Evaluating the e-PASS: the psychometric properties and user experience of an online clinical diagnostic assessment program for mental disorders

Diagnostic assessment is critical for identifying and treating mental disorders. However, consumers face numerous barriers to effective diagnostic assessment in traditional clinical practice, including: stigma and poor mental health literacy; high costs and shortage of clinicians; and inadequate assessment resources and training for clinicians. With recent internet usage growth, web-based diagnostic assessment programs offer a means of overcoming these barriers while also providing benefits such as added accuracy and efficiency through automated item and scoring administration. However, they also raise issues which emphasise the need to ensure such programs are effective. At present, various web-based programs are publically available, although most present limited diagnostic functionality (e.g., narrow scope of disorders, limited feedback) and lack supporting research. This prompted development of the e-PASS, a freely available web-based clinical assessment program providing diagnostic and referral feedback for 21 disorders. To inform its appropriate use and address the dearth of empirical literature regarding similar applications, this thesis reports an evaluation of the e-PASS's performance in real-world circumstances. Participants consisted of Australian adults (N = 616, M age = 38.1 years) recruited online and representing actual e-PASS users. In a psychometric study, participant subsets underwent: a structured clinical interview (n = 158); standardised clinical questionnaires (n = 173); and e-PASS re-testing (n = 39). With the clinical interview as the gold standard, the e-PASS's diagnostic results displayed: strong agreement for some disorders (e.g. panic disorder) but not others (e.g. social phobia, OCD); high specificity, and varying sensitivity; similar or better accuracy compared to certain clinical questionnaires; and limited accuracy in primary diagnosis. Test-retest reliability was found to be high, while logistic regression indicated that reducing the diagnostic threshold and adding particular items could improve e-PASS validity. Two studies evaluated the e-PASS's user experience, with participants (recruited from the psychometric study) completing an online survey (n = 88) or in-depth qualitative interviewing (n = 15). Results showed high e-PASS acceptability, including motivation to re-take the e-PASS for future needs. The e-PASS was widely praised for being more comprehensive and less-judgmental than previous assessment experiences, promoting accessibility, anonymity, self-reflection, and disclosure. Group differences in experiences, as well as criticisms regarding length, restrictive response options, and lack of support in some cases highlighted the need to tailor online assessment to optimise outcomes. In conclusion, the e-PASS clearly demonstrates the potential to enhance traditional clinical assessment practice, and improve consumer mental health awareness and access to appropriate services.