Role of Narrowband Symptom Measures in Diagnostic Assessment of Adult ADHD

Background and Purpose: In response to heightened awareness of psychostimulant abuse among college students, university medical centers and primary care offices are becoming increasingly cautious in their diagnostic and treatment decisions for adults with suspected ADHD. A certain level of caution is appropriate when treatment might include the prescription of controlled substances however, a lack of standardized evidence-based protocols for accurately diagnosing Attention Deficit Hyperactivity Disorder (ADHD) in adults leads to the creation of arbitrary standards and practices that can present unnecessary barriers to diagnosis and appropriate treatment for adults who do have ADHD.

The purpose of the current study is to compare diagnostic outcomes across several commonly used narrow-band measures of adult ADHD symptoms to the DSM-V criteria for adult ADHD. Agreement in outcomes on these measures with the DSM-V criteria would support the use of such measures as a valid and time-efficient way for clinicians and prescribers to render accurate ADHD diagnoses in adult populations.

Methods: Eighty adults were referred for and completed an ADHD evaluation at an outpatient mental health agency located across the street from a large urban university. Participants all completed a battery of assessment measures (Conners Adult ADHD Rating Scale Self-Report - CAARS, CAARS Observer Report, Wender Utah Rating Scale - WURS, WURS Observer Report, Barkley Deficits in Executive Functioning - BDEFS, and BDEFS Observer Report) and completed three one-hour clinical interview sessions with a licensed mental health clinician.  

Agreement between categorical diagnostic outcomes for the narrowband measures of ADHD (CAARS Self- and Observer Reports, WURS Self- and Observer Reports, and BDEFS Self- and Observer Reports) and DSM-V Diagnostic criteria were tested with Cohen’s kappa statistic. Agreement in continuous scores among the different narrowband measures of ADHD were tested with intraclass correlations.

Results: All Kappa coefficients for agreement among narrowband self-report instruments and DSM-V criteria were non-significant except for the CAARS Self-Report which showed very low agreement (kappa = .270, p=.001). Kappa coefficients for the relationship between Observer Report Instruments and DSM-V criteria showed better agreement with two of the measures reaching statistically significant agreement. Finally, intraclass correlation coefficients for measures yielding continuous scores had higher levels of agreement with DSM-V criteria however, none were above the standard cut-off of .7 for adequate agreement.

Conclusions and Implications: Although some of the relationships between narrowband measures of ADHD and DSM-V criteria showed statistically significant levels of agreement, none were above standard cutoffs for strong agreement. Results from this study should inform diagnostic best practices and urge clinicians and prescribers to resist temptation to trade comprehensive diagnostic interviews for seemingly time-efficient measures that have been validated mostly in research settings.