Commentary: Facts and Values in Competency Assessment

The author of “Conceptualizing and Characterizing Accuracy in Assessments of Competence to Stand Trial,”¹ has contributed more than anybody else I know to the ability of forensic psychiatrists to understand the statistical aspects of their work, particularly their work in the assessment of risk and in the evaluation of civil and criminal competency. Those with an interest in the value of what we do could do worse than carry with them at all times the explanation he has provided here of why an 80 percent agreement between raters, on its own, means nothing.

Dr. Mossman reports that, “no study has reported on the accuracy of competence assessments” (Ref. 1, p 348). That these assessments nevertheless affect what happens to so many defendants should give us all pause for thought. That defendants facing serious charges should be imprisoned, and even put to death, contingent on the outcome of competency assessments raises professional and moral questions that deserve our attention. One such question is whether psychiatrists and psychologists should put these serious potential consequences of a finding of competence out of their heads when they conduct these evaluations. I think that it is unreasonable to expect them to do so.

Unreasonable and wrong in principle. The idea that the seriousness of the charge should affect both the conduct of competency evaluations and the ways in which the results of those assessments are used is not mine.² The important question is how, exactly, seriousness should be taken into account. I argued that, all other things being equal, more serious charges should lead to greater caution before the conclusion is reached that someone is fit to proceed.³ Put another way, when the charges are serious, a wider margin of error is justified before someone is declared competent. I paid less attention to the kaleidoscopic array of forms that human abilities can take. Utah's list, as Mossman points out, is long but not exhaustive. Perhaps we should be doing more to define what is required.

But perhaps not. To turn ourselves in this direction would be to imply that there is a gold standard: a real quality and quantity of competence at which a defendant should be allowed to proceed. I cannot see that such a thing exists. Mossman¹ offers three reasons to think that it might. First, he argues, the criminal courts already use what could be called a common-sense cutoff: is the defendant at a “substantial disadvantage” compared with most defendants? My difficulty here is that common-sense cutoffs have not been shown to be more valid than any other kind.

The second is that, even in the absence of a gold standard, one can still offer level-of-confidence responses that are useful to other people. I am not sure that couching our judgment in these terms takes us further forward. Mossman¹ uses the example of mammography. I suspect that terms such as “probably benign” and “highly suggestive of malignancy” are useful only to the extent that radiological diagnosis has been shown to be accurate. In other words, I suspect that the results of mammography, whether they are offered as level-of-confidence responses or are expressed in other ways, are useful only to the extent that the results of the mammogram agree with the results of a biopsy.

Finally, Mossman¹ argues, modern statistical approaches allow inferences about diagnostic accuracy in the absence of a gold standard for diagnosis. One of the recent papers he cites describes a nonparametric, maximum likelihood, method of estimating ROC curves for the sensitivity and specificity of pathology findings in the absence of a clinical diagnosis of cancer.⁴ Another paper employs a random-effects model to the same end.⁵ Both are based on a premise with which we can all agree: that there is something called cancer and that it is different from the healthy state.

Except in extreme cases, I am not sure that we can all agree that there is something called incompetence to stand trial and that it is different from normal. Incompetence to stand trial strikes me as more of a value judgment. Perhaps the most obvious way in which reasonable people might disagree as to a defendant's competence stems from the value they place on his autonomy.

Some of the aims of punishment, such as deterrence and incapacitation, can be achieved without the defendant's participating in his trial. General deterrence, for instance, seems to require only that future potential offenders know that previous offenders have been apprehended and punished. In fact, since some incompetent defendants will never be punished, requiring competence reduces the deterrent effect of punishment: the future potential offender will see a greater chance of his avoiding punitive consequences. Similarly for incapacitation. The only requirement here would seem to be that the offender be placed in a hospital or prison. The defendant's participation in his trial does not appear to be essential.

In addition to punishment and deterrence, however, we also value defendants' autonomy. While trials with competent defendants are probably more reliable than trials with incompetent ones,⁶ this may not be the reason we require trial competence. The gain in reliability is unknown and may be small. I suspect that our requirement that defendants be competent derives more from our desire for fairness⁷ and our wish to see people's dignity be respected.⁸ Crucially, we seem prepared to sacrifice some deterrence and incapacitation to achieve these ends. The difficulty, from the point of view of establishing a gold standard for competence to stand trial, is that there must be a range of opinions as to how much deterrence and incapacitation it is appropriate to give up in an individual case in order properly to respect someone's autonomy.

Goldstein and Stone⁹ divided assessors of competence to stand trial into Guardians, “worried that the worst might happen and must be averted” and Green Lighters, “inclined to let the chips fall where they may” (Ref. 9, p 95). Neither group is wrong, but they do have different priorities. There must be many such distinctions to be made among those of us who conduct these evaluations. Not everyone will agree with me, for instance, that the seriousness of the charge should be taken into account.

I do not think that competence to stand trial is a “feature of defendants” (Ref. 1, p 343). Rereading “Conceptualizing and Characterizing Accuracy in Assessments of Competence to Stand Trial” persuaded me that it is better seen as an inference, an inference based on an evaluation conducted in a particular context. There are many contexts and just as many ways of looking at them. In these circumstances I see no prospect of a gold standard. I am not even convinced that it is a laudable goal.

• American Academy of Psychiatry and the Law