Catching the Right Fish

Designers of descriptive diagnostic criteria for mental disorders face some of the same problems as fishermen. Fishermen, like nosologists, want to capture not just any fish but a particular kind. Fishermen deal with this problem in various ways. One is finding where to look for their kind of fish—they don’t look for tuna in the neighborhood pond. Another is selecting a fishing line to break with the ones that are too big. Some fishermen choose nets with a particular mesh, to allow the wrong critters to escape and at least the (mostly) right ones to be captured. Still others tailor their trap to the mechanics and habits of the fish—think of lobster traps.

Similarly, DSM nosologists write their diagnostic criteria with descriptions of what kinds of clinical features to include—inclusion criteria—and which ones to exclude—exclusion criteria. Inclusion criteria are intended to capture the positive features of the disorder, and exclusion criteria to exclude the clinical features that are unnecessary or confounding. Taken together, inclusion and exclusion criteria make the “net” of diagnosis more selective—we get a kettle of disorders that truly resemble each other.

The DSM architects since DSM-III have embraced a “polythetic” approach to diagnostic inclusion criteria. This means that multiple positive features are listed, but not all are required to make the diagnosis. Hence a DSM-IV-TR Major Depressive Episode can be diagnosed with 5 out of 9. A criteria met. Other diagnostic criteria exclude confounds: not mixed bipolar, not due to another disorder or condition. The practical impact of inclusion and exclusion diagnostic criteria is that they narrow the population of people diagnosed with a given disorder. The polythetic “Chinese menu” approach allows for a substantial amount of individual variation of disorder expression while still capturing individuals who share common clinical features, and hopefully, the same disorder.

The Holy Grail for descriptive criteria design is for the resulting disorder to be useful in predicting the course of illness, responses to treatment, etiological features, etc. If the diagnostic criteria cast too wide a net, false-positive diagnoses occur, and if the mesh is not right, false-negatives occur. Both consequences enfeeble the potential for treatment, validation, and etiological studies to detect the signals that deepen our understanding and treatment of the disorder. One way to enhance the potential for detecting useful signals is to have a rich and detailed set of diagnostic criteria. By increasing the number of good inclusion and exclusion criteria, we reduce the confounding and irrelevant conditions that are captured.

Unfortunately, the past iterations of the DSM have not consistently drawn the diagnostic criteria with rich and detailed criteria sets. For classical mental disorders like Schizophrenia, Bipolar Disorder, and Dementia of the Alzheimer’s Type, the descriptive criteria have multiple and detailed descriptions of the clinical phenomenology and carefully composed exclusionary criteria. However, the example set by these criteria sets are not always followed elsewhere, and many disorders have impoverished criteria sets, with single symptoms or symptom clusters and few exclusionary criteria.¹ Good examples of DSM-IV-TR disorders with impoverished criteria sets include all the Paraphilias and the Impulse Control Disorders. (Interestingly, these are also “vice-laden” disorders as discussed in my earlier “Vice Squad” piece.) It should be no surprise that these disorder concepts suffer from rampant comorbidities and crossed directions for etiological formulation.

To their credit, the Sexual and Gender Identity Disorders Work Group for DSM-5 appears to be taking the problem of impoverished criteria sets seriously, by adding new clinical descriptors and considering the use of laboratory measures.² Emil Coccaro and colleagues, studying Intermittent Explosive Disorder, have offered an enriched set of research criteria that have already generated a rich and multidimensional set of validation studies.³ Curiously, Coccaro’s contributions have not yet found their way into the DSM-5 draft criteria online. We should hope all DSM-5 Work Groups take the problem of impoverished criteria sets as seriously as these authors have.

References
1. Sadler JZ. Vice and the diagnostic classification of mental disorders: A philosophical case conference. Philos Psychiatr Psychol. 2008;15:1-17.
2. Zucker KJ. Reports from the DSM-V Work Group on Sexual and Gender Identity Disorders. Arch Sex Behav. 2010;39:217-220.
3. Coccaro EF. Intermittent Explosive Disorder. Curr Psychiatry Rep. 2000;2:67-71.