(PS8-83) Predictors of engagement with an internet-based CBT program implemented in routine care

Abstract Text: Internet-based cognitive behavioral therapies (iCBTs) are an efficacious and accessible treatment option for a variety of emotional concerns. Unfortunately, in the context of real-world implementations, studies suggest that only 20-50% of patients engage in iCBTs at the recommended frequency and/or complete the full course of treatment. The aims of these analyses were (a) to characterize program engagement and (b) to evaluate the relationship between level of program engagement and both demographic variables and baseline symptom severity. Analyses included all adult patients in a US-based healthcare system referred by their health care team to an iCBT program between 6/1/20-1/26/22 for symptoms of depression and/or anxiety. The program was an adapted version of an evidence-based iCBT program developed at University of New South Wales and included 6 CBT-informed modules, along with homework worksheets and weekly coach contact. Engagement was measured using three metrics: (1) starters: number of patients who used their assigned course once within 1 month of referral; (2) completers: among those who started their assigned course, number who completed all 6 modules; (3) number of modules completed (among starters only). Patients referred were majority female (73.06%) and non-Hispanic white (78.74%), with an average age of 43.69 years (SD = 15.22). N = 1777 (76.46%) were starters, N = 705 (39.67%) were completers, and the average number of modules completed was 3.54 (SD = 2.31). Baseline symptom severity was only recorded for patients that started their assigned course. Among these patients, mean PHQ-9 = 8.85 (SD = 5.52) and mean GAD-7 = 9.53 (SD = 4.85). Across legal sex, chi-squares indicted no significant differences in proportion of starters (χ² = .845, p = .358) or completers (χ² = .170, p = .680) and ANOVA indicated no difference in modules completed (F = 1.44, p = .230). Across minority status, chi-squares indicated that minority patients (non-white and/or Hispanic) were significantly less likely to be starters (χ² = 11.16, p = 0.004) or completers (χ² = 14.26, p < 0.001). Additionally, ANOVA indicated that minority patients completed significantly fewer modules than non-minority patients (F = 10.78, p < .001). With regard to age, ANOVA indicated that both patients who started (F = 4.98, p = .026) and completed (F = 50.20, p < .001) their assigned course were significantly older than those who did not. Similarly, linear regression indicated a significant positive relationship between age and modules completed (ß = .16, p < .001). With regard to baseline symptom severity, ANOVA indicated that completers had significantly lower baseline depression (F = 14.24, p < .001) and anxiety (F = 3.90, p = .048) than non-completers. Similarly, linear regression indicated a significant negative relationship between modules completed and both baseline depression (ß = -.093, p < .001) and anxiety (ß = -.056, p = .018). Findings contribute further evidence suggesting that low rates of iCBT engagement occur in routine care settings. Additionally, results indicate that minority groups, younger patients, and those with more severe baseline symptoms are at greater risk for poor engagement. Future research on how to enhance engagement and outcomes is recommended.