Suicide risk assessment and management remains a daunting task. Arguably, now more than ever, clinicians, families, and communities are making both independent and coordinated efforts to address the ongoing problem of suicide. Yet we have made little in the way of progress in terms of lessening the frequency of death by suicide. Such deaths represent tragedies for patients, families, and providers, frequently impact communities at large, and are often a source of adverse medicolegal outcomes for the health care systems and professionals that render care. The gravity of this public health issue warrants ongoing efforts to improve suicide risk assessment and management, including novel approaches to the problem. Certainly, the idea of suicide-specific diagnoses represents a new approach, and one worthy of further discussion and consideration. Although we are tremendously appreciative of the diagnostic entities and arguments in favor of these entities proposed by Joiner et al1 in the first column in this 2-part series, we are also compelled to offer competing arguments suggesting that there are potentially many unintended consequences related to these proposed diagnoses and suggest alternative solutions that might afford greater benefits.
ENHANCING CLINICAL CARE AND PATIENT SAFETY
Joiner et al1 argued that a suicide-specific diagnosis would enhance clinical care, and thereby increase patient safety. It does seem plausible that a diagnosis specific to suicide might help call attention to suicide risk in a manner that precipitates enhanced treatment planning around suicide risk. However, the existence of such a diagnosis, or its use in medical records, is neither necessary nor sufficient to accomplish this goal. In other words, the diagnosis itself does not guarantee appropriate clinical action, which may occur even without a suicide-specific diagnosis. Take, for example, the clinical scenario offered by Joiner et al1 involving a patient with schizophrenia who died by suicide hours after discharge from a psychiatric unit: “Nine days postadmission, the patient is discharged with a standard discharge plan, including some safety planning and a scheduled appointment with a social worker within 2 weeks of discharge. Approximately 12 hours postdischarge, suicidal ideation with lethal intent (as shown by the patient’s communications to others) develops and escalates rapidly, culminating in the patient’s death by suicide.” How, in and of itself, would a suicide-specific diagnosis change this outcome?
As illustrated in this clinical scenario, it seems that efforts to identify and mitigate risk are far more important than diagnostic entities, and such efforts may exist independent of diagnostic labels. It is entirely plausible that, in the proposed scenario, a suicide-specific diagnosis might afford no meaningful change in the care provided, and no change in terms of outcome. Of course, the intent behind placing a suicide-specific diagnosis is that the diagnosis will serve as the vehicle to bring hyperawareness of suicide risk, so that clinicians will be less likely to overlook suicide risk and therefore put in place appropriate interventions to mitigate that risk. But might it be more effective to pursue that goal via an approach that is not based on a novel diagnosis, which surely will lead to both false-positives and false-negatives?
Returning to the scenario offered by Joiner et al,1 the identified shortcoming in the assessment and treatment afforded this hypothetical individual comes down to suicide risk assessment and safety planning. “The most thorough version of suicide risk assessment occurred at intake, and the notes from that intake assessment indicated that ∼3 years before admission, the patient attempted suicide. Beyond noting its occurrence, the attempt is not further characterized … Had the parameters of the past suicide attempt been foremost in the minds of discharge clinicians, discharge planning might have proceeded differently, in such a way that this patient’s suicide might have been prevented.” Certainly, appreciating the details of the patient’s prior attempts often affords knowledge which might help to anticipate what future suicidal crises may look like, and thereby facilitate efforts to anticipate and prevent future suicidal behaviors. The suicide-specific diagnosis is, however, extraneous to that process, such that enhanced suicide risk assessment and management can, and arguably should, occur entirely independent of such proposed diagnostic entities.
