 3). Finally, in their analysis of treatment failures that expose practitioners to liability, Packman and colleagues15 urge clinicians to obtain a thorough self-harm history.
Standard 1 recognizes that a reasonable clinician's efforts are subject to a patient's cooperation and to the treatment context. There is no guarantee that the patient will disclose suicidal thinking when asked, and patients may intentionally mislead the clinician to avoid detection.28 Some conditions may interfere with reliable reporting as well (e.g., psychosis or substance intoxication).29 When patients knowingly withhold pertinent information (like a history of self-harm), courts have not found clinicians liable.30,31 On the other hand, courts may expect clinicians to know more about long-term patients. In Perez v. United States, the court concluded that, in longer treatments, “the provider has (or should have) greater knowledge of a patient's specific psychiatric status and suicidal intentions and can better prescribe and administer a course of action” (Ref. 18, p 120).
Probable Standard 2: Gathering Data from Other Sources
Whenever relevant and possible, the clinician reviews pertinent documentation, makes reasonable attempts to obtain past records, and collects collateral reports from other professionals, family, or significant others.
The Restatement Third,6 the widely respected reference in tort law, states that, sometimes, anticipating harm requires engaging in an effort to gain full appreciation of the present dangers. Courts “take into account the likely benefit in risk reduction the actor could have achieved by endeavoring to gather more information before engaging in conduct, and also the burden the actor would have borne in making such an effort” (Ref. 18, p 33). This notion bears out in cases involving suicidal behavior: negligence is often found when records have not been obtained or reviewed, partly because failing to do so allowed critical information to go undetected. For example, in Bell v. New York City Health & Hospitals Corp.,20 a psychiatrist was found negligent, in part because he did not obtain past treatment records, which documented three previous suicide attempts. Similarly, failing to review available records, which contained a history of hospitalization for thoughts of self-harm, supported a finding of negligence.18,32 Thus, courts appear to take the perspective that the burden on clinicians to inform themselves costs little in time and effort, especially when compared with the severity of possible harm to the patient.
Standard 2 includes critical qualifiers (“whenever relevant and possible,” “pertinent,” and “reasonable attempts”). On the face of it, obtaining past records seems a shoo-in for any standard on gathering clinical data. However, courts seem to recognize that what practitioners should know depends on the circumstances, and not reviewing records has not always been found to be a proximate cause of injury (for a review of illustrative cases, see Roach et al.33). The reality of day-to-day practice is messy. Obtaining collateral reports in low-risk situations, for example, is often unnecessary. Higher-risk situations, of course, demand that clinicians be more thorough before formulating a diagnosis and plan. However, even here, the reasonableness of a clinician's efforts to obtain and review records should be judged in the light of the patients' willingness to sign releases of information, the urgency of the situation (emergency rooms visits, home visits by psychiatric emergency response teams), and the length of clinical contact (e.g., hospitalizations lasting 72 hours). In some cases, obtaining past records quickly is difficult if not impossible.
Even when records are readily available, Standard 2 requires that clinicians review only pertinent records. Of course, what constitutes pertinent is not always straightforward. As Rogers et al. highlight: “Do psychiatrists have time to sift through hundreds of pages of documents of patients' medical records? Do psychiatrists need to obtain records from 1 year ago, 5, 10, or 20? Should just mental health records be obtained? What about from a patient's primary care provider?” (Ref. 34, p 453). Tellingly, psychiatrists do not commonly request records older than one year for moderate-risk patients.35
In addition, clinicians must balance the need for information with preserving therapeutic rapport. In the early sessions, when past records are typically requested, rapport is tenuous with difficult and disturbed patients and must be weighed against the intrusion of privacy that patients may experience when pressured to involve collaterals. The matter is simpler when working with children and adolescents, but parental cooperation can be an issue. On the other hand, clinicians must balance confidentiality with patient safety. In cases where the clinical picture suggests that collateral information is critical to a complete SRA, clinicians must weigh the risks and benefits of preserving privacy and safeguarding the patient.36 That said, Simpson and Stacy warn: “Comments after the fact that one didn't call relatives or prior caregivers for information because of ‘confidentiality’ ring hollow to a jury when it is obvious that the patient was in danger” (Ref. 27, p 188).
