In the Journals Plus

Prior nonadherence, addition of insulin therapy increase noncompliance risk in diabetes, depression

Show Citation

March 17, 2017

Patients with diabetes and comorbid depression are at increased risk for nonadherence to their diabetes therapy if they have a history of treatment noncompliance before the depression diagnosis or experience a change in their treatment plan during follow-up, according to an analysis of Canadian public health insurance data.

“Depression is associated with low diabetes self-care, such as adherence to antidiabetic drug treatment,” Carlotta Lunghi, MSc, PhD, a postdoctoral fellow with the faculty of medicine and health sciences at the Université de Sherbrooke in Quebec, told Endocrine Today. “Among patients with diabetes and comorbid depression, some of them are at higher risk of nonadherence, such as younger patients and those having changed their initial antidiabetic treatment to a combination of another oral antidiabetic drug plus insulin.”

Carlotta Lunghi
Carlotta Lunghi

In a population-based retrospective study, Lunghi and colleagues analyzed administrative data from 3,106 patients with type 2 diabetes who were prescribed at least one new diabetes agent between 2000 and 2006, using data from the health insurance board of Quebec (RAMQ), the Institut de la statistique du Quebec and the Quebec Registry of Hospitalizations (57.5% women). Included patients also had a diagnosis of depression during follow-up (from 2000 to 2007; average age at depression diagnosis, 67 years). The dependent variable was nonadherence to diabetes therapy in the year after the diagnosis of depression; nonadherence was defined as having a proportion of days covered below 90% for one or more diabetes agents (assessed by drug claims data). Researchers analyzed utilization of health care resources, including different physicians visited, prescriptions filled and any hospitalizations before depression diagnosis, and assessed variables related to the presence of comorbidities and the severity of diabetes.

Within the cohort, metformin therapy was the most commonly prescribed initial diabetes treatment (72.9%); 63.8% had no changes made to their initial therapy regimen. Median proportion of days covered for diabetes therapy before a depression diagnosis was 91.1%; median proportion of days covered fell to 89.1% in the first year after depression diagnosis.

In logistic regression analysis, researchers found that baseline nonadherence to diabetes treatment was strongly associated with nonadherence to treatment after a depression diagnosis (adjusted OR = 7.86; 95% CI, 6.63-9.33). Other factors influencing treatment adherence included receiving a diabetes prescription from a provider other than an endocrinologist or internist (adjusted OR = 1.76; 95% CI, 1.13-2.72), the addition of insulin therapy (adjusted OR = 1.58; 95% CI, 0.68-3.65) and the addition of insulin plus a change in oral diabetes agent (adjusted OR = 1.78; 95% CI, 1.03-3.07). Factors related to better therapy adherence included older age (aged at least 44 years), low socioeconomic status or having another diabetes agent added to or replacing the current therapy regimen, according to the researchers.

“The factors associated with nonadherence that we identified in this study may help clinicians to detect and pay special attention to patients at higher risk for nonadherence,” Lunghi said. “For instance, patients with previous nonadherence behaviors could benefit from closer follow-up by the clinician so that he/she can quickly notice symptoms of depression and a lack of adherence to antidiabetic treatment.”

Lunghi noted that the factors associated with nonadherence found in the study are mostly unmodifiable factors, such age or sex.

“However, these results could help other researchers to target a high-risk population to conduct a study aiming to identify modifiable risk factors of nonadherence that are not present in administrative databases,” Lunghi said. – by Regina Schaffer

For more information:

Carlotta Lunghi, MSc, PhD, can be reached at the Université de Sherbrooke, K1, 2500 Boulevard de I’Universite, Sherbrooke, QC, J1K, 2R1, Canada; email:

Disclosure: The researchers report no relevant financial disclosures.