Anticipating DSM-V: should psychosis risk become a diagnostic class?

In this issue, Woods et al1 report data analyses that support the validity of the criteria for identifying the prodrome stage of psychotic disorders. We have previously placed emphasis on this issue with a special theme.2–8 Questions related to the development of Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) (DSM-V) now emerge. Should clinical practice move in the direction of early detection of risk status? Do research findings warrant interventions at the prodrome stage in general clinical practice? If the answers are yes, then DSM-V and International Classification of Diseases, Twelfth Revision, will be challenged to provide clinicians with a diagnostic category to support this shift in practice. This will be an essential step for providing education, encouraging clinical intervention, and supporting research on these clinical activities. The DSM-V workgroup responsible for psychotic disorders is considering creating a new diagnostic class for this purpose. Although I chair the DSM-V Psychoses Workgroup, this editorial represents personal views, not recommendations, of the Psychosis Workgroup or official decisions of the DSM-V Task Force. These views are informed, however, by discussions with workgroup members (http://www.psych.org/MainMenu/Research/DSMIV/DSMV.aspx), especially Ming Tsuang, and advisors Tom McGlashan and Scott Woods.

A number of questions must be resolved before moving forward.

Does a critical clinical need exist?

If so, does Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) (DSM-IV) provide diagnostic classes that adequately address this need?

If not, can the reliability and validity achieved by investigators be accomplished in clinical practice by primary care and mental health diagnosticians?

When prodromal or high-risk cases convert to an established DSM-IV illness, is it predominantly psychosis? Is it specifically schizophrenia? Should the new class be placed with the group of psychoses or placed elsewhere?

Do the diagnostic criteria reflect behaviors so common among adolescents and young adults that a valid distinction between ill and nonill persons is difficult, resulting in too many false-positive diagnoses?

Will placing individuals in this diagnostic category do more harm than good because of stigma or the unwarranted administration of treatments with a poor benefit/risk ratio?

Studies in a number of countries document the validity of prepsychosis risk detection by observing nontrivial rates of conversion to psychosis in a short time period (Woods et al1). Data from controlled studies suggest that clinical intervention may be effective in delaying or preventing exacerbation into psychosis, but evidence to date is very weak.9,10 Screening for psychosis risk is certainly controversial.11,12 Nonetheless, there is a clinical need for methods of identification and intervention preceding fully manifest psychosis. Predictive power for progression along this path is enhanced by family history of psychotic illness, but the gain in predictive accuracy is offset by a sharply reduced sensitivity to detect cases, thus creating many false-negative classifications. Enhancing risk determination using endophenotype and biomarker assessment is a future goal, but current consideration of risk criteria for clinical application depends largely on manifest psychopathology. Success with these criteria to date suggests a compelling need for a diagnostic class, a view supported by the help-seeking status of cases and families in these studies.

An adequate approach is not provided in DSM-IV. The state-like and duration-related criteria exclude these cases from trait-like personality diagnoses including schizotypal personality disorder. Cases are also excluded if they manifest fully formed psychotic symptoms. This represents “converting,” and psychosis diagnoses are then applicable. Over time, if the prodromal symptom criteria are stable, the case may meet schizotypal personality disorder criteria, but this is not useful for early intervention. Eventually, cases may meet other criteria such as those for mood disorder, or they may simply recover and not have a definitive diagnosis. Hence, there is space for a risk syndrome class that allows for various outcomes but is redundant with no other diagnostic class.

The prodrome or high-risk concept (of which there are several versions) can be assessed in individuals with sufficient reliability and validity to justify potential application as a diagnostic class. However, it remains to be determined if utility in research will generalize to clinical practice. Work to date has been based on formal diagnostic assessment interviews and supplemental information often not available in general practice. Reliable and valid application of criteria may be especially challenging in primary care and in areas with inadequate numbers of trained psychiatrists. For DSM-V, this means that reliability of proposed criteria must be evaluated in field trials.