Finally, Standard 2 is informed by the common forensic wisdom: Do not over-rely on patient's report.37 Past records and collaterals can resolve discrepancies in the patient's report, reveal clinically significant behavior, and confirm denials of suicidality. Even courts have opined that relying solely on a patient's denial of intent is not acceptable.18
Probable Standard 3: Estimating Suicide Risk
The clinician estimates the degree of suicide risk based on collected information.
The law expects the reasonable clinician to anticipate harm, that is, to exercise foreseeability. According to the Restatement Third: “Primary factors to consider in ascertaining whether the person's conduct lacks reasonable care are the foreseeable likelihood that the person's conduct will result in harm, the foreseeable severity of any harm that may ensue, and the burden of precautions to eliminate or reduce the risk of harm” (Ref. 6, p 29). The suicide risk estimate is akin to foreseeable likelihood because it summarizes risk data that the clinician can use to anticipate harm. Greater risk indicates more foreseeability. In SRAs, the foreseeable severity of harm is always serious injury or death that can occur as a result of a suicide attempt. Therefore, it is the suicide risk estimate that aids clinicians in determining the burden of precautions they must take to safeguard patients from suicide.
Courts accept that suicide is not predictable38 and that harm is best foreseen by assessing risk. Not only do courts understand what risk factors are (i.e., they frequently describe relative risk and are not causal), they appreciate that a consideration of risk factors is a logical approach to determining the degree of danger “because of the difficulty of tracing exactly whether and how a given action combines with other factors to directly ‘cause’ a particular death” (Ref. 39, p 17). Furthermore, courts grasp concepts such as chronic factors, acute factors, and warning signs (although may not refer to them as such). For example, in Keebler v. Winfield Caraway Hospital, the Alabama Supreme Court found that suicide is foreseeable when a person has a “history of suicidal proclivities,” has “manifested suicidal proclivities” to the defendant, or was hospitalized for a suicide attempt and treated by the defendant (Ref. 40, p 845).
Standard 4 avoids specifying how clinicians should estimate risk. Estimating suicide risk is an artful exercise in applying scientific knowledge, partly because the science of risk estimation for suicide is still young and partly because it will always require clinicians to extrapolate empirical findings based on groups of patients to a particular patient. In addition, we know little about how risk factors combine to elevate risk. To standardize how clinicians describe risk estimates, some researchers have proposed guidelines for rating overall suicide risk,41 whereas others have proposed estimating chronic and acute risk levels separately.42 Current risk classifications rely heavily on the progression from suicidal ideation, to intent, to plans, for which research shows mixed support,29 and they do not yet incorporate imminent warning signs. Existing classification schemes are compelling syntheses of science and clinical wisdom, but no scheme has been shown to correlate with suicide attempts or suicides. Moreover, the clinical estimation of risk is still evolving,43 and even suicidologists are poor at identifying imminent risk.44 Nevertheless, Standard 3 expects clinicians to assess the degree of suicide risk because the consequences of not doing so are potentially grave.
Probable Standard 4: Treatment Planning
When there is substantial risk of suicide, the clinician formulates and follows through on a treatment plan, the components of which reasonably correspond to the severity of the suicide risk estimate.
The law views the doctor–patient relationship as one that imposes an affirmative duty, that is an obligation, to protect patients from harm.6 When suicide risk is elevated, courts expect clinicians to take reasonable steps to prevent suicide,45 regardless of the treatment setting.46 As discussed earlier, in determining what preventive responses are “reasonable,” courts expect clinicians to consider the likelihood of harm (degree of suicide risk) together with the gravity of resulting injury (serious self-injury or suicide). More simply, the rule is “the greater the danger, the greater the care” (Ref. 6, p 46). In an SRA, this means that a higher suicide risk estimate should result in more intensive or invasive clinical intervention.