The ultimate goal ought to be improved risk assessment and patient-centered risk management (eg, personalized safety plans and means safety practices specific to the individual and the unique and dynamic nature of his or her risk). Anchoring that goal to an imperfect diagnostic entity is both unnecessary and runs the risk of failing to attend to the needs of individuals whose manifestations of suicidal thoughts and behaviors depart from that dictated by proposed diagnostic entities. Surely, the suicide-specific diagnostic criteria proposed will capture the majority of people likely to suffer impending suicidal crises. But equally surely, many such identified persons will not go on to engage in self-directed violent behavior, and there will be some persons whose qualitatively atypical suicidal crises will escape detection or fall outside of any proposed suicide-specific diagnoses. Of course, not all high-risk individuals are prepared or able to acknowledge the nature and extent of their suicidal ideation and/or intent. Sometimes, the decision to die by suicide is followed by a relative period of calm. Suicide risk assessment frequently occurs during intervals in which the chronically at-risk person is not acutely in crisis, so that, even with optimal assessment in the context of a strong therapeutic relationship, the true extent of future risk may not be discernible. The point we are making is that suicide-specific diagnoses may potentially interfere with efforts to provide individualized suicide risk assessment and management plans that attend to the innumerable and idiosyncratic manners in which suicidal thoughts and behaviors may manifest. Hence, rather than confuse the process with imperfect diagnostic entities, we might instead attend directly to the risk assessment and management that is ultimately most likely to yield life-saving interventions, and other methods for communicating risk.
There are many potential ways to communicate suicide risk in the medical record or to other clinicians. Are the proposed diagnoses any better than risk assessments articulating increased risk for suicide? How does a diagnosis further facilitate prediction of an impending suicidal crisis once a risk assessment has already revealed increased likelihood for such a crisis? Another alternative method for communicating risk, and one seemingly without the potential perils and implications of an imperfect diagnostic entity, is the use of patient record flags or notifications. For example, the Veterans Health Administration requires that facilities use patient record flags which feature prominently in the electronic medical record as a method for communicating to other treatment providers that the individual is at high risk for suicide.2 The flag also requires that certain clinical actions take place (eg, outreach following a missed appointment). The appropriateness of that flag and high-risk designation are then reevaluated on a regular basis. Unlike a suicide-specific diagnosis, the flag communicates only perceived risk, and avoids the potential pitfalls of diagnostic error and the needless constraints dictated by any categorical diagnosis.
Another approach to communicating suicide risk was described in an earlier Law and Psychiatry column in this journal,3 as part of our proposed model for achieving therapeutic risk assessment of the suicidal patient.4 The method entails stratifying risk according to both severity and temporality, with those designations being associated with indicated clinical actions. Risk designations thereby inform disposition and/or treatment plans aimed specifically at the nature of the individual’s suicide risk, whether it be acute, chronic, or both. The model proposes that risk assessment and mitigation strategies, and attendant documentation, be tailored to the individual’s drivers of risk, to realize patient-centered and individualized care in a medicolegally informed manner. For example, in the case example provided by Joiner et al,1 the documentation might identify the patient as being at low acute risk and high chronic risk for suicide at the time of discharge, simultaneously denoting the appropriateness for transitioning to outpatient care, while also highlighting the tendency to become acutely suicidal if and when particular stressors occur. Family members and providers may then be enlisted for routine check-ins following discharge, and be placed on look-out for potential events that have historically precipitated acute crises (eg, relapse, relationship dissolution). The safety plan might include individualized warning signs, facilitating the ability to recognize impending crises and a tendency toward rapid escalation, followed by concrete steps for self-soothing, enlisting support, or obtaining emergency care. A suicide-specific diagnosis, by virtue of limitations inherent in a categorical scheme, is less adaptable to such individualized assessment and treatment planning, or for optimal communication concerning such assessment and treatment planning. Once again, although a diagnosis could be a part of this process, it arguably imposes needless diagnostic constraints while only adding superfluous labels.
Could the use of such diagnoses complicate patients’ ability to obtain the very treatment needed to mitigate suicide risk? A patient seeking treatment in the community might encounter various providers who are reluctant to offer services to persons carrying a suicide-specific diagnosis. In our experience, it is already not uncommon to encounter clinicians or programs that cite suicide risk as a barrier to entry; we have seen suicide risk used as the ostensible reason for turning patients away from housing programs, substance abuse programming, dialectical behavioral therapy, or even evidence-based treatment for posttraumatic stress disorder. Such instances are all the more frustrating because long-term suicide risk reduction mandates aggressively targeting the drivers of suicide risk (eg, homelessness, substance abuse, borderline personality disorder, or posttraumatic stress disorder) with precisely such interventions. Might insurance carriers utilize such diagnoses to declare suicidal thoughts and behaviors as preexisting conditions, and thus refuse coverage for subsequent acts of self-directed violence, or services provided in the context of an acute suicidal crisis? It seems possible that an unintended consequence of suicide-specific diagnoses might be making it more difficult for those individuals who are most in need to find and pay for needed care.
