January 2010
Reviewer: Jeff C. Huffman, MD
Diabetes and poor disease control: is comorbid depression associated with poor medication adherence or lack of treatment intensification?
Background: The majority of patients with diabetes do not have their blood sugar, blood pressure, and lipids all under good control, despite the impact of poor control of these parameters on major diabetes complications. Diabetes patients with depression are especially likely to have poor control of these diseases. It is assumed that two major factors in poor disease control in depressed patients are: (1) poor adherence to treatment (e.g., taking medications), and (2) clinicians’ lack of aggressive treatment intensification due to a sense of futility or being waylaid by the patient’s mental health issues.
Methods: This investigation used data from an epidemiologic study/database that mailed questionnaires to patients with diabetes enrolled in the large Group Health healthcare plan in Western Washington (30 primary care clinics); patients completed initial questionnaires that inquired about baseline characteristics and clinical status (e.g., duration of diabetes, current treatment), and contained the PHQ-9 for depression; depression was defined as having 5 of the 9 depression symptoms (including depressed mood or anhedonia) at least half the days in the prior two weeks.
Over the 5-year follow-up period, records from Group Health were used to assess medical comorbidity and to measure the primary outcomes related to disease control in three specific domains: (1) hemoglobin A1C (=8.0% considered poor control), (2) blood pressure (SBP=140mm Hg considered poor control), and (3) LDL cholesterol (=130 mg/dL considered poor control). Treatment adherence was measured using pharmacy refill records from Group Health (=20% of total days that a patient did not have medication available for the condition in question was considered nonadherence), and treatment intensification was also assessed using pharmacy records (intensification occurred when there was an increase in dose, switch of medications, or addition of medication for the given disorder).
Results: Patients with depression at baseline had substantially higher rates of poor disease control in all three domains, compared to those without depression. Treatment nonadherence was also higher among depressed patients for diabetes control (37% depressed cohort vs. 23% nondepressed; adjusted odds ratio 1.98; p<0.001), hypertension control (29% vs. 17%; adjusted OR 2.06; p<0.001), and lipid control (39% vs. 24%; adjusted OR 2.43; p<0.01). However, there were no significant differences in treatment intensification between the groups, and in fact there were numerically higher rates of intensification in depressed patients for two of the three conditions.
Discussion: In case there was any doubt of the importance of depression in the medically ill, depressed patients with diabetes were twice as likely (or more) to have poor disease control in multiple key medical domains, compared to nondepressed patients, and this association was independent of other factors. As expected, diabetes patients with depression were less likely to take their medication (or, more accurately, refill their prescriptions) than nondepressed patients, and this certainly plays some role in their poor disease control and high risk for major complications. However, in contrast to expectations, physicians intensified the treatment of depressed patients at a rate equal to that in nondepressed patients—the existence of depression did not impede these patients from having their treatment optimized. The authors note that treatment intensification may be more possible in depressed patients because they had more visits to physicians than nondepressed patients, giving their doctors more opportunities to adjust their treatment.
There were some limitations of this work—for example, it focused on a single disease in a single healthcare plan in a single regional area, pharmacy refills do not exactly assess medication adherence (and other forms of adherence were not evaluated)—but this was a very large, well-designed study that asked and answered important questions in psychosomatic medicine. It leaves us wondering: what can be done to improve adherence in medically-ill patients with depression, and is treatment of depression enough to improve adherence and other important mediators of medical health?