gms | German Medical Science

15. Deutscher Kongress für Versorgungsforschung

Deutsches Netzwerk Versorgungsforschung e. V.

5. - 7. Oktober 2016, Berlin

Long-term persistence of depression treatment in Germany and associated factors – A four year cohort study using claims data

Meeting Abstract

  • Christoph Wagner - Uniklinik, Institut für Gesundheitsökonomie und klinische Epidemiologie, Köln, Deutschland
  • Charalabos-Markos Dintsios - Uniklinik, Institut für Versorgungsforschung und Gesundheitsökonomie, Düsseldorf, Deutschland
  • Florian G. Metzger - Uniklinik, Abteilung für Psychiatrie und Psychotherapie und geriatrisches Zentrum, Tübingen, Deutschland
  • Helmut L'hoest - BARMER GEK, Wuppertal, Deutschland
  • Ursula Marschall - BARMER GEK, Wuppertal, Deutschland
  • Björn Stollenwerk - Helmholz Zentrum München, Institut für Gesundheitsökonomie und Management im Gesundheitswesen, Neuherberg, Deutschland
  • Stephanie Stock - Uniklinik, Institut für Gesundheitsökonomie und klinische Epidemiologie, Köln, Deutschland

15. Deutscher Kongress für Versorgungsforschung. Berlin, 05.-07.10.2016. Düsseldorf: German Medical Science GMS Publishing House; 2016. DocP148

doi: 10.3205/16dkvf208, urn:nbn:de:0183-16dkvf2082

Published: September 28, 2016

© 2016 Wagner et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. See license information at



Background: The number of surveys studying time to recovery from depression disease beyond two years of follow-up is limited. Surveys are at risk for recall-, response-, and attrition bias. Most observational studies analyse long-term persistence of pharmacological treatment and a subgroup of about 45% of patients with depression taking antidepressants. Tracing inpatient and outpatient and drug treatment using linked micro-level claims data would allow to (i) follow a more comprehensive population suffering depression, (ii) possibly overcome some survey limitations, and (iii) explore associations with real world baseline treatment choice.

Objectives: We aimed to measure long-term persistence of depression treatment-person-quarters (DTPQ) and to explore baseline patient characteristics and treatment groups associated with it.

Methods: We analysed a population-based closed cohort of German residents fulfilling a case definition of new-onset diagnosed depression in 2007 for 16+1 quarters beginning with the first quarter of depression treatment. We excluded psychiatric comorbidity to avoid measurement bias and concurrent risk. We used 2006-2011 enrolment/outpatient/inpatient/drug data of the “BARMER GEK” statutory health insurance (SHI) fund linked per person and quarter. We built categories/combinations of inpatient/outpatient/drug services related to depression and thereof classified DTPQ. DTPQ were validated by multiple provider treatment or at least one confirmed outpatient diagnosis and at least the provision of one service related to depression.

We defined long-term-DTPQ-persistence as 16+1 consecutive DTPQ, and long-term-DTPQ-absence as ≥12 consecutive quarters without depression treatment following a closed first period within ≤4 DTPQ (subgroup). We predicted both outcomes using logistic regression. We defined and disjunctively assigned five groups of baseline treatment according to maximum service utilisation in the first two quarters: 1 "hospital", 2 "psychotherapy services and antidepressants", 3 "psychotherapy services", 4 "antidepressants", 5 "physician". Because of partial unavailability of historical data for insured of the former “GEK” SHI fund, we limited prediction to patients insured by the former “BARMER” SHI fund before the year 2010 “BARMER GEK” merger.

Results: Totally 28,348 patient’s first observed period persisted for mean (Percentile 25/50/75) 5.44 (2/3/8) DTPQ, and they had totally 8.68 (3/8/14) DTPQ within the first 16 quarters. Significant (p<.05) baseline predictors of long-term-DTPQ-persistence (former “BARMER” SHI fund, N=23,229, 14% affected) were the treatment groups “Hospital” (OR=1.797) and “Psychotherapy services AND Antidepressants” (OR=1.814), age/year (OR=1.033), East German states (OR=1.176), low spatial psychotherapist density (OR=0.844), unemployment (OR=1.208), retirement (OR=1.311), insured as a dependent (OR=1.323), levels of diagnosed depression severity, diagnosed recurrent depression (OR 1.701). Significant baseline predictors of long-term-DTPQ-absence (former “BARMER” SHI fund, subgroup, N=14,656, 34% affected) were the treatment groups “Psychotherapy services” (OR=1.398) and “Antidepressants” (OR=0.541), unemployment (OR=1.175), retirement (OR=1.175), insured as a dependent (OR=0.879), levels of diagnosed depression severity, diagnosed recurrent depression (OR=0.717).

Discussion: 14% with DTPQ-persistence beyond four years confirmed the sole hitherto identified survey’s result analysing four years of depression symptoms [1]. Found first period’s persistence of DTPQ exceeded the duration of first depression episode reported in surveys mostly because of a rigid case definition and other survey’s limited time horizon and cut off distribution.

Missing psychiatric outpatient clinic data possibly induced DTPQ underestimation. However, analysed psychiatric mono-morbid patients normally don’t qualify for psychiatric outpatient clinics. Population selectivity by rigid case definition, exclusion of comorbidity, and continuous insurance limited external validity of results, but were chosen to facilitate a methodologically valid DTPQ measurement.

Regression analysis was subject to prescription “bias” and overt bias by means of e.g. missing information on education. However, “choice” of psychotherapy might favour long-term-DTPQ-absence and long term cost savings. Antidepressant studies’ time horizon possibly not suffices.

Practical implications: Despite methodologically necessary patient selectivity and still possible misclassification bias for outpatient treatment files, we suggest further development of proposed method for a raw population-level measure of “treatment-free-time” to cheaply and tightly trace large populations.


Keller MB, Lavori PW, Mueller TI, Endicott J, Coryell W, Hirschfeld RM, Shea T. Time to recovery, chronicity, and levels of psychopathology in major depression. A 5-year prospective follow-up of 431 subjects. Arch Gen Psychiatry. 1992 Oct;49(10):809-16.