MITIGATING CLINICIAN LIABILITY
Part of the argument in support of a suicide-specific diagnosis is predicated on the notion that the use of such a diagnosis will afford clinicians some sort of medicolegal protection. “The bona fide utilization of a suicide-specific diagnosis would clarify discharge planning and improve safety planning, thus mitigating legal liability for clinicians. Indeed, the rigorous, routine, and scientifically backed assessment and management of suicide risk is the clinician’s most reliable protector of patient safety and mitigator of legal liability.”1 Consistent with the arguments we have raised regarding enhancing clinical care and patient safety, a suicide-specific diagnosis is neither necessary nor sufficient when it comes to issues of liability. Offering such a diagnosis absent adequate risk assessment and commensurate risk management practices will surely not afford meaningful protection from liability, nor should it. Conversely, a thoughtful and well-documented risk assessment followed by a patient-centered, individualized treatment plan that involves safety planning and means safety should afford protection from liability irrespective of a suicide-specific diagnosis.
The proposed protection afforded by a suicide-specific diagnosis would appear to be predominantly predicated upon the presence of such a diagnosis in the patient’s medical record. “When a lawyer initially reviews a potential case, all he or she typically has are the medical records. He or she is looking for a well-documented chart reflecting careful and thoughtful suicide assessments, a formulation of risk, and a plan to protect the patient from that risk. The inclusion of a suicide-specific diagnosis within the Diagnostic and Statistical Manual of Mental Disorders would indicate to a lawyer that the clinician has actually considered suicide as a possible outcome and is appropriately addressing an intervention to prevent this outcome. Including a suicide-specific diagnosis can only help a clinician to avoid a malpractice suit by showing that at least the clinician considered the issue.”1 But there are many ways to reflect consideration for suicide risk other than a suicide-specific diagnosis. Once again, our desired endpoint ought to be thoughtful and well-documented risk assessment and management. The provision of a diagnosis does not actually get us to that critical endpoint, and for the reasons previously discussed, it may actually interfere with reaching that endpoint on occasion.
The medicolegal argument in support of a suicide-specific diagnosis also inadequately attends to the potential perils for clinicians when such a diagnosis is missing from medical records among persons who die by suicide. Joiner et al1 contend that “If the suicide-specific diagnosis is absent, that fact does not change the analysis of the potential case.” Respectfully, we must disagree. Once a suicide-specific diagnosis is firmly established and becomes part of the standard of care, failing to offer that diagnosis will become a major potential source for liability.
First, it is necessary to recognize that many individuals who die by suicide will not acquire such a diagnosis, and often through no fault of clinicians. It is well established that nearly half of suicide deaths involve individuals who had not previously been recognized as being at significant risk for suicide.5–7 When it comes to suicide risk assessment, our best efforts remain very imprecise in terms of predicting death by suicide, and there is widespread agreement regarding our collective inability to do so. As discussed in a recent Law and Psychiatry column, suicide risk assessment still matters,5 not because it accurately determines who will, and who will not, go on to die by suicide, but because it is part of a therapeutic process that promotes collaborative efforts to recognize and mitigate suicide risk. In short, false-negatives will be inevitable when it comes to the proposed suicide-specific diagnoses.
When such a false-negative occurs, it is not difficult to imagine the medicolegal scenario which may ensue. By virtue of having failed to offer a suicide-specific diagnosis for an individual who died by suicide, claims involving diagnostic error seem inevitable. This represents a major risk for mental health clinicians, or any provider working with potentially suicidal persons, and a very dangerous, albeit unintended, consequence of introducing such diagnostic entities. The medicolegal risk attendant to diagnostic error is very real.