Treaters commonly defend their treatment decisions as based on clinical judgment. Standing alone, however, such an assertion is not evidence of reasonableness. To courts, clinical judgment is evidenced by a “balancing test.” For example, in Johnson v. United States, the court concluded that “the treating physician must exercise his judgment and balance the various therapeutic considerations together with the possible dangers” (Ref. 38, p 1293). Thus, clinicians are expected to weigh the cost of preventive measures against the benefit they provide, where cost refers to the amount of effort or expense required by the treater to accomplish the preventive step, and benefit refers to the degree of risk reduction a particular preventive step is expected to yield. In general, legal scholars believe that, as serious harm becomes more foreseeable, arguments of cost become less defensible.6 Based on this, common clinical responses to suicide risk (e.g., hospitalization, communicating risk to family or significant others, means restriction or ordering close observation) are likely to be viewed as incurring little cost to clinicians compared with the substantial reduction in risk that such responses could produce.
Forensic experts have been less specific about the SOC and treatment planning for at-risk patients, likely because it is impractical to address every type of clinical intervention. What they often recommend, however, is that clinicians make treatment decisions based on a risk–benefit analysis.47,48 Standard 4 also requires clinicians to use a common principle in health care: prescribe care that is commensurate with the severity of symptoms and risk of future harm (e.g., as heart disease worsens, interventions become more aggressive and invasive). This type of measured response is also widely accepted in the treatment of suicidal thinking and behavior.49,50 Reid is especially emphatic: “First, protect the patient.”51
Probable Standard 5: Documentation
The clinician documents the findings of the suicide risk assessment and, when substantial suicide risk exists, the rationale for the selected course of treatment.
Many court decisions involving patient suicide clearly show that documentation is necessary to prove that reasonable care occurred. In Abille v. United States, the court concluded that a psychiatrist's “failure to keep contemporary progress notes reflecting his exercise of judgment, and the basis for it, was below the standard of care” and that this was the proximate cause of a patient's suicide (Ref. 52, p 14; see also Bell20). In Campbell v. Kelly,53 the court was skeptical that a psychiatrist's habit of documenting only “positive” findings was evidence that an SRA was completed. In other words, the absence of notes is not a reliable indicator that an SRA was made. Finally, in Perez v. United States,18 the court asserted that neglecting to note risk factors for suicide was a deviation from the standard of care in the treatment of a suicidal patient.
Thus, without documentation clinicians cannot establish whether they met any of the previously described standards. As Simpson and Stacy observe: “Since suicide is one of the worst possible outcomes for a psychiatric patient, most juries conclude that if a psychiatrist actually conducted a suicide assessment, he or she surely would have documented it” (Ref. 27, p 186). Because hindsight bias can distort how experts and juries understand clinical events,54 the clinical record establishes precisely what data clinicians relied on and how they used it to arrive at a suicide risk estimate. The goal is to show that reasoned judgment was exercised, not that the suicide risk estimate was “right” or correctly predicted suicidal behavior.55
Probable Standard 6: Monitoring
The clinician updates the suicide risk estimate when there are clinically significant changes in the patient's circumstances or condition and reassesses risk at significant treatment junctures.
As discussed in Standard 3, the law insists that the reasonable person foresees harm. However, legal scholars point out that, as circumstances change, so does the likelihood of harm.6 Consequently, so long as there is a duty to care, the reasonable person is obligated to monitor the risk picture. Courts have applied this reasoning to the care of the suicidal patient. In Perez v. United States, the court stated: “To the extent that a mental health patient continues to receive care from a provider, the duty to render a proper diagnosis is ongoing” (Ref. 18, p 86). In addition, findings of negligence frequently occur when suicide risk was not reassessed at critical treatment transitions such as psychiatric discharge20 or decisions to lower safety precautions.52
As numerous experts have pointed out, SRAs are not static: psychiatric symptoms fluctuate, suicidal urges wax and wane, contributing factors worsen or fade, and new events can improve or exacerbate the clinical picture.25,56,57 Rudd and colleagues recommend that clinicians: “Routinely monitor, assess, and document a patient's initial and ongoing suicide risk and document interventions for maintaining outpatient safety until suicidality has clinically resolved” (Ref. 58, p 442). Reid goes even further: “A single risk assessment is often not enough to meet the standard of care” (Ref. 59).
