The Multidimensional Assessment of Dangerousness: Competence Assessment
in Patient Care and Liability Prevention
Bull Am Acad Psychiatry Law, Vol. 14, No. 2,1986
Thomas G. Gutheil, MD; Harold Bursztajn, MD; and Archie Brodsky,
BA
Drs. Gutheil. Burwtajn, and Brodsky are afiliated
with !hi Program in Psychiatry and the Law. Massachusetts Mental Health
Center and Haward Medical School. Boston. MA. Address reprint requests
to Dr. Gutheil. 74 Fenwood Rd., Boston. MA 02115.
We offer a multidimensional model for
assessing dangerourness of patients in relation to their competence
to engage in informative dialogue with clinicians. This model, though
based on clinical considerations, may offer some degree of liability
protection. Current clinical and legal trends are reviewed as they
bear upon this determination.
Clinicians attempting to assess and then predict dangerousness in psychiatric
patients have approached the task with serious reservation [1-3]
based on the clear predictive limitations demonstrated by empirical studies.
[4,5] Empirical findings demonstrate
relatively low accuracy of prediction and low foreseeability of dan-
gerous acts by patients. Nevertheless, civil courts in some recent case
[6-8] and even the Supreme Court [9]
have continued to act as though dangerousness were clearly predictable.
Proceeding from this premise, courts then rule as though failure to predict
and then to prevent dangerousness constituted negligence. [10]
The result has been a marked expansion of liability for clinicians. [11]
Determination of liability is subject to numerous variables—nature
of case, publicity given to the dangerous act, quality of lawyers involved,
expert witness testimony, wisdom and attitude of the judge, etc.—that
would appear to defeat any attempt to derive some guidelines for clinicians.
We would like to suggest, however, that increased understanding of one
legal current may actually aid the clinician in decreasing liability
risk and in improving patient assessment. To understand the point at
issue, we must first identify a latent assumption in the law.
Legal Views
In cases involving liability for suicide, [12] courts
have often tended to view suicidal mental patients as globally incompetent,
often without any evidence for this view. While the courts in question
often appear to be unaware that this process is occurring. the patient
is essentially portrayed as a child in the hands of parentally responsible
clinicians. [13] In such a characterization. the "child-patient" cannot
really be seen as responsible for his/her actions, since he/she is "just
a child": instead, the "parent-clinicians" (the responsible
adults) must be held responsible or, in this case, liable (note, parenthetically,
the paradox that, when the issue is the right to refuse treatment instead
of malpractice, similar patients have been portrayed as globally competent!).
Starting from the premise of global incompetence, courts then employ a
unidimensional model of dangerousness and its control (Fig. 1). The degree
of control over or restraint of the patient expected to be exercised
by the clinician varies directly and simply with the degree of patient
dangerousness to self or others; greater danger calls for greater controls.
Within this model, if the clinician exerts too much control relative
to the danger (point A), the patient's civil rights may be seen as compromised;
if too little control is exerted (point B), the clinician may be deemed
to have been negligent if the patient harms someone else or himself/herself.
As earlier noted,the clinicians appear in a number of court decisions
to be the only actors: the patient is inert. Clearly, in any legal case
this model operates by hindsight, with the attendant illusion of certainty
and conviction of accuracy inherent in that vantage point. [12]

Some courts, however, have grasped the important distinction, usually
intuitively understood by clinicians, between outpatients and inpatients
and the varying degrees of control that can be exercised over these disparate
groups. Outpatients are viewed as more broadly competent for their actions;
hence. clinician control is diminished and the resulting liability when
bad results occur is decreased. A suicide liability case [14]
recently made this explicit, capturing an important clinical refinement
in the assessment of the suicidal patient (one form of dangerousness).
In a presumably competent outpatient, the patient's suicidal intent itself,
not medical negligence, was deemed to be the causal factor in the suicide;
hence, the clinician was not liable.
Proposed Model
A model that hews closer not only to actual practice but to clinical applicability—and
perhaps to liability reduction—is the multidimensional model pictured
in Figure 2, designed by members of the Program in Psychiatry and the
Law at the Massachusetts Mental Health Center. This model is based on
the clinical observation that two essential elements of clinical work,
informed consent and the therapeutic alliance, [15]
assume and require some level of patient competence: specifically, the
patient's preserved ability to engage in a dialogue with the clinician
to weigh the risks and benefits of his or her actions—this weighing
constituting a reasonable definition of socially valid responsibility.
Such capacity to engage in this process with another human being is essential
for the sort of deliberate, mature decision making which represents the
patient's competence to inform the clinician of potential self-harm or
violence or, comparably, competence to handle responsibly a pass or some
other increase in freedom.

