Abstract
Objective:
We examined emotional distress in relation to metformin adherence, overall diabetes self-management, and glycemic control among adults with early type 2 diabetes (T2DM) enrolled in the GRADE study.
Methods:
Linear regression models examined cross-sectional associations of baseline depression symptoms and diabetes distress with adherence to metformin, self-management, and HbA1c, adjusting for covariates. Cognitive-affective (e.g., sadness) and somatic (e.g., sleep/appetite disturbance) depression symptoms and diabetes distress subscales were also examined.
Results:
This substudy of 1,739 GRADE participants (56% Non-Hispanic White, 18% Non-Hispanic Black, 17% Hispanic, 68% male, mean[SD] age=57.96[10.22] years, diabetes duration=4.21[2.81] years, and HbA1c=7.51[0.48]). The prevalence of clinically significant depression and diabetes distress was 8.7% and 25%, respectively. Fully adjusted models showed that depression symptoms were associated with lower self-management (p<0.0001); this effect was only significant for somatic symptoms. Diabetes distress was associated with lower adherence (p=0.0001) and self-management (p<0.0001); effects were significant for all subscales, except physician-related distress. No significant relationships of total depression symptom severity or diabetes distress with HbA1c were found.
Conclusions:
Depression symptoms and diabetes distress were robustly associated with problematic diabetes self-management among participants in GRADE. These findings highlight the need for routine assessment of depression symptoms and diabetes distress early in T2DM care.
Keywords: Depression, diabetes-related distress, treatment adherence, metformin, self-management
1.0. INTRODUCTION
Depression is among the most consistent predictors of problematic self-management across a diversity of chronic illnesses, (1) including diabetes. (2) Depression symptoms are consistently associated with hyperglycemia, (3) complications, (4,5) mortality, (6) and poorer quality of life among individuals with diabetes. (7) Based on this evidence, depression is recognized by standards of care as an important patient-reported outcome. (8) Controlled studies indicate that depression symptoms are more common among individuals with type 2 diabetes mellitus (T2DM) as compared to those without (17.6% vs. 9.8%). (9) However, studies often inadequately control for treatment regimen and health-related differences. For example, depression symptoms are more common among insulin-treated adults with T2DM as compared to those not taking insulin (10) and are associated with the onset of diabetes-related complications. (4,5) These differences may confound associations between depression symptoms and self-management or glycemic control in T2DM.
Fundamental questions remain unresolved as to what construct is being captured by commonly used self-report depression symptom measures. (11,12) Such self-reports were designed to screen for major depressive disorder (MDD). However, most cases identified by screening will not meet diagnostic criteria. (13) Some research suggests that screening cutoffs should be increased to avoid biases introduced by somatic symptoms that can overlap between depression and diabetes (e.g., fatigue, sleep and appetite disturbance). (14) Other research suggests differential effects of somatic versus cognitive-affective (e.g., sadness, loss of interest, guilt) symptoms in relation to cardiovascular outcomes, (15) diabetes prevalence, (16) and diabetes treatment adherence. (17) Rare studies that use the gold standard diagnostic interview method to identify cases of depression find no difference in the glycemic control of these individuals compared to those who do not meet diagnostic criteria. (18,19) In contrast, the presence of depression symptoms below MDD diagnostic thresholds is associated with worse diabetes self-management both cross-sectionally (20) and longitudinally. (21) These findings suggest that subsyndromal levels of severity are reflective of emotional distress that can correlate with problematic self-management. Prior studies report similar incremental relationships for diabetes complications and mortality, observable even at minor elevations in depression symptoms. (22)
Diabetes-related distress has been conceptualized as an outcome of living with the stress and burdens of diabetes and its management. (23,24) A meta-analysis of available studies indicates that prevalence of significant diabetes distress is common (36%) in T2DM. (25) Some studies suggest a closer relationship between diabetes distress and suboptimal diabetes outcomes than for depressive disorders. (18,19) Whether levels of depression symptoms continue to relate to diabetes self-management and glycemic control after accounting for the presence of diabetes distress is less clear. (17,18,26–29) Most available studies evaluating both constructs rely on relatively small samples with substantial heterogeneity in diabetes progression, complications, and diabetes treatment regimens.
The overall objective of this study is to document baseline levels of emotional distress, an umbrella term we use to refer to both depression symptoms and diabetes distress, in the Emotional Distress Substudy (EDS) of the Glycemia Reduction Approaches in Diabetes: a Comparative Effectiveness (GRADE) Study (NCT01794143). GRADE compared the effectiveness of four glucose-lowering medications on HbA1c in patients who were treated with metformin. (30,31) Using baseline data, we examine demographic, socioeconomic, and patient health characteristics associated with elevations in depression symptoms and diabetes distress. We further evaluate the cross-sectional association of these factors with diabetes medication adherence, overall diabetes self-management, and glycemic control prior to randomization. We hypothesized that both depression symptoms and diabetes distress would show independent associations with treatment nonadherence, suboptimal self-management, and higher HbA1c. We also explore sex, ethnicity, and socioeconomic status as potential moderators of these relationships.
2.0. RESEARCH DESIGN AND METHODS
A total of 5,047 patients were enrolled from 36 clinical centers and 9 additional subsites across the United States. Inclusion criteria include ≥ 30 years of age, except for American Indian and Alaska Natives eligible if ≥ 20 years, diagnosis of T2DM within 10 years, HbA1c from 6.8–8.5% (51–69 mmol/mol) and being treated with metformin alone. Screening procedures included a review of medical history, medication use, and laboratory tests of HbA1c, liver function tests, hematocrit, and serum creatinine levels. After completing the screening process, eligible participants completed a run-in period that lasted between 6–14 weeks, during which the metformin dose was titrated to 1,000–2,000 (goal) mg/day. HbA1c at the final run-in visit was between 6.8%–8.5% (51–69 mmol/mol) for study entry. Full details of the GRADE study and the EDS designs have been published elsewhere. (31,32) GRADE was a randomized clinical trial that compared the metabolic effects of four common glucose-lowering medications (sulfonylurea glimepiride, DPP-4 inhibitor sitagliptin, GLP-1 agonist liraglutide, or basal insulin glargine U-100) in metformin-treated patients. For the current study, analyses are limited to baseline data collected prior to initiation of any glucose-lowering medication beyond metformin.
