Telemental Health Treatment of Patients Diagnosed with Anxiety Disorders

Abstract

Background. A prominent rationale for telemental health (TMH) adoption is that TMH can help overcome treatment barriers; however, TMH adoption has progressed more slowly than expected and questions about its utility persist. In 2013, the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) the anxiety disorder category changed and excluded disorders that had been the focus of videoconferenced (VC) TMH study. VC TMH anxiety disorder treatment outcome assumptions, therefore, warranted reconsideration. Methods. The Knowledge to Action framework underpinned this retrospective, double cohort study of clinic-based individual treatment, measured by the Generalized Anxiety Disorder 7-item scale (GAD-7). Candidates were selected from military treatment-eligible patient records documenting DSM-5 anxiety disorders. Initial TMH encounters occurred 1/1/2015 - 12/31/2015 and had ≥ 2 follow-up encounters with GAD-7 scores prior to 1/1/2017. Treatment episodes were delivered via either VC TMH or in-person treatment as usual (TAU) conditions. Treatment encounters (n = 854) for both TMH (n = 612) and TAU (n = 233) candidates were engaged by eight independent mental health professionals assigned to a single clinic. Candidates were randomized and, while maintaining 15.63% female representation, 32 subjects per platform group were matched (n = 64) on age group, sex, race, ethnicity and service branch. Of 490 visits, a total of 391 GAD-7 scores for TMH (v = 230, 58.82%) and TAU (v = 161, 41.18%) cohorts were evaluated. The average TMH patient distance from the consulting provider was 1,377 (SD = 314) miles. Multiple linear regression was used to examine outcome predictors and a linear mixed effects model was fitted to examine estimated group means by visit. Results. Unadjusted GAD-7 change scores improved by -2.77 (SD = 5.25) points (t(63) = 6.50, p < .001). In the TMH and TAU conditions, GAD-7 scores improved by -3.06 (SD = 5.45) points (t(31) = 4.21, p < .001) and -2.47 (SD = 5.11) points (t(31) = 4.95, p < .001), respectively. In the medication and psychotherapy treatment groups, initial-to-last GAD-7 scores significantly improved by -2.43 (SD = 5.05) and -3.22 (SD = 5.58) points respectively (ps = .01). These improvements were not significantly different between treatment platform or between therapy type conditions. Multiple linear regression showed that initial GAD-7 scores uniquely predicted outcome scores, b = 3.97, SE = 0.65, t(58) = 6.08, p < .001. Subjects who were not married outperformed married subjects by 3.78 points, b = 1.89, SE = .66, t(58) = 2.87, p < .01, and there was a significant interaction between marital status and initial GAD-7 scores on outcomes, b = 1.41, SE = .66, t(58) = 2.14, p < .05. Linear mixed effects regression showed that the fixed main effects of marital status (F(1,52.07) = 5.23, p < .05) and visit measure changes (F(4,176.42) = 5.82, p < .001) over the first five visits were significant. Again, no significant overall platform-therapy type change rate differences were found; however, two within group outcome improvements were significant. On average, at the fifth visit, TAU subjects who were not married improved by -10.14 (SE = 2.17) points and the TMH married subject group improved by -3.26 (SE = 1.08) points (ps < .001). Conclusion. The lack of differences between treatment-platforms in symptom improvement rates supported continuation of outcome assumptions based on a previously held conceptualization of anxiety disorders. GAD-7 symptoms improved significantly among treatment platform cohorts with a large effect (f2 = .60) regardless of the therapy type engaged. When added to the full model, marital status predicted an outcome advantage for patients who were not married with a significant small-to-medium effect (f2 = .14) that remained significant when initial GAD-7 scores were higher (+1 SD) than average (f2 = .07). Over the first five visits, GAD-7 outcomes for TAU subjects who were not married (p < .001) and married TMH subjects improved significantly (p < .05). Fixed effects of visit measure change (f2 = .33) and marital status (f2 = .29) on GAD-7 outcomes had medium-large local effect sizes. These findings support VC TMH intervention improvements such as referral decision guidance and increased efficiency of behavioral health resource allocation. Tailoring treatment in these ways may improve access to care in locations where resources are limited and thereby advance treatment options to improve outcomes for military service members and their families.