Confirmatory bias in the diagnosis of ADHD : an exploratory study and survey of New Zealand clinicians' protocols & practices in the diagnosis of attention-deficit hyperactivity disorder : a thesis presented in partial fulfilment of the requirements for the degree of Master of Arts in Psychology at Massey University, Palmerston North, New Zealand

Abstract

Attention-deficit hyperactivity disorder (ADHD) is a heterogeneous syndrome of childhood, with primary symptoms of inattention, hyperactivity, and impulsivity. In recent years, numbers of children diagnosed with ADHD have increased. While many factors may be associated with this increase, one possibility is increased false positive diagnoses due to confirmatory bias (CB) in the diagnostician. CB occurs when a clinician pays attention to positive symptoms with disregard of disconfirmatory symptoms. The present study used a quasi-experimental approach to investigate whether CB was present in the diagnosis of ADHD. Diagnostic decision making was examined in three hypothetical case studies where the ratio of positive to negative ADHD symptoms varied. Results demonstrated CB in the diagnosis of ADHD for many participants. Forty-three percent of clinicians gave no indication of considering disconfirmatory symptoms. Additionally, for all symptoms but one, more attention was paid when they were positive rather than negative. Gaining knowledge from psychological literature and completing an internship increased the likelihood of considering disconfirmatory data. CB was related to clinicians' realworld belief of ADHD prevalence, although this was limited to a statistical trend. The majority of clinicians gave a tentative ADHD diagnosis for all case studies. For clear (i.e. not tentative) diagnoses, clinicians who demonstrated CB were significantly more likely to give a positive diagnosis than a negative diagnosis, whether or not this diagnosis was correct. Results suggest possible misdiagnosis of ADHD in some cases, with concerns of this study being support for the potential of overdiagnosis as a function of CB. Some additional hints of underdiagnosis by a few clinicians merit further research, with the phenomenon of a possible disconfirmatory bias raised and discussed. ln addition, clinicians were surveyed regarding ADHD assessment and treatment in actual practice. Clinicians indicated using an average of 7 assessment steps, with school information, parent or family interview, and rating scales being the most popular tools. Clinicians who took disconfirmatory data into account used more assessment steps in actual practice than the CB group. There was a mean of 4 treatment options listed, with the most utilised being medication and behavioural treatment. Findings are limited by the survey-based, correlational nature of the study. The ability to generalise findings to actual practice is considered and discussed.