2.1. GRADE Emotional Distress Sub-study
All participating GRADE sites were invited but not required to participate in the sub-study; 26 of the 36 GRADE centers and 8 subsites chose to participate. Recruitment for the substudy began more than halfway through the GRADE recruitment period. Ten centers did not participate because too few participants would be enrolled owing to slow enrollment or sites were nearing completion of enrollment, or because of limitations in site staffing. Training and certification were required of research coordinators at participating sites prior to local EDS implementation. Center enrollment ranged from 4 to 138 participants with a total of 1,739 participants enrolled from 2015–2017. The parent study informed consent was amended to include information regarding EDS and presented to participants as an embedded substudy. Participants who consented and enrolled in GRADE at participating sites were simultaneously enrolled in EDS. Participants received compensation for completion of the additional assessments and were offered a copy of the ADA Booklet, Diabetes and Your Emotional Health in English or Spanish. All site IRBs approved the study procedures prior to initiation and informed consent was obtained from all participants. Participants were approached and consented at the final run-in visit and completed the EDS assessments consisting of a self-administered questionnaire battery and collection of a blood sample at the final run-in visit or at the baseline visit, prior to the initiation of their randomly assigned glucose-lowering medication.
2.2. Measures
2.2.1. Depression symptoms Severity.
The Patient Health Questionnaire (PHQ-8) was used to assess participants’ depression symptoms. (33) The total PHQ-8 scores range between 1–24, with a higher score indicating greater depression symptom severity. Clinical levels of depression symptoms, or a positive screen, were defined as a PHQ-8 total score ≥ 10. Three items assessing little interest or pleasure, feeling down, depressed or hopeless, and feeling bad about yourself were summed to create a cognitive-affective symptom score. Five items assessing problems with sleep, fatigue, appetite, concentration, and psychomotor slowing were summed into a somatic symptom score. (15,16,17)
2.2.2. Diabetes Distress.
Diabetes distress was assessed using the 17-item Diabetes Distress Scale (DDS). (24) The measure includes four sub-scales: 1) emotional burden captures emotional and overwhelming feelings related to living with diabetes, 2) regimen distress assesses burden related to diabetes self-management, 3) interpersonal distress represents perceived lack of support and empathy from friends and family, and 4) physician distress captures perceived inadequacy of expertise, positive regard, and clear direction from health providers. The full scale and sub-scales each have good internal reliability and are sensitive to treatment regimen differences in type 2 diabetes. (24) The mean score ranges from 1 to 6 and a score of 2 or more is used as a cut-off for significant diabetes distress. (34)
2.2.3. Medication Adherence.
We measured adherence to metformin using three items referencing the last 30 days: “on how many days did you miss at least one dose of any of your diabetes medicines?”, “how good a job did you do at taking your diabetes medicines in the way you were supposed to?” and “how often did you take your diabetes medicines in the way you were supposed to?”. (35,36) Similar items were validated in T2DM against electronically monitored adherence data and HbA1c. (37) A total score was calculated for these 3 items that ranged from 0 to 100, with higher scores signifying better adherence.
2.2.4. Diabetes Self-Management.
Overall diabetes self-management was assessed using a five-item measure that has been found to be associated with HbA1c and is sensitive to change in the context of a self-management intervention. Items ask participants to rate their level of difficulty completing each behavior over the past 30 days with medication taking, blood glucose monitoring, diet, exercise, and foot self-care. Responses are scored on a scale between 1 (SO DIFFICULT – I couldn’t do it at all) and 5 (NOT DIFFICULT AT ALL – I did it exactly right). The total score was scaled from 0 to 100, with higher scores indicating better self-management. (38)
2.2.5. HbA1c.
Whole blood samples were collected at quarterly visits for measurement of glycated hemoglobin (HbA1c) by the central laboratory at the Advanced Research and Diagnostic Laboratory (ARDL) at the University of Minnesota, using a high-performance liquid chromatography method. HbA1c provides an indication of participants’ average blood glucose levels over the past 3 months, and elevated HbA1c is defined as HbA1c ≥ 7% (53 mmol/mol).
2.2.6. Demographics and other Participant Characteristics.
Medical history, current medications, alcohol intake, smoking status, race/ethnicity, and educational attainment were self-reported and obtained through interviews conducted by research staff. All physical and metabolic measurements were performed by centrally trained certified staff. Medications for depression and anxiety was self-reported. Any diabetes-related complication at baseline included: prior amputation, retinopathy, nephropathy, or neuropathy (diabetic peripheral neuropathy (DPN) or cardiovascular autonomic neuropathy (CAN)); myocardial infarction (MI) from self-report and from automatically scored ECGs; stroke by self-report; and any macrovascular disease (prior MI or stroke). Nephropathy was defined by the presence of either ACR≥30 mg/gram or eGFR<60 ml/min/m2. DPN was defined using a composite cut-off including both the MNSI questionnaire score and the clinical examination. CAN was defined using two HRV indices: standard deviation of normally conducted R-R intervals <8.2 ms, and root mean square of successive differences between normal-to-normal R-R intervals <8 ms.
2.2.7. Data Analysis.
EDS participants had very low levels of missing data at baseline. Of 1,739 participants assessed, 4 had incomplete PHQ-8 questionnaires and 17 had incomplete DDS questionnaires. Analyses for a specific variable used all available participants with non-missing data for that variable. Characteristics of the participants are described using mean (standard deviation, SD) for quantitative factors (e.g., age) and percent for discrete factors (e.g., sex, income). Spearman rank correlation described the association between the PHQ-8 total score and the DDS average score, and medication adherence, diabetes self-management and HbA1c. Separate linear regression models assessed the associations of the PHQ-8 and DDS total scores, and of their subdomains (the independent variables) with medication adherence, diabetes self-management and HbA1c (the dependent variables). The models were unadjusted (Model 1) or with various levels of adjustment as follows: adjusted for age, sex, race/ethnicity, education, income (Model 2); Model 2 + duration of T2DM, any diabetes-related complications and any macrovascular disease (Model 3); and Model 3 + DDS for PHQ-8, or Model 3 + PHQ-8 for DDS (Model 4). Moderation analyses evaluated the heterogeneity of the associations between PHQ-8 and DDS with medication adherence, diabetes self-management and HbA1c across levels of race/ethnicity, sex, and income using the closed testing principle. (39) Given the exploratory nature of our analyses, no adjustment for multiplicity was used. However, a more stringent significance level of 0.01 was determined a priori and employed, and all p-values less than this cutoff value were considered nominally significant. Data were analyzed centrally at the GRADE Coordinating Center at the George Washington University Biostatistics Center.