The form of competence here envisioned begins with the clinician conveying
to the patient, "I can only help you if you level with me"; "I
can't know what you are concerned about unless you tell me in words or
actions"; and similar communications. The clinician then determines
whether the patient understands this issue. This determination is essentially
identical to deciding if the patient is capable of informed consent;
the documentation here might convey that the patient appeared to be capable,
by specific assessment, to weigh the risks and benefits of giving or
withholding information. Should the patient's assurances of safety later
prove to be deceptions, the record will demonstrate that the patient
could have revealed his/her destructive plan, but competently elected
not to; the patient was not incompetent, so as to be beyond electing
anything.
This approach is rooted solidly in clinical observation. For patients
who are incompetent to participate as reliably in a decision-making alliance,
the tie to the clinician (as both someone to live for and someone to
inform about, say, homicidal or suicidal pressures or future plans) is
not as available to aid in preventing those clinical states that lead
to dangerous actions-states which represent greater risk in part just
because the patient feels isolated in this way. Thus, assessing this
specific competence has direct clinical utility in measuring the strength
of the doctor-patient relationship as it relates to important judgments.
We define axes for both the staff's assessment of the patient's dangerousness
(i.e., short-term dangerousness within the limits of clinical predictability) and the
patient's level of impairment of competence to inform clinical staff
about suicidality or danger to others. Patients in specific clinical
states of varying dangerousness or incompetence may be "mapped" or
located on this schema. Such "mapping," of course, is intended
as an heuristic device; we cannot claim that incompetence can be precisely,
quantitatively assessed, nor that assessment of danger and incompetence
can be strictly biaxial or can be measured in comparable "units," implying
that these two parameters necessarily have equal standing.
We can then identify two decision areas—the area of patient assumption
of risk and the area of clinicians' assumption of risk; a border zone
of negotiation and two boundary demarcations, the hospitalization threshold
and the commitment threshold. Thus, although there are only two axes,
a multidimensional model is produced.
The inner white area labeled "Patient Assumes Risk" represents
a clinical state wherein the patient's level of danger and level of impairment
of competence to inform are both low. A patient whose clinical condition
can be described within this region is "his/her own master"—a
responsible adult, despite the presence of mental illness. Here, the
patient's own decisions should be respected and negotiations regarding
expansion of liberties, privileges, and the like should take place within
the parameters for informed consent; [15-18] the patient's
participation and consent should be treated as valid. Many patients seen
in practice, particularly those in the phases of recovery and on the
verge of discharge, fall within this realm.
In contrast, patients whose level of assessed danger and/or whose
level of incompetence locates them beyond the gray zone, in
the outer white zone in Figure 2, are either too dangerous or
too incompetent to make their own decisions as to their actions. Such
patients require active, even unilateral, intervention by their caretakers;
these interventions may take the form of legitimate constraints, restrictions,
and substitute dedsion making (by a guardian, for instance).
Finally, the gray zone defines an area where further observation,
data gathering, or negotiation with the patient and others should take
place to "locate" the patient more precisely on this schema.
This area is deliberately pictured as broad, in order to convey the idea
that there is no "bright line" separating those clinical states
where the clinician must
"take over" from those in which the patient may safely act
autonomously. Because this zone is also an area of uncertainty, close
monitoring of patients
"situated" in this region is clearly essential.
The two thresholds bounding this gray zone are straightforward.
As either danger or incompetence increase, a patient may require hospitalization,
preferably voluntary but negotiable; as danger or incompetence increase,
the patient's condition may reach the threshold where involuntary commitment
is required (Fig. 2).
The value of this multidimensional conceptualization of a patient's level
of safety or danger is that it offers the clinician a surer sense of
what is required, as well as a systematic, clinically usable framework
within which to make difficult decisions that may have tragic outcomes.
For example, a patient at A in Figure 3 is highly dangerous
despite little competence impairment. Examples might include the violent
psychopath, perhaps, or dangerous paranoid patient. The need for the
clinician's control of the situation under these conditions is clear,
no matter how "together" and "sane" the patient may
appear. Comparably, the patient at B in Figure 3 is not acutely
dangerous, but is so competence impaired as not to be able safely to
make his/her own decisions: some aggressive, inarticulate patients might
be located here. In such cases, again, the caretakers must accept their
responsibilities.

To see the value of this schema in more problematic cases, consider the
patient at C in Figure 3. The patient's moderate level of danger,
when coupled—as it were, synergistically—with moderate incompetence,
is just sufficient to require the caretakers to take the responsibility.
An example would be a patient just beginning to recover from a violent
acute psychotic episode, who has not fully reconstituted.
In contrast, a patient at D in Figure 3, although moderately
dangerous, remains sufficiently competent to exercise judgment and thus
to be expected to bear the responsibility for his/her actions, like the
average citizen. Some impulsive borderline patients might fit this category.