Further Qualifications
Discussions of reasonable care run the risk of imposing impossibly high expectations upon practicing clinicians, so it bears repeating that the probable SOCs discussed above are an attempt to apply the legally defined qualities of the reasonable careful person to the assessment of suicide risk. The probable SOCs are inferences, based on legal scholarship, court cases, and the opinions of forensic experts, about how judges and juries might apply the legal standard of reasonable care to SRAs. The probable SOCs are not a substitute for a jurisdiction's legal definition of reasonable care, are not standards of ethics for professional behavior, and are not medical standards for conducting SRAs. The probable SOCs can assist but cannot replace the reasoned analysis that forensic experts must apply when preparing to testify about what the medical standard of care is in a given circumstance.
Furthermore, I have intentionally used the qualifier “probable.” The probable SOCs are a set of clinical activities that could meet, not assuredly meet or the only way to meet, a legal standard of reasonable care. Numerous variables could lead fact finders to other interpretations of what is reasonable care for SRAs. For example, juries may be affected by hindsight bias and their own notions of right and wrong, or they can adopt a strict liability perspective believing that a clinician could not have provided reasonable care if the patient died by suicide.60 Appellate courts, on the other hand, are more likely to appreciate how forensic testimony corresponds to finer points of law. Variations in law across jurisdictions and the facts and circumstances of a case can also substantially alter the meaning of reasonable care.
The probable SOCs should not be rigidly applied or used in all cases involving suicidal behavior. They are most appropriate for analyzing clinical situations in which there is a clear need for a formal SRA (e.g., upon admission to an inpatient unit for self-harm). In outpatient settings, clinicians commonly conduct a brief screening (e.g., ask about current and past suicidal ideation) to determine whether a full SRA is needed. In this case, the probable SOCs that I have described are not relevant. Even when a formal SRA is indicated, there may be circumstances where the probable SOCs cannot be straightforwardly applied. For example, the probable SOCs do not address situations in which the care of a suicidal patient is shared by several treatment teams across shifts or days.
Finally, the probable SOCs are not clinical practice guidelines for SRAs. Clinical practice guidelines prescribe treatment that is directly informed by the best available medical evidence and are usually associated with best practice.61 The probable standards were derived from legal and forensic sources not from medical evidence. As such, they comport with the legal definition of reasonable care rather than with medically optimal care. Although I hope that forensic experts use the probable standards that I identified as a guide for analyzing cases and preparing testimony, the standards themselves, strictly speaking, are not treatment guidelines. They are reference points for forensic experts to use when determining whether a formal SRA could meet the legal standard of reasonable care. Forensic experts must still address whether the methods used in an SRA were consistent with accepted practice and science, form an opinion about how the defendant's actions caused injury, and guard against using their clinical preferences as the medical standard of care.
Conclusion
Any attempt to define standards will naturally be controversial: some experts will view the clinical activities in some standards as excessive, whereas others will assert that key activities were left out. Nor are the probable standards for SRAs that I have described intended to resolve all ambiguities in defining the standard of care; forensic experts must still study the facts of a case to form an opinion about whether care was negligent. My hope is that these probable standards can support objective forensic analysis and, at the same time, help practitioners understand what a legally adequate assessment of suicide risk involves.
Acknowledgments
The author wishes to thank Skip Simpson and William H. Reid for their incisive feedback.
Footnotes
Disclosures of financial or other potential conflicts of interest: None.
- © 2017 American Academy of Psychiatry and the Law
