Two opinions about one child--what's the clinician to do?

How does a clinician proceed when she or he gets conflicting information from parent, child, and school? The special section in this issue entitled, ‘‘Information Variance and Its Diagnostic Implications,’’ tackles a conundrum that confounds evaluations in youth (Achenbach, et al. 1987, Youngstrom, et al. 2003) and adults (Achenbach, et al. 2005) alike. After years of wrestling with the reality that any two information sources correlate only moderately with each other, and often disagree, there has been a concerted effort to move beyond solutions like counting symptoms only if there is agreement, or counting everything for lack of a better plan (De Los Reyes Kazdin 2005). The five articles in this section address basically two different but related aspects of informant relationship—that between parent and teacher and that between parent and offspring. Studies by Rettew et al. and Carlson and Blader examine parent/ teacher agreement and disagreement. The former uses a community sample in the Netherlands (N= 1,730 children, mean age 11 years) and asks how often parents report clinical levels of problems when teachers do not and vice versa. Further analyzed is whether parent personality, parental stress, number of children in the household, child gender, socio-economic status (SES), intelligence quotient (IQ), and an aspect of child temperament called ‘‘effortful control’’ (the ability to inhibit something you want to do in favor of something you should do) contribute to our understanding of the agreement versus site specific differences. The authors found that disagreement between informants was more the rule than the exception, but with only a few of these variables distinguishing between symptoms reported only at home or only at school. Using a multivariate analysis, they concluded that parental stress predicts greater symptomatology in all the symptom domains studied (i.e., aggression, rule breaking behavior, attention symptoms, and attention-deficit/hyperactivity disorder (ADHD) behaviors). They also reported that effortful control (or its absence) was associated more with parent-reported (with or without teacher agreement) symptoms than teacher reported symptoms. Gender played a somewhat different role depending on the type of behavior studied. Female gender was most predictive of inattentive and hyperactive symptoms reported only at home, while for rule breaking behavior, the home specific group was primarily boys. Children with crosssituational problems had lower IQ and lower SES along with greater parental stress and poorer effortful control, though for statistical reasons, these variables did not contribute to the final model. The Carlson and Blader study uses a large outpatient clinic sample (N= 911, mean age about 12 years), rather than a community sample, to examine parent/teacher agreement and disagreement specifically on manic symptoms. They found the usual modest correlation between parent and teacher scores to be r= 0.27 ( p< 0.000), this time on the Child Mania Rating Scale (Pavuluri et al., 2006). This is similar to what others have found between parent and teacher ratings using other scales. The article, though, is more focused on the question of the diagnostic implications of informant agreement and disagreement about manic symptoms. Using ‡ 75%ile of scores on the Child Mania Rating Scale (Pavuluri et al., 2006) to determine most symptomatic and £ 25%ile for the least symptomatic children, the study hypothesizes that bipolar disorder, where it occurs, will be most associated with cross situational endorsement of high mania symptom scores. The study also examines what diagnoses are most prominent where there is parent/teacher agreement, and where a parent only reports symptoms but the child is completely asymptomatic in school. In this report, logistic regression revealed a 10-fold greater odds of an externalizing disorder (ADHD with or without oppositional defiant disorder/compulsive disorder being present) where there was high parent/teacher concordance on the presence of manic symptoms. Although bipolar disorder did occur more often where the child was significantly symptomatic at home and at school, the overall low base rate of bipolar disorder made it a less useful explanatory diagnosis for parent/teacher concordance. Perhaps more importantly, in situations where parents only reported significant manic symptoms, and the child was fine in school both academically and behaviorally, anxiety disorders, and not ultradian cycling bipolar disorder was the diagnosis observed. The next three articles focus on parents and youths seeking services from a large outpatient mental health sample. Keep in mind that the question of what to do with disparate information between parents and others has usually been solved by combining positive symptoms from both informants. The article by Youngstrom et al. instead examines interviewer ratings of the ‘‘credibility’’ of the caregiver or youth to decide how to combine information. At the end of the diagnostic interview, the interviewer noted whether the credibility of information seemed ‘‘good,’’ ‘‘fair’’ or ‘‘poor’’ for the caregiver and for the youth. Credibility is a clinical decision about how much to credit or discount the information provided by an individual and differs from reliability in that reporters can be consistent in their information but not necessarily credible. Youngstrom and colleagues found that informants with ‘‘good’’ or ‘‘fair’’ credibility groups tended to perform similarly, with a more pronounced drop in validity for the ‘‘poor’’ group measured against criteria such as consensus diagnoses (based on all information) or observational ratings. What makes a person more credible during a clinical interview? On average, caregiver (i.e., the adult most involved in the child’s life, regardless of who it was) credibility was higher in better functioning families with higher levels of education, and when youths were younger versus adolescent. Youth credibility was strongly connected to increasing age, cognitive ability, caregiver credibility, and