Letter to the editor of the Journal of Affective Disorders: Supporting the guideline's quest for real direct costs of depression care

Health economic evaluation of an internet intervention for depression (deprexis), a randomized controlled trial

BACKGROUND

Depressive disorders often remain undiagnosed or are treated inadequately. Online-based programs may reduce the present treatment gap for depressive disorders and reduce disease-related costs. This study aimed to examine the potential of the internet intervention "deprexis" to reduce the total costs of statutory health insurance. Changes in depression severity, health-related quality of life and impairment in functioning were also examined.

METHOD

A total of 3805 participants with, at minimum, mild depressive symptoms were randomized to either a 12-week online intervention (deprexis) or a control condition. The primary outcome measure was statutory health insurance costs, estimated using health insurers' administrative data. Secondary outcomes were: depression severity, health-related quality of life, and impairment in functioning; assessed on patient's self-report at baseline, post-treatment, and three-months' and nine-months' follow-up.

RESULTS

In both groups, total costs of statutory health insurance decreased during the study period, but changes from baseline differed significantly. In the intervention group total costs decreased by 32% from 3139€ per year at baseline to 2119€ in the study year (vs. a mean reduction in total costs of 13% in the control group). In comparison to the control group, the intervention group also showed a significantly greater reduction in depression severity, and impairment in functioning and a significantly greater increase in health-related quality of life.

CONCLUSION

The study underlines the potential of innovative internet intervention programs in treating depressive disorders. The results suggest that the use of deprexis over a period of 12 weeks leads to a significant improvement in symptoms with a simultaneous reduction in the costs of statutory health insurance.

Longterm persistence and nonrecurrence of depression treatment in Germany: a four-year retrospective follow-up using linked claims data

OBJECTIVES

To measure persistence and nonrecurrence of depression treatment and investigate potential risk factors.

METHODS

We retrospectively observed a closed cohort of insurees with new-onset depression treatment in 2007 and without most psychiatric comorbidity for 16 quarters (plus one to ascertain discontinuation). We linked inpatient/outpatient/drug-data per person and quarter. Person-quarters containing specified depression services were classified as depression-treatment-person-quarters (DTPQ). We defined longterm-DTPQ-persistence as 16 + 1 continuous DTPQ and longterm-DTPQ-nonrecurrence as 12 continuous quarters without DTPQ and used multivariate logistic regression to explore associations with these outcomes.

RESULTS

Within first 16 quarters, 28,348 patients' first period (total time) persisted for a mean/median 5.4/3 (8.7/8) quarters. Fourteen percent had longterm-DTPQ-persistence, associated (p < .05) with baseline hospital (odds ratio, OR = 1.80), psychotherapy/specialist-interview and antidepressants (OR = 1.81), age (years, OR = 1.03), unemployment (OR = 1.21), retirement (OR = 1.31), and insured as a dependent (OR = 1.32). Thirty-four percent had longterm-DTPQ-nonrecurrence, associated with psychotherapy/specialist-interview (OR = 1.40), antidepressants (OR = 0.54), female sex (OR = 0.84), age (years, OR = 0.99), retirement (OR = 1.18), and insured as a dependent (OR = 0.88). Women differed for episodic and not chronic treatment.

CONCLUSION

Treatment measures compared to survey's symptoms measures. We suggest further research on "treatment-free-time." Antidepressants(-) and psychotherapy/specialist-interview(+) were significantly associated with longterm-DTPQ-nonrecurrence. This was presumably moderated by possible short-time/low-dosage antidepressants use(-) and selective therapy assignment(+). Sample selectivity limited data misclassification.