The question of where to place a new class is complicated. The work to date has been oriented to risk for psychotic disorders, and the criteria are mainly minor expressions of psychotic-like psychopathology. Case conversion is generally based on increasing symptom severity resulting in recognition of psychosis. Diagnosis on conversion usually falls not only within the Schizophrenia and Related Disorders section but also mood disorders with psychotic features. The field will, no doubt, develop risk criteria for other diagnostic groups, but at present the work relates to the psychoses. Thus, the natural home for a new diagnostic class is with the psychotic disorders. However, the resulting potential for stigma and unwarranted treatment merits caution. Psychotic-like phenomena are fairly common in the general population, as Woods et al1 (this issue) point out. Establishing a mental illness diagnosis and initiating treatment are stigmatizing and can cause harm to individuals. Initiation of treatment can have adverse effects, especially associated with antipsychotic drugs. Clinicians could be caught between 2 compelling, but mutually exclusive, goals: identify illness and institute treatment at the earliest and most effective time vs the traditional mandate in medicine—“first of all, do no harm.”

It seems possible to reconcile these goals. The first task is to assure that the diagnosis is not applied to nonill persons. In this regard, it is worth keeping in mind that most criteria for any DSM-IV diagnosis are based on behaviors on a continuum of human experience. Consider anxiety, or the experience of sadness, aggressiveness, or memory problems. The validity of psychiatric diagnosis rests heavily on the general criteria of distress, dysfunction, and/or disability. It is these attributes that prompt individuals or families to seek help. If DSM-V is to succeed with this class, it is crucial that the diagnosis be applied in the context of illness as defined by distress, dysfunction, and/or disability. Suspiciousness without impaired function or dysphoria will not qualify for the diagnosis.

The second aspect, do no harm, is also a valid concern. If this new diagnosis leads to therapeutic procedures in which cost and/or adverse effects outweigh benefit, harm results. Therefore, it will be important to provide effective treatment guidelines supported with scientific evidence. A presumption, eg, that all or most cases of prodrome should be placed on antipsychotic medication would not be evidence based and would cause harm. At present, clinical care approaches will be based on stress reduction while monitoring for symptomatic exacerbation. Studies now underway may provide more specific information on the effectiveness of psychosocial treatments, judicious use of medications, and integration of therapeutic technique.

The field will be challenged to manage this new class prudently. But, without the class, a framework for early detection and intervention is lost, and the opportunity to develop evidence-based treatment will be minimized. In my view, DSM-V must represent that which is scientifically valid even if concern for potential misuse is serious. The prodrome, as a class within the psychoses framework, may parallel attention deficit/hyperactivity disorder in potential for public misunderstanding and in the possibility of therapeutic approaches in some individuals where risks are greater than benefits. Establishing a clear boundary between ill and nonill and restricting therapy to the former cases and implementation of prudent, evidence-based care will be essential in meeting our clinical care obligations. It will also be the most effective way to combat the expected criticism that another category is being created to justify the marketing of drugs to nonill populations.

The issue of stigma must be addressed in DSM-V. Words have connotations, and naming things creates an initial orientation. Prodrome is not a good choice for the new class. It has a different meaning as defined in DSM-IV, and the “prodrome of what” question is not adequately answered. Of psychosis generally, yes, but not to a specific class. S. Heckers, MD (personal communication, 2009) has made the interesting suggestion that DSM-V create a new grouping that would comprise risk syndromes. “Risk syndrome for psychosis” is less problematic than “prodrome” that has a different definition in DSM-IV and may address aspects of stigma if shortened to “risk syndrome.” Such a grouping would allow space for other risk syndromes to be developed including risk dimensions that are not disease per se. Elevated blood pressure and abnormal glucose indices are risk syndromes with multiple possible disease outcomes. Assessment and treatment of these conditions are essential because of their inherent risk. The same logic holds true for the risk syndrome associated with psychosis. However, caution is necessary,11–13 and quasi-psychotic features are not uncommon in the nonill population.14

Concerning schizophrenia, we anticipate an increasing ability to detect early morbid manifestations (eg, impaired cognition or avolition) with endophenotypes or biomarkers that relate assessment to potential disease class. As therapies are developed, it will be imperative to determine if early intervention (perhaps years in advance of psychosis) is effective and favorably alters the future course of illness. The rubric of risk syndromes would facilitate evolution in the understanding of disease trajectories and refinement of early treatment. Furthermore, DSM-V may have a dynamic format in order to facilitate timely communication based on advances in information technology. An evidence-based approach to psychiatric diagnosis requires the integration of new information as it is established rather than waiting for the next DSM publication.