[19] In all of these cases, the clinician would retain
the burden of actually performing and especially documenting the assessment
of the patient's competence. Relevant data might include the patient's
history of openness or guardedness, honesty or duplicity, ability to
consider the value of informing caretakers of his/her intentions, and
the like.
Discussion
This diagram is a heuristic model rather than a graph of empirical measurements;
it is intended to offer a way of thinking about a complex subject. We
propose that this schema permits more useful and accurate assessment
of actual patient dangerousness and permits the clinician to determine
more accurately when to intervene unilaterally and when to seek the patient's
active participation in decision making through informed consent—improvements
that clearly redound to the patient's benefit.
This framework holds additional value for liability prevention. If the
patient's competence is documented and available in the chart for use
in litigation it is more difficult for the court to portray the patient
as a helpless, incompetent child under the exclusive control of the caretakers.
Such a vision paints the clinician's actions as entirely detemerminative
of any outcome, as though the clinician's actions (and, presumably, negligence) "caused" the
dangerous act that led to the lawsuit. [20]
Conclusion
We offer a multidimensional model for assessment of dangerousness in relation
to incompetence, based on theoretical considerations, clinical experience,
and empirical study of decisions in legal cases. We suggest that this
model offers a clinically useful guide for the clinician's approach to
decision making with difficult patients. In addition, whereas legal outcomes
can never be guaranteed because of the multiplicity of variables involved,
this schema may offer some protection against inappropriate assignment
of liability.
Acknowledgment
We acknowledge our indebtedness to Ms. Elyse
Littaye for assistance with the manuscript and to Dr. Philip Brown and
Ms. Joyce Nevis-Olsen for valuable suggestions.
References
-
Steadman HJ: Predicting dangerousness among
the mentally ill: Art, magic and science. Int J Law Psychiatry 6:381-90,
1983
-
Cocozza JJ, Steadman HJ: Prediction in Psychiatry:
An example of misplaced confidence in experts. Soc Problems 25:265-76,
1978
-
Carter CA: Some conditions predictive of suicide
at termination of psychotherapy. Psychother Theory Res Pract 8:156-7,
1971
-
Monahan J: The prediction of violent behavior:
Toward a second generation of theory and policy. Am J Psychiatry
141:10-5, 1984
-
Motto JA. Heilbron DC, Juster RP, Bostrom
AG: Suicide risk assessment: development of a clinical instrument,
in Proceedings of the Fourteenth Annual Meeting of the American Association
of Suicidology. Albuquerque, NM, 1981
-
Jablonski v. U.S., No. 81-5786
(9th Cir. June 14, 1983)
-
Hedlund v. Superior Court of Orange
County, 194 Cal. Rptr. 805 (Cal. 1983)
-
Petersen v. Washington, 671 P.2d
230 (Wash. Sup. Ct. 1983)
-
Barefoot v. Estelle, 103 S. Ct.
3383 (1983)
-
Davis v. Lhim, 335 N.W. 2d
481 (Mich. App. 1983)
-
Appelbaum PS: The expansion of liability
for patients' violent acts. Hosp Community Psychiatry 35: 13-4, 1984
-
Gutheil TG, Bursztajn H, Swagerty E, Brodsky
A: Magical thinking in suicide litigation: I. Causality, modality
and confusion in reasoning about negligence. Presented at the Annual
Meeting of the American Academy of Psychiatry and Law, Nassau, The
Bahamas. October 24-28. 1984
-
Gutheil TG: Malpractice liability for a
patient's suicide. Legal Aspects Psychiatr Pract 1:l-4, 1985
-
Speer v. U.S., 512 F. Supp.
670 (N.D. Tex. 1981)
-
Gutheil TG, Bursztajn H, Brodsky A: Malpractice
prevention through the sharing of uncertainty: Informed consent and
the therapeutic alliance. N Engl J Med 311:49-51, 1984
-
Gutheil TG, Appelbaum PS: Clinical Handbook
of Psychiatry and the Law. New York, McGraw-Hill, 1982
-
Bursztajn H, Feinbloom RI, Hamm RM, Brodsky
A: Medical Choices, Medical Chances. New
York, Delacorte Press, 1981
-
Bursztajn H, Gutheil TG, Hamm RM, Brodsky
A: Subjective data and suicide
assessment in the light of recent legal developments: II. Clinical
uses of legal standards in the interpretation of subjective data.
Int J Law Psychiatry 6:331-50, 1983
-
Gutheil TG: Medico-legal pitfalls in the
treatment of borderline patients. Am J Psychiatry 142:9-14, 1985
-
Gutheil TG, Bursztajn H, Hamm RM, Brodsky
A: Subjective data and suicide
assessment in the light of recent legal developments: I. Malpractice
prevention and the use of subjective data. Int J Law Psychiatry
6:317-29, 1983