3.0. RESULTS
3.1. Description of the Sample.
The baseline characteristics of the full GRADE EDS cohort are reported in Table 1. This was a racially/ethnically and socioeconomically diverse sample, that was predominantly middle-aged adult, non-Hispanic White (56%) and male (68%) with a relatively short average diabetes duration of 4.2 years. Prevalence of individual diabetes complications were relatively low, but 41% of participants had at least one complication. Overall, the sample’s characteristics were similar to the overall GRADE cohort, although some differences were observed related to later initiation of EDS enrollment, resulting in slightly older, retired, male, participants who were slightly more likely to have a history of smoking, and have DPN. Use of medications for depression/anxiety was substantially higher in the EDS sub-cohort than in GRADE overall (18% vs. 5%); this likely reflects a more comprehensive assessment of psychotropic medications in EDS (Supplemental Table 1).
Table 1.
Baseline characteristics of the EDS participants both overall and separately for PHQ-8 < 10 vs. ≥ 10, and for DDS < 2 vs. ≥ 2.
| Overall EDS Cohort | PHQ-8 < 10 (n=1584) | PHQ-8 ≥ 10 (n=151) | p-value | DDS < 2 (n=1291) | DDS ≥ 2 (n=431) | p-value | |
|---|---|---|---|---|---|---|---|
| Age (years) | 57.96 (10.22) | 58.10 (10.26) | 56.44 (9.59) | 0.0455 | 58.76 (10.22) | 55.48 (9.81) | <0.0001 |
| T2DM Duration (years) | 4.21 (2.81) | 4.24 (2.83) | 3.93 (2.64) | 0.1763 | 4.14 (2.83) | 4.40 (2.74) | 0.0882 |
| Sex (%) | |||||||
| Male | 68 | 67 | 68 | 71 | 56 | ||
| Female | 32 | 33 | 32 | 0.9277 | 29 | 44 | <0.0001 |
| Race/Ethnicity (%) | |||||||
| Hispanic | 17 | 17 | 19 | 16 | 20 | ||
| Non-Hispanic Black or African American | 18 | 18 | 22 | 17 | 21 | ||
| Non-Hispanic White | 56 | 56 | 47 | 59 | 47 | ||
| Non-Hispanic Other | 9 | 9 | 13 | 0.1391 | 9 | 11 | <0.0001 |
| Education (%) | |||||||
| High school/GED or less | 30 | 29 | 32 | 30 | 28 | ||
| Some college | 29 | 29 | 34 | 29 | 31 | ||
| College/Graduate School | 41 | 42 | 34 | 0.1540 | 41 | 41 | 0.5540 |
| Income (%) <10k | 6 | 6 | 10 | 6 | 7 | ||
| 10–20k | 11 | 10 | 18 | 11 | 10 | ||
| 20–50k | 32 | 31 | 37 | 31 | 34 | ||
| 50k+ | 51 | 53 | 35 | 0.0001 | 52 | 48 | 0.3714 |
| Living (%) Alone | 17 | 16 | 23 | 16 | 19 | ||
| With another adult | 79 | 80 | 75 | 80 | 75 | ||
| With children only | 4 | 4 | 3 | 0.1236 | 3 | 5 | 0.0475 |
| Employment (%) | |||||||
| Employed | 56 | 57 | 44 | 55 | 60 | ||
| Retired | 26 | 26 | 26 | 29 | 20 | ||
| Other | 18 | 16 | 30 | 0.0001 | 17 | 20 | 0.0013 |
| Smoking (%) | |||||||
| Never | 52 | 52 | 48 | 49 | 58 | ||
| Past | 35 | 36 | 32 | 37 | 32 | ||
| Current | 13 | 12 | 21 | 0.0118 | 14 | 10 | 0.0028 |
| BMI (kg/m2) | 34.10 (6.48) | 33.82 (6.29) | 37.02 (7.66) | <0.0001 | 33.93 (6.40) | 34.54 (6.69) | 0.1004 |
| Hypertension (%) | 72 | 72 | 78 | 0.1081 | 74 | 68 | 0.0308 |
| Amputation (%) | 1 | 1 | 0 | NA | 1 | 0 | NA |
| Nephropathy (%) | 18 | 18 | 21 | 0.3668 | 18 | 19 | 0.8672 |
| Retinopathy (%) | 1 | 1 | 0 | 0.4265 | 1 | 1 | 0.7698 |
| DPN (%) | 23 | 22 | 33 | 0.0043 | 23 | 25 | 0.3644 |
| CAN (%) | 10 | 10 | 13 | 0.2576 | 11 | 8 | 0.1347 |
| Any Diabetes Complications (%) | 41 | 40 | 47 | 0.1278 | 41 | 40 | 0.7217 |
| MI (%) | 6 | 5 | 7 | 0.4455 | 6 | 5 | 0.3620 |
| Stroke (%) | 2 | 2 | 2 | 1 | 2 | 3 | 0.7080 |
| Macrovascular (%) | 7 | 7 | 9 | 0.6577 | 8 | 7 | 0.7842 |
| Medication Adherence | 89.88 (11.06) | 90.34 (10.72) | 85.01 (13.23) | <0.0001 | 90.85 (10.67) | 87.26 (11.49) | <0.0001 |
| Diabetes Self-Management | 78.81 (11.42) | 79.48 (11.33) | 71.92 (10.03) | <0.0001 | 80.62 (11.14) | 73.56 (10.56) | <0.0001 |
| Depression/Anxiety Medication (%) | 18 | 16 | 47 | <0.0001 | 17 | 22 | 0.0384 |
| HbA1c(%) | 7.51 (0.48) | 7.51 (0.49) | 7.49 (0.44) | 0.5372 | 7.50 (0.48) | 7.56 (0.47) | 0.0106 |
| HbA1c (mmol/mol) | 58.6 (5.2) | 58.6 (5.4) | 58.4 (4.8) | 0.5372 | 58.5 (5.2) | 59.1 (5.1) | 0.0106 |
The entries are mean (standard deviation) for quantitative factors (e.g., age) or percent for discrete factors (e.g., sex, race/ethnicity).
