The results of this study indicate that in this indigent population with type 2 diabetes, health-related QOL differed in patients with depression compared to nondepressed patients after adjustment for relevant covariates. Patients who were depressed reported lower health-related QOL than nondepressed patients. In contrast, measures of metabolic control did not differ significantly between depressed and nondepressed patients after adjustment for relevant covariates.
There are three important findings that emerged from this study. First, depressed patients with diabetes reported decreased QOL when compared to nondepressed patients. QOL generally includes physical and mental functional functioning and the associated feelings of well-being.18
The differences between the adjusted means of the PCS (2.9) and MCS (5.2) scores of the depressed and the nondepressed patients were not only statistically but also clinically meaningfully different. Edelman et al.19
reported that mean differences of 2.0–2.5 points on QOL life scales among patients with diabetes are clinically meaningful differences on SF-12 subscales and other health-related QOL scales. These values were based on Cohen's standardized effect size for determining minimally clinically important differences in health-related QOL measures.20,21
These findings agree with at least two previous studies. Eren et al.22
found that the presence of depression resulted in a significant reduction in QOL in individuals with diabetes. Additionally, a 2004 population-based survey of patients with diabetes also revealed lower health-related QOL in patients with diabetes and those at risk for diabetes (three to five diabetes-related risk factors).23
It is also believed that comorbid depression has an additive effect on chronic diseases and negatively affects QOL.24
Therefore it is tenable that depressed patients with diabetes are more likely to rate their QOL lower than nondepressed patients.
Complications due to improper management of diabetes are the most likely determinants of negative patient-reported QOL. To test this hypothesis, Goldney et al.24
evaluated the association between depression and QOL in individuals with (1) depression only, (2) diabetes only, and (3) depression and diabetes and found significant differences in both the physical and mental measures of QOL. The lowest QOL scores was reported by patients with diabetes and coexisting depression, thereby highlighting the effects of depression on patient-reported assessments of QOL. The findings of our study and those of Goldney et al.24
underscore the need to consider patient perceptions of disease management and control as depression has a negative influence on QOL and overall outcomes.
Second, measures of metabolic control did not differ between depressed and nondepressed patients after adjustment of relevant covariates. These findings agree with Van Tilburg et al.,25
who also reported a lack of meaningful association between depressive symptomology and glycemic control in patients with type 2 diabetes, although associations were present in patients with type 1 diabetes. However, we note that a number of studies report positive associations between depression and poor glycemic control.26–30
Gary et al.26
found that depressive symptoms were associated with suboptimal levels of hemoglobin A1C and LDL cholesterol levels. Ciechanowski et al.27
also found significant associations between depressive symptoms and hemoglobin A1C levels in patients with type 1 diabetes, although not in patients with type 2 diabetes. Richardson et al.28
assessed the longitudinal effects of depression on glycemic control and found that over 4 years of follow-up there was a significant longitudinal relationship between depression and glycemic control and that depression was associated with persistently higher hemoglobin A1c levels over the time period. Similarly, Wagner et al.29
found higher hemoglobin A1C and more diabetes complications in African Americans with higher depressive symptoms after controlling for confounders. Finally, Hailpern et al.30
found significant associations between depression and hemoglobin A1C in an inner city male population with diabetes having severe depression, but the same associations were not observed in female patients with severe depression. Confirming the interrelationship between depressive symptomology and metabolic control is of considerable interest for clinicians managing patients with diabetes because proper glycemic control is critical to self-management of diabetes and generally associated with better QOL in patients with diabetes.24,31
Third, total cholesterol levels were higher (although not statistically significant) in depressed patients. These findings are supported by previous observations by Gary et al.,26
who also found higher total cholesterol levels in African American patients with type 2 diabetes and depressive symptoms. It is unclear why depressed patients have higher total cholesterol levels. Poor diet, limited physical activity, and poor medication adherence may be one explanation for the observed differences between depressed and nondepressed patients.
Despite what we believe are very interesting findings, we are left with more questions regarding what specific roles patient-specific characteristics play in determining the diabetes outcomes measured in this study. Few studies have given full consideration to the influences low incomes and lack of a usual source of care have on diabetes management. Poor and underserved populations suffer from limited access to adequate healthcare services, resulting in disparities in outcomes.32
Low income patients have a higher prevalence of diabetes9,10
and greater likelihood of more adverse long-term complications.33
Rabi et al.34
proposed that clinical, biologic, and behavioral characteristics differ by income levels, resulting in more negative diabetes outcomes. However, there is evidence that when provided proper disease management, indigent patients can have outcomes similar to patients at higher income levels.35
For example, indigent patients seen by a pharmacist certified in diabetes education for diabetes education were compared to patients seen in a primary care clinic who were seen by a nonpharmacist educator where 88% of the patients had insurance.35
After 3 years, total cholesterol, LDL cholesterol, hemoglobin A1C, and triglyceride levels declined equally in both groups. Additionally, the indigent patients seen by the pharmacist faculty clinician reached the American Diabetes Association target for A1C (<7%), which suggests that even though patients differ by income, equivalent outcomes can be achieved with patient-focused diabetes care.
We recognize a number of limitations of our study. First, as noted above, our patient sample represented only indigent patients. Approximately 25% of our sample had incomes less than $5,000 and an average educational levels of 11 years. We acknowledge that education levels are frequently tied to poor health literacy and associated poor diabetes knowledge, poor self-management practice, and poor outcomes.36
Second, because we did not stratify by severity of depression, we were unable to identify or characterize the specific mechanism by which severity of depression affects the variables of interest. Third, cross-sectional studies by nature are limited in their ability to adequately discuss direction of causality.10
In summary, in this study, we obtained critical information related to the effects of depression on adequate metabolic control and health-related QOL. Our results demonstrate that depressed and nondepressed indigent patients with diabetes may exhibit differing patterns of QOL.