Document Type


Date of Degree


Degree Name

PhD (Doctor of Philosophy)

Degree In


First Advisor

Carney Doebbeling, Caroline

Second Advisor

Torner, James


Depression is common and disproportionately affects those with chronic medical comorbidity, such as diabetes mellitus (DM). Only a limited amount of information is available concerning the quality of guideline-concordant treatment of depression and its influence on glycemic control among those with DM. This is the first study to address these issues in a veteran population with DM.

This is a retrospective cohort study (1997-2005) of veterans with and without DM from the Roudebush Veterans Affairs Medical Center in Indianapolis. Veterans with and without DM and a new episode of depression were identified. Administrative, clinical, and pharmacy data were linked to assess initiation of treatment, follow-up care, antidepressant dosage and duration, and change in antidepressant agent based on the Veterans Health Administration (VHA) clinical practice guidelines for depression. HbA1c levels were assessed following initiation of antidepressant therapy.

Treatment of depression was not consistent with guideline recommendations. Only 60% of subjects received treatment within 30-days of the depression diagnosis. Veterans with DM were more likely to have received treatment within the first two weeks than veterans without DM. Few subjects received appropriate follow-up care for depression (<40%) or an adequate duration of antidepressant therapy (<9%), although most (88%) received a dosage consistent with guideline recommendations when treatment was provided. Most subjects (>75%) were treated with a serotonigenic agent and only 23% experienced a change in therapy during the treatment period, almost 84% of which received an adequate trial of therapy prior to the change or augmentation in agent. Presence of DM was associated with significantly increased odds for receipt of guideline-concordant care for depression in most multivariate analyses. Receipt of guideline-concordant care for depression was not a significant predictor of glycemic control but was associated with a clinically meaningful reduction of 0.5% in HbA1c levels.

This research demonstrates that under-treatment of depression is common and may influence at least one medical outcome. Findings also support that the relationship between depression and DM is complex and that further research is necessary to help align current practice with evidence-based practices in the VHA.


depression, diabetes mellitus, quality of care, clinical practice guidelines, epidemiology, Veterans Health Administration


xix, 330 pages


Includes bibliographical references (pages 303-330).


Copyright 2006 Laura Elizabeth Jones

Included in

Epidemiology Commons