DPN=diabetic peripheral neuropathy; CAN=cardiovascular autonomic neuropathy.
3.2. Prevalence of Depression Symptoms.
Overall, 8.7% of EDS participants met the screening threshold of ≥ 10 on the PHQ-8; the average score was 3.45 (3.96), within the range interpreted to indicate no significant depression. (33) Average scores for somatic symptoms were higher (0.51 [0.54]) than for cognitive-affective symptoms (0.30 [0.52]). Table 1 shows, compared to participants who screened negative for depression, those who screened positive were younger, had significantly higher BMI, more DPN, lower income, and were more likely to be retired or employed less than full-time (see: data columns 2–4). Participants who screened positive for depression were also more likely to be taking medication for depression and/or anxiety than those who did not. Metformin adherence and overall diabetes self-management were significantly lower among those who screened positive for depression. No differences were observed for HbA1c between these groups.
3.3. Prevalence of Diabetes Distress.
Baseline total DDS score was 1.68 (0.74), within the range interpreted to indicate little or no diabetes distress(34) and midway between the response scale options of ‘not a problem’ and ‘a slight problem’ of the DDS; 25% met the screening threshold of ≥ 2 (clinically significant DDS) and 6.4% had an average DDS score 3 or above (high diabetes distress). Average scores were highest for regimen distress (1.97 [0.96]) and emotional burden (1.74 [0.89]) and lower on physician-related distress (1.36 [0.84]) and interpersonal distress (1.50 [0.88]) sub-scales. Data columns 5–7 in Table 1 show that age was significantly younger among those who screened positive for diabetes distress, with females, racial/ethnic minorities, and those who were employed full-time more likely to screen positive. Current smoking and smoking history were less common among those who screened positive for diabetes distress. Metformin adherence and overall diabetes self-management were significantly lower among those positive for diabetes distress. A modestly higher HbA1c was observed among those who screened positive for diabetes distress, but this fell short of our threshold for significance.
3.4. Overlap between depression symptoms and diabetes distress.
The zero-order correlation between PHQ-8 total scores and overall DDS scores was significant (r = .38; p <0.0001). Among DDS subscales, emotional burden (r = .40; p <0.0001), regimen distress (r = .38; p <0.0001), and interpersonal distress (r = .29; p <0.0001) subscales showed correlations with PHQ-8 total of similar magnitude; physician-related distress showed the smallest association with PHQ-8 total (r = .11; p <0.0001).
3.5. Depression symptoms and medication adherence, self-management, and glycemic control.
Various models were estimated to sequentially adjust the relationship of depression symptoms with the outcomes of interest for demographic and diabetes-related factors, with a final adjustment for diabetes distress to examine the independence of the relationships of depression symptoms with these outcomes. These models showed relationships between depression symptoms severity and metformin adherence were robust to multivariable adjustment (See Models 1–3; First 5 data columns in Table 2). Once overlap between depression symptoms and diabetes distress was adjusted for, the independent effect of depression symptoms for medication adherence was not significant; however, an independent and significant effect on diabetes self-management was retained (See Model 4; First 5 data columns in Table 2). The relationship between higher depression symptoms severity and worse self-management was also robust to adjustment. Even after accounting for overlap with diabetes distress, depression symptoms severity remained significantly associated with worse self-management. Importantly, depression symptom severity was not significantly associated with baseline HbA1c in either the unadjusted model or in any of the multi-variate models. As seen in Table 3, when examining the independent effects of PHQ-8 somatic and cognitive-affective depression symptoms severity scores in relation to metformin adherence, and self-management, results showed that effects were more robust and significant for somatic symptoms than for cognitive-affective symptoms of depression. One exception is that cognitive-affective symptom severity was significantly associated with HbA1c in the opposite direction than we hypothesized after adjustment for diabetes distress; this effect was marginal (but still negative) in models that did not adjust for overlap with diabetes distress.
Table 2.
Associations of PHQ-8 and DDS with medication adherence, diabetes self-management, and HbA1c.
| PHQ-8 Total | DDS Average | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Est | LL | UL | z-value | p-value | Est | LL | UL | z-value | p-value | |
| A. Medication Adherence | ||||||||||
| Model 1 | −0.45 | −0.58 | −0.32 | −6.81 | <0.0001 | −2.68 | −3.37 | −1.99 | −7.64 | <0.0001 |
| Model 2 | −0.40 | −0.53 | −0.26 | −5.65 | <0.0001 | −2.36 | −3.13 | −1.59 | −6.03 | <0.0001 |
| Model 3 | −0.33 | −0.48 | −0.18 | −4.39 | <0.0001 | −2.21 | −3.02 | −1.40 | −5.34 | <0.0001 |
| Model 4 | −0.20 | −0.36 | −0.04 | −2.45 | 0.0146 | −1.77 | −2.65 | −0.90 | −3.95 | 0.0001 |
| B. Diabetes Self-Management | ||||||||||
| Model 1 | −0.92 | −1.05 | −0.79 | −13.89 | <0.0001 | −4.53 | −5.23 | −3.83 | −12.68 | <0.0001 |
| Model 2 | −0.96 | −1.10 | −0.82 | −13.76 | <0.0001 | −4.87 | −5.64 | −4.10 | −12.43 | <0.0001 |
| Model 3 | −0.98 | −1.13 | −0.83 | −13.21 | <0.0001 | −4.69 | −5.50 | −3.88 | −11.34 | <0.0001 |
| Model 4 | −0.76 | −0.91 | −0.60 | −9.54 | <0.0001 | −3.02 | −3.88 | −2.17 | −6.93 | <0.0001 |
| C. HbA1c | ||||||||||
| Model 1 | −0.01 | −0.01 | 0.00 | −0.64 | 0.5206 | 0.04 | 0.01 | 0.06 | 2.46 | 0.0140 |
| Model 2 | −0.01 | −0.01 | 0.00 | −1.14 | 0.2556 | 0.03 | −0.00 | 0.07 | 1.89 | 0.0593 |
| Model 3 | −0.00 | −0.01 | 0.00 | −1.20 | 0.2286 | 0.03 | −0.01 | 0.06 | 1.45 | 0.1478 |
| Model 4 | −0.01 | −0.01 | 0.00 | −1.82 | 0.0683 | 0.04 | 0.00 | 0.08 | 2.06 | 0.0395 |
Model 1: unadjusted model; Model 2: age, sex, race/ethnicity, education, income; Model 3: Model 2 + duration of T2DM, any diabetes-related complications and any macrovascular disease; Model 4: Model 3 + DDS for PHQ-8, or Model 3 + PHQ-8 for DDS.
