The DSM-5 Field Trials: What They Mean for Child Psychiatry

January 1, 2013

Editor-in-Chief, Caroline Fisher, PhD, MD

The DSM-5 Field Trials: What They Mean for Child Psychiatry

in advance of the official release of DSM-5 this May, the field trial data has been reported in three papers in the American Journal of Psychiatry. The first is a methods paper, the second reports on the reliability of the diagnostic cri- teria, and the third on the reliability of patient (and parent) cross diagnostic symptoms. The three papers have been summarized here together, as they func- tion more as a unit than as three separate pieces. The clinical trials involved both adult and pediatric diagnoses, but only the pediatric information is reported in this article.

How Were the Trials Designed?

The trials were conducted at several large academic sites, where a sample of 689 children were evaluated at least once (98% of them were evaluated twice). symptom questionnaires were filled out in the waiting room by patients overthe age of 10 and their parents. These questionnaires were a broad psychiatric review of symptoms, including depres- sion, anxiety, anger, inattention, mania, somatic symptoms, substance use, dis- sociative symptoms, repetitive thoughts/ behaviors, psychosis, suicidality, sleep/ wake problems, and cognitive/memory problems. if, based on this questionnaire, a child met a threshold for a potential diagnosis, additional severity question- naires were added. The patient was then seen by a clinician who had access to the completed questionnaires. The clinician reviewed the questionnaires and com- pleted a dsM diagnostic checklist, a con- dition-specific severity checklist, several procedural questionnaires, and, some- where in the midst of that, addressed the needs of the patient. at the next visit, the patient saw a second clinician who fol- lowed the same procedure. The first cli- nician was the “test” and the second the “retest.”

How Do These Trials Compare to Previous DSM Trials?

The focus of the DSM-5 field trials shows how the manual has evolved over the years. in the DSM-III, the focus was on inter-rater reliability of the diagnoses: in other words, can two people cometo the same conclusion about a patient using DSM-III criteria? however, the sam- pling of patients was random, so rare diagnoses often were poorly represent- ed in the mix. in the DSM-IV, the focus was on prevalence changes: if the diag- nostic criteria were changed, would that alter the number of people who qualified for the diagnosis? To do that, select clin- ical populations and structured research interviews were used. The authors compare the first strategy to Nih effec- tiveness trials (what happens in a natural setting) and the second to pharma-fund- ed efficacy trials (what happens in a diag- nostically pure sample). For the DSM-5, the approach returned to the natural setting, with some statistical strategies, like stratified sampling, to make the rarer diagnoses frequent enough for evalua- tion. The clinicians were no more trained in DSM-5 criteria than any clinician would be.

What Did We Learn?

of the tested diagnoses of children and adolescents, many did not reach sta- tistical relevance, so no conclusions were made. adhd and autism both reached the level of “very good” inter-rater reli- ability. avoidant/restrictive food intake disorder, a new diagnosis, and opposi- tional defiant disorder both reached the “good” level. Major depressive disorder and disruptive mood dysregulation disor- der, another new diagnosis, only made it to the “questionable” range, while mixed anxiety-depressive disorder and non-sui- cidal self injury disorder were considered “unacceptable.” Bipolar I and II, PTSD, and conduct disorder, with or without the additional callous/unemotional speci- fier, failed to reach sufficient sample sizes to come to any conclusions.

differences between sites suggest some additional conclusions. For example, there was good inter-rater reliability for disruptive mood dysregulation disorder on inpatient sites, but not for outpatient sites. perhaps there is a certain faddishness to diagnoses, making disruptive mood dysregulation disorder a handy description for the general chaos that

presents in an inpatient unit. or perhaps the severity of children presenting for inpatient level of care makes the subtle- ties of the diagnosis more evident. The truth remains to be seen.

The multisymptom questionnaire that each patient over 10 and each parent filled out prior to meeting with the cli- nician was not to aid in diagnosis butto address the problem that categorical diagnoses in the dsM are hobbled by the frequent comorbidities that occur. each child patient was given a 25-question symptom inventory with a 5-point likert scale (scored 0 to 4) and their parents or guardians were given something similar. if patients screened in with a score of2 or more, they were asked to fill out a second, more symptom-specific question- naire to assess severity.

The results were as follows: first, kids are pretty reliable reporters in most cases, except about mania, anxiety, hal- lucinations, and somatic distress. parents are also pretty reliable reporters, except about legal drug misuse. parents do a better job than kids reporting psychot-ic symptoms, somatic distress, and sleep. children do a better job of reporting sui- cidality. The most common symptoms that cut across disorders in children were sleep problems, attention problems, and irritability/anger. That comes as no partic- ular surprise to most readers but it is val- idating nevertheless.

CCPR’s Take: These three papers shed some light on the much-contest-ed, but little-understood, dsM process. added to that were some unexpected plot twists on diagnosis, and a little vali- dation that we should listen to kids and to their parents, and then do it again. From a research perspective, it’s hard to argue that you really have nailed downa diagnosis when patients have so many additional symptoms, yet any of us that practice for more than a minute and a half recognize that there are precious few patients that fit into a single category. Perhaps the reason we don’t all agree on a child’s diagnosis is that the common- alities overshadow the specifics at some times, and the specifics become more prominent at others.

Child Psychiatry