A publication exploring 25 years of malpractice claims for diagnostic error in the United States from the years 1986 to 2010 is illustrative.8 Diagnostic error is defined as a diagnosis that is missed, incorrect, or delayed, as established by subsequent definitive tests or findings. Misdiagnosis-related harm is that which may have been prevented had the diagnosis been correct, or in the case of suicide-specific diagnoses, not have been missed. Tehrani and colleagues reported that “the most frequent types of diagnostic error identified were failure to diagnose … The most frequent outcomes were death, significant permanent injury, major permanent injury, and minor permanent injury.” Needless to say, legal advocates can offer the compelling argument that a death by suicide represents a rather definitive test establishing that the deceased individual was in fact suffering from a suicide-specific diagnosis of the sorts proposed. Of course, in the case of a death by suicide wherein a suicide-specific diagnosis was not offered, legal arguments suggesting that diagnostic error resulted in death seem inevitable, claims which will carry tremendous inherent risk to health care providers and health care systems. Establishing such diagnoses for routine use will create the potential for tremendous medicolegal liability for providers even when clinical care was commensurate with applicable standards. Many clinicians and health care entities will find themselves in the unenviable and perilous position of trying to explain why they did not offer a suicide-specific diagnosis to an individual who is now dead as a result of suicide, or one who has suffered harm as the consequence of a suicide attempt. Even when clinical care is otherwise commensurate with the standard of care, it may prove very difficult to overcome the superficial implications of not having offered a suicide-specific diagnosis to an individual who went on to experience a suicide death, or even injury as a consequence of attempted suicide. As we are all aware, some individuals will fail to disclose their suicidal thoughts and intentions despite worthy efforts at risk assessment. Some persons will simply suddenly and unpredictably become acutely suicidal in a manner which neither they, nor the best of clinicians, can possibly predict.
Faced with this unavoidable dilemma, many providers will logically elect to err on the side of offering a suicide-specific diagnosis rather than face the potential liability associated with not offering such diagnostic formulations described above. That circumstance arguably will render the very diagnostic entity meaningless, defeating the very purposes for which it was originally intended. In other words, once providers start to liberally dispense the diagnosis, the frequency with which it is encountered will then serve to diminish the intended effect—prompting heightened attention to the individual’s suicide risk and creation of patient-centered plans to mitigate that risk. The idea of a suicide-specific diagnosis is cursorily compelling and not without merit, but it also would appear to be very premature upon further analysis. Best efforts at suicide risk assessment continue to suffer from our very modest ability to discern who will and who will not die by suicide. That reality, combined with the unacceptable medicolegal implications of creating suicide-specific diagnoses, offers compelling reasons to maintain these proposed diagnoses as interesting talking points and worthwhile subjects for ongoing research, but which remain far from ready for clinical applications.
1. Joiner TE, Simpson S, Rogers ML, et al. Whether called acute suicidal affective disturbance or suicide crisis syndrome, a suicide-specific diagnosis would enhance clinical care, increase patient safety, and mitigate liability. J Psychiatr Pract. 2018;24:274–278.
2. Office of Heathcare Inspections. Evaluation of Suicide Prevention Programs in Veterans Health Administration Facilities. Report No. 16-03808-215. Washington, DC: Department of Veterans Affairs, Office of Inspector General; 2017. Available at: www.va.gov/oig/pubs/VAOIG-16-03808-215.pdf
. Accessed July 12, 2018.
3. Wortzel HS, Homaifar BY, Matarazzo BB, et al. Therapeutic risk management of the suicidal patient: stratifying risk in terms of severity and temporality. J Psychiatr Pract. 2014;20:63–67.
4. Wortzel HS, Matarazzo BB, Homaifar BY. A model for therapeutic risk management of the suicidal patient. J Psychiatr Pract. 2013;19:323–326.
5. Wortzel HS, Nazem S, Bahraini NH, et al. Why suicide risk assessment still matters. J Psychiatr Pract. 2017;23:436–440.
6. Large M, Kaneson M, Myles N, et al. Meta-anlaysis of longitudinal cohort studies of suicde risk assessment among psychiatric patients: heterogeneity in results and lack of improvement over time. PLoS One. 2016;11:e0156322. Available at: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0156322
. Accessed July 12, 2018.
7. Chan MK, Bhatti H, Meader N, et al. Predicting suicide following self-harm: systematic review of risk factors and risk scales. Br J Psychiatry. 2016;209:277–283.
8. Tehrani S, Lee H, Mathews SC, et al. 25-year summary of US malpractice claims for diagnostic errors 1986-2010: an analysis from the national Practitioner Data Bank. BMJ Qual Saf. 2013;22:672–680.