Est=the coefficient from the corresponding linear regression model; LL/UL=95% confidence interval.
Table 3.
Associations of depression symptoms with medication adherence, diabetes self-management, and HbA1c when somatic and cognitive-affective symptom scores are included as independent variables in the same models.
| PHQ-8 Somatic Depression Score | PHQ-8 Cognitive-Affective Depression Score | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Est | LL | UL | z-value | p-value | Est | LL | UL | z-value | p-value | |
| A. Medication Adherence | ||||||||||
| Model 1 | −2.78 | −4.07 | −1.48 | −4.19 | <0.0001 | −0.75 | −2.11 | 0.61 | −1.08 | 0.2787 |
| Model 2 | −2.73 | −4.11 | −1.34 | −3.87 | 0.0001 | −0.35 | −1.80 | 1.10 | −0.47 | 0.6362 |
| Model 3 | −2.23 | −3.71 | −0.75 | −2.95 | 0.0033 | −0.34 | −1.88 | 1.21 | −0.43 | 0.6698 |
| Model 4 | −1.49 | −2.99 | 0.014 | −1.94 | 0.0524 | −0.03 | −1.58 | 1.53 | −0.04 | 0.9702 |
| B. Diabetes Self-Management | ||||||||||
| Model 1 | −5.73 | −7.03 | −4.43 | −8.65 | <0.0001 | −1.46 | −2.81 | −0.10 | −2.11 | 0.0353 |
| Model 2 | −5.69 | −7.05 | −4.32 | −8.17 | <0.0001 | −1.86 | −3.29 | −0.43 | −2.55 | 0.0108 |
| Model 3 | −5.93 | −7.38 | −4.47 | −8.00 | <0.0001 | −1.76 | −3.27 | −0.26 | −2.29 | 0.0219 |
| Model 4 | −4.88 | −6.33 | −3.42 | −6.57 | <0.0001 | −1.02 | −2.51 | 0.48 | −1.33 | 0.1827 |
| C. HbA1c | ||||||||||
| Model 1 | 0.04 | −0.01 | 0.10 | 1.53 | 0.1306 | −0.07 | −0.13 | −0.01 | −2.21 | 0.0270 |
| Model 2 | 0.03 | −0.03 | 0.10 | 1.10 | 0.2726 | −0.07 | −0.13 | −0.01 | −2.18 | 0.0298 |
| Model 3 | 0.04 | −0.03 | 0.11 | 1.22 | 0.2239 | −0.08 | −0.15 | −0.01 | −2.36 | 0.0184 |
| Model 4 | 0.03 | −0.04 | 0.10 | 0.87 | 0.3825 | −0.09 | −0.16 | −0.02 | −2.65 | 0.0082 |
Further adjusted for: Model 1: no further adjustment; Model 2: age, sex, race/ethnicity, education, income; Model 3: Model 2 + duration of T2D, any diabetes-related complications and any macrovascular disease; Model 4: Model 3 + DDS.
Est=the coefficient from the corresponding linear regression model; LL/UL=95% confidence interval.
3.6. Diabetes Distress and medication adherence, self-management, and glycemic control.
Results showed relationships between higher total diabetes distress and lower adherence and self-management were robust in the unadjusted model and after multivariable adjustments (See Models 1–3; Second 5 data columns in Table 2). Furthermore, diabetes distress remained significantly associated with these outcomes after adjusting for overlap with depression symptoms (See Model 4; Second 5 data columns in Table 2). However, as reported for depression symptoms, total diabetes distress was not significantly associated with baseline HbA1c.
Among the DDS Subscales, the emotional burden and regimen-related distress scales were most robust to covariate adjustment in relation to medication adherence and self-management (Table 4). Regimen distress remained significantly associated with lower medication adherence and self-management after adjustment for overlap with depression symptoms. The emotional burden subscale’s relationship with medication adherence was marginal after adjustment for depression symptoms but remained significant in relation to self-management. Higher regimen distress was significantly associated with higher HbA1c in the unadjusted model and marginal in other models, including after adjustment for depression symptoms. The physician-related distress and interpersonal distress subscales were more modestly associated with lower metformin adherence and overall self-management and these effects were less robust to covariate adjustment; neither subscale was significantly associated with HbA1c.
Table 4.
Associations of DDS subdomains with medication adherence, diabetes self-management, and HbA1c.
| DDS Emotional Burden | DDS Regimen Related Distress | DDS Physician Related Distress | DDS Interpersonal Distress | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Est | LL | UL | p-value | Est | LL | UL | p-value | Est | LL | UL | p-value | Est | LL | UL | p-value | |
| A. Medication Adherence | ||||||||||||||||
| Model 1 | −1.98 | −2.56 | −1.40 | <0.0001 | −2.32 | −2.85 | −1.79 | <0.0001 | −1.19 | −1.81 | −0.57 | 0.0002 | −1.53 | −2.12 | −0.94 | <0.0001 |
| Model 2 | −1.56 | −2.20 | −0.92 | <0.0001 | −2.15 | −2.72 | −1.58 | <0.0001 | −0.80 | −1.49 | −0.12 | 0.0217 | −1.26 | −1.90 | −0.62 | 0.0001 |
| Model 3 | −1.39 | −2.06 | −0.72 | 0.0001 | −2.00 | −2.60 | −1.40 | <0.0001 | −0.80 | −1.51 | −0.09 | 0.0274 | −1.23 | −1.90 | −0.57 | 0.0003 |
| Model 4 | −0.92 | −1.66 | −0.18 | 0.0145 | −1.73 | −2.39 | −1.08 | <0.0001 | −0.66 | −1.37 | 0.06 | 0.0710 | −0.90 | −1.59 | −0.21 | 0.0111 |
| B. Self-management | ||||||||||||||||
| Model 1 | −2.91 | −3.51 | −2.32 | <0.0001 | −4.90 | −5.42 | −4.39 | <0.0001 | −1.24 | −1.89 | −0.59 | 0.0002 | −2.05 | −2.66 | −1.45 | <0.0001 |
| Model 2 | −3.10 | −3.75 | −2.45 | <0.0001 | −5.02 | −5.56 | −4.47 | <0.0001 | −1.11 | −1.83 | −0.40 | 0.0023 | −2.11 | −2.76 | −1.45 | <0.0001 |
| Model 3 | −2.90 | −3.59 | −2.22 | <0.0001 | −4.98 | −5.56 | −4.41 | <0.0001 | −0.99 | −1.73 | −0.24 | 0.0095 | −1.89 | −2.57 | −1.21 | <0.0001 |
| Model 4 | −1.26 | −1.99 | −0.54 | 0.0006 | −4.02 | −4.63 | −3.41 | <0.0001 | −0.47 | −1.18 | 0.23 | 0.1905 | −0.67 | −1.34 | 0.01 | <0.0001 |
| C. HbA1c | ||||||||||||||||
| Model 1 | 0.03 | 0.00 | 0.05 | 0.0421 | 0.03 | 0.01 | 0.06 | 0.0076 | 0.00 | −0.02 | 0.03 | 0.7409 | 0.03 | 0.01 | 0.06 | 0.0110 |
| Model 2 | 0.02 | −0.01 | 0.05 | 0.1142 | 0.03 | 0.01 | 0.06 | 0.0194 | −0.00 | −0.03 | 0.03 | 0.8679 | 0.03 | −0.00 | 0.05 | 0.0829 |
| Model 3 | 0.02 | −0.01 | 0.05 | 0.1706 | 0.02 | −0.00 | 0.05 | 0.0823 | −0.01 | −0.04 | 0.03 | 0.7115 | 0.02 | −0.01 | 0.05 | 0.1629 |
| Model 4 | 0.04 | 0.00 | 0.07 | 0.0384 | 0.03 | 0.01 | 0.07 | 0.0177 | −0.00 | −0.04 | 0.03 | 0.7873 | 0.03 | −0.00 | 0.06 | 0.0702 |
Model 1: unadjusted model; Model 2: age, sex, race/ethnicity, education, income; Model 3: Model 2 + duration of T2DM, any diabetes-related complications and any macrovascular disease; Model 4: Model 3 + PHQ-8.
Est=the coefficient from the corresponding linear regression model; LL/UL=95% confidence interval
3.7. Moderation Analyses.
Moderation Analyses (using covariates listed for Model 3 in Tables 2–4, and a heterogeneity p-value of 0.01) for the associations of total PHQ-8 and DDS scores with medication adherence, diabetes self-management, and HbA1c across race/ethnicity, sex, and income showed no significant heterogeneity for PHQ-8 effects. However, significant heterogeneity was found for the effect of DDS on worse diabetes self-management across race/ethnicity levels (p<0.0001). The associations were similar and highly significant (p<0.0001) for all Non-Hispanic groups (β±SD: −6.83 ± 0.65 for Non-Hispanic White, −5.63 ± 0.99 for Non-Hispanic Black, and −5.21 ± 1.15 for Non-Hispanic Other), but the association was not significant at the 0.01 level among Hispanics (−1.43 ± 0.72). No other potential moderation effects were significant.
4.0. DISCUSSION
The GRADE EDS examines the interrelationships among emotional distress and diabetes treatment adherence, self-management, and glycemic control in the context of a landmark randomized controlled trial comparing the effectiveness of second-line treatment regimens early in the progression of T2DM. These baseline findings provide important evidence on participant characteristics that are associated with elevations in depression symptoms and diabetes distress among individuals with early T2DM and document robust and largely independent relationships between these problems related to emotional distress and self-reported problems with diabetes self-management among patients taking only metformin with relatively well controlled T2DM.
Participants with higher BMI, lower income, and those employed in less than fulltime positions were more likely to screen positive for depression. Although previous research has consistently shown a higher prevalence of depression in women with T2DM than among men, (9) we did not find a significant difference. Among health-related variables, only the presence of diabetic peripheral neuropathy was associated with screening positive for depression. Consistent with prior evidence, the prevalence of significant diabetes distress (25%) was markedly higher than the prevalence of significant depression symptoms (8.7%) among these adults with early T2DM. Participants who were younger, female, of race/ethnicity other than Non-Hispanic White, and those who were employed were more likely to screen positive for diabetes distress. No health-related variables were associated with likelihood of screening positive for diabetes distress.
The observed correlation between total depression symptoms and overall diabetes distress demonstrates that these constructs of emotional distress in diabetes are not substantially overlapping: baseline scores in our study shared only 14% of their variance. Results from multivariable linear regression modeling showed that depression symptoms and diabetes distress were both consistently and robustly associated with lower levels of medication adherence and overall diabetes self-management. Observed effects for total depression symptoms and diabetes distress scores were unaffected by covariate adjustment. For the most part, these effects were also independent of each other, except in the case of depression symptoms’ relationship to medication non-adherence, which was marginal once adjusting for overlap with diabetes distress. Although prior, smaller studies have often yielded conflicting evidence, (17,18,26–29) the independence of the effects observed in our large sample suggests that both depression symptoms and diabetes distress would be important to assess in individuals experiencing problems with taking medication and self-management of T2DM, consistent with current recommendations. (8) An important novel finding from this study came from moderation analyses showing that diabetes distress was robustly associated with worse self-management among all ethnic/racial groups except for Hispanics/Latinos. Whether this reflects evidence of protective factors among this group or problematic measurement validity of the DDS for this group deserves further exploration. When depression symptoms were examined in models that adjusted for the overlap between somatic and cognitive-affective symptoms to examine their independent effects, only somatic symptoms of depression remained associated with lower metformin adherence and overall self-management. This is consistent with earlier research that showed only somatic symptoms of depression were significantly associated with lower electronically monitored adherence, whereas cognitive-affective symptoms showed no relationship. (17) Other research has suggested that the often-observed relationship between depression symptoms and cardiovascular disease outcomes may also be explained by somatic symptoms. (15) These symptoms overlap with hyper- and hypo-glycemia, treatment side-effects, and other health problems, raising the possibility of confounding in the assessment of depression symptoms among individuals with T2DM. (14) Differences between these symptom dimensions deserve further attention in future studies. DDS subscales were largely similarly associated with lower adherence and self-management, with regimen distress and emotional burden showing the most robust associations.
In contrast to early studies, (3) we found no evidence for a relationship between depression symptoms and HbA1c. When we examined the independent effects of somatic and cognitive-affective symptoms, we found a significant association for higher levels of cognitive-affective depression symptoms being associated with modestly lower HbA1c. However, this finding was opposite to the hypothesized direction and should be interpreted with caution. For diabetes distress, the total score was marginally associated with higher HbA1c in the expected direction, but these differences were modest in magnitude and did not meet the threshold set for statistical significance. Examination of subscale relationships suggested this effect was primarily driven by regimen distress. These findings are also inconsistent with prior studies reporting an association between diabetes distress and higher HbA1c. (17–20,26–29) One strength of the current study that might explain our lack of HbA1c findings is the greater homogeneity of diabetes progression, diabetes complications and diabetes treatment relative to other available studies. Our design allowed us to control for these possible confounding factors. It is possible that relationships between emotional distress and higher HbA1c levels reported by prior studies is driven by diabetes progression, treatment regimen intensification, and limitations in health and functioning related to the onset of diabetes complications, or other uncontrolled factors. However, it is also possible that the restricted baseline range of HbA1c per the GRADE protocol attenuated our ability to demonstrate relationships between emotional distress and higher HbA1c. This relationship may become clearer in forthcoming studies using longitudinal follow-up data, since over time with ongoing follow-up these patients will experience a progressive increase in HbA1c levels.
The strengths of our study include its large and ethnically/racially and socioeconomically diverse sample, comprehensive assessment of depression symptoms, diabetes distress and potential confounding variables, and inclusion of participants who were relatively homogenous with respect to T2DM duration, treatment, and glycemic control. The diversity and size of our sample also allowed us to explore evidence for effect moderation by sex, race/ethnicity, and socioeconomic status. Limitations include our cross-sectional design. Forthcoming longitudinal analyses will evaluate the effect of time, initial GRADE treatment, treatment intensification, and onset of complications on emotional distress over time and will also examine whether changes in emotional distress predict changes in glycemic control over time. Our use of self-reports for medication adherence and self-management may be vulnerable to bias. However, self-ratings for diabetes medication adherence similar to those used here have previously been shown to have equivalent relationships with glycemic control as compared to those observed for more costly and burdensome electronic monitoring. (37) Levels of depression symptoms and diabetes distress in this sample were lower than prior studies of individuals with more heterogeneous diabetes progression and treatment regimens. We are also unable to identify which cases of screen-detected depression are true positives vs. false positives because we lack diagnostic interviews for depression. Thus, our results should be generalizable to individuals with depression symptoms that may or may not meet criteria for a psychiatric diagnosis and we cannot speak to the role that meeting criteria for major depression disorder or other psychiatric conditions played in the reported associations.
In conclusion, results from the baseline assessment visit of the GRADE EDS indicate that the prevalence of clinically significant diabetes distress was markedly higher than that for clinically significant depression. Variations in both indicators of emotional distress were more related to demographic factors than to illness- or health-related variables. These indicators of emotional distress were robustly and mostly independently associated with metformin nonadherence and problems with overall diabetes self-management but were not associated with baseline glycemic control. Moderation analyses suggested that the relationship between total diabetes distress and worse self-management is weaker for Hispanics than for other ethnic/racial groups. This finding deserves further exploration in future research. Future studies will evaluate the directionality of these relationships longitudinally and will examine potential differences in emotional distress over time across GRADE treatment arms.
Supplementary Material
Acknowledgments.
The GRADE Study Research Group is deeply grateful to our participants whose loyal dedication made GRADE possible.
Funding.
The GRADE Study was supported by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the National Institutes of Health under Award Number U01DK098246. The planning of GRADE was supported by a U34 planning grant from the NIDDK (U34-DK-088043). The American Diabetes Association supported the initial planning meeting for the U34 proposal. The National Heart, Lung, and Blood Institute and the Centers for Disease Control and Prevention also provided funding support. The Department of Veterans Affairs provided resources and facilities. Additional support was provided by grant numbers P30 DK0111022, P30 DK017047, P30 DK020541, P30 DK020572, P30 DK072476, P30 DK079626, P30 DK092926, U54 GM104940, UL1 TR000170, UL1 TR000439, UL1 TR000445, UL1 TR001102, UL1 TR001108, UL1 TR001409, 2UL1TR001425, UL1 TR001449, UL1 TR002243, UL1 TR002345, UL1 TR002378, UL1 TR002489, UL1 TR002529, UL1 TR002535, UL1 TR002537, UL1 TR002541 and UL1 TR002548. Educational materials have been provided by the National Diabetes Education Program. Material support in the form of donated medications and supplies has been provided by Becton, Dickinson and Company, Bristol-Myers Squibb, Merck & Co., Inc., Novo Nordisk, Roche Diagnostics, and Sanofi. The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflict of Interest:
JSG reports grants from NIH and JDRF, outside the submitted work. CJH reports grants from NIDDK, NIA, and JDRF outside the submitted work. EAW reports grants from NIH/NIDDK, participation in a data safety monitoring board or advisory board for a NIH P20 center grant, and leadership as a board of director for Chronic Care International, outside the submitted work. HKS, IB, GCR, AND, AW, DE, JBF and AC have nothing to disclose.
Guarantor Statement:
IB and JG are the guarantors of this work and as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Clinical Trial registration number: NCT01794143
Prior Presentation: No prior presentation
Data Availability:
The Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE) is being conducted with funding from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). This manuscript is based on the baseline (pre-treatment) data from the 5047 participants enrolled into the study. This baseline data will be archived with the NIDDK data repository and will be available for sharing with other investigators in 2022.
References
- 1.DiMatteo MR, Lepper HS, Croghan TW. (2000) Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherence. Archives of Internal Medicine 160(14):2101–7. DOI: 10.1001/archinte.160.14.2101. [DOI] [PubMed] [Google Scholar]
- 2.Gonzalez JS, Peyrot M, McCarl LA, Collins EM, Serpa L, Mimiaga MJ, Safren SA (2008) Depression and diabetes treatment nonadherence: a meta-analysis. Diabetes Care 31(12):2398–403. DOI: 10.2337/dc08-1341. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lustman PJ, Anderson RJ, Freedland KE, De Groot M, Carney RM, Clouse RE (2000) Depression and poor glycemic control: a meta-analytic review of the literature. Diabetes Care 23(7):934–42. DOI: 10.2337/diacare.23.7.934. [DOI] [PubMed] [Google Scholar]
- 4.De Groot M, Anderson R, Freedland KE, Clouse RE, Lustman PJ (2001) Association of depression and diabetes complications: a meta-analysis. Psychosomatic Medicine 63(4):619–30. DOI: 10.1097/00006842-200107000-00015. [DOI] [PubMed] [Google Scholar]
- 5.Nouwen A, Adriaanse MC, van Dam K, Iversen MM, Viechtbauer W, Peyrot M, Caramlau I, Kokoszka A, Kanc K, de Groot M, Nefs G (2019) Longitudinal associations between depression and diabetes complications: a systematic review and meta analysis. Diabetic Medicine 36 (12):1562–1572. DOI: 10.1111/dme.14054. [DOI] [PubMed] [Google Scholar]
- 6.Park M, Katon WJ, Wolf FM (2013) Depression and risk of mortality in individuals with diabetes: a meta-analysis and systematic review. General Hospital Psychiatry 35(3):217–25. DOI: 10.1016/j.genhosppsych.2013.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Schram MT, Baan CA, Pouwer F (2009) Depression and quality of life in patients with diabetes: a systematic review from the European depression in diabetes (EDID) research consortium. Current Diabetes Reviews 5(2):112–9. DOI: 10.2174/157339909788166828. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Young-Hyman D, De Groot M, Hill-Briggs F, Gonzalez JS, Hood K, Peyrot M (2016) Psychosocial care for people with diabetes: a position statement of the American Diabetes Association. Diabetes Care 39(12):2126–40. DOI: 10.2337/dc16-2053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ali S, Stone MA, Peters JL, Davies MJ, Khunti K (2006) The prevalence of co-morbid depression in adults with Type 2 diabetes: a systematic review and meta-analysis. Diabetic Medicine 23(11):1165–1173. DOI: 10.1111/j.1464-5491.2006.01943.x. [DOI] [PubMed] [Google Scholar]
- 10.Bai X, Liu Z, Li Z, & Yan D (2018). The association between insulin therapy and depression in patients with type 2 diabetes mellitus: a meta-analysis. BMJ Open 8(11), e020062.DOI: 10.1136/bmjopen-2017-020062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gonzalez JS, Fisher L, Polonsky WH (2011) Depression in diabetes: have we been missing something important?. Diabetes Care 34(1):236–9. DOI: 10.2337/dc10-1970. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Fisher L, Gonzalez JS, Polonsky WH (2014) The confusing tale of depression and distress in patients with diabetes: a call for greater clarity and precision. Diabetic Medicine (7):764–72. DOI: 10.1111/dme.12428. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Roy T, Lloyd CE, Pouwer F, Holt RI, & Sartorius N (2012) Screening tools used for measuring depression among people with Type 1 and Type 2 diabetes: a systematic review. Diabetic Medicine 29(2):164–175. DOI: 10.1111/j.1464-5491.2011.03401.x. [DOI] [PubMed] [Google Scholar]
- 14.Twist K, Stahl D, Amiel SA, Thomas S, Winkley K, Ismail K (2013) Comparison of depressive symptoms in type 2 diabetes using a two-stage survey design. Psychosomatic Medicine 75(8):791–7. DOI: 10.1097/PSY.0b013e3182a2b108. [DOI] [PubMed] [Google Scholar]
- 15.de Miranda Azevedo R, Roest AM, Hoen PW, & De Jonge P (2014) Cognitive/affective and somatic/affective symptoms of depression in patients with heart disease and their association with cardiovascular prognosis: a meta-analysis. Psychological Medicine 44(13), 2689–2703. DOI: 10.1017/S0033291714000063. [DOI] [PubMed] [Google Scholar]
- 16.Wiltink J, Michal M, Wild PS, Schneider A, König J, Blettner M, Münzel T, Schulz A, Weber M, Fottner C, Pfeiffer N, Lackner K, Beutel ME (2014) Associations between depression and diabetes in the community: do symptom dimensions matter? Results from the Gutenberg Health Study. PLoS One 9(8):e105499. DOI: 10.1371/journal.pone.0105499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Gonzalez JS, Kane NS, Binko DH, Shapira A, & Hoogendoorn CJ (2016) Tangled up in blue: unraveling the links between emotional distress and treatment adherence in type 2 diabetes. Diabetes Care 39(12), 2182–2189. DOI: 10.2337/dc16-1657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Fisher L, Glasgow RE, Strycker LA (2010) The relationship between diabetes distress and clinical depression with glycemic control among patients with type 2 diabetes. Diabetes Care 33(5):1034–6. DOI: 10.2337/dc09-2175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Fisher L, Skaff MM, Mullan JT, Arean P, Mohr D, Masharani U, Glasgow R, Laurencin G (2007)Clinical depression versus distress among patients with type 2 diabetes: not just a question of semantics. Diabetes Care 30(3):542–8. DOI: 10.2337/dc06-1614. [DOI] [PubMed] [Google Scholar]
- 20.Gonzalez JS, Safren SA, Cagliero E, Wexler DJ, Delahanty L, Wittenberg E, Blais MA, Meigs JB, Grant RW (2007) Depression, self-care, and medication adherence in type 2 diabetes: relationships across the full range of symptom severity. Diabetes Care 30(9):2222–7. DOI: 10.2337/dc07-0158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gonzalez JS, Safren SA, Delahanty LM, et al. (2008) Symptoms of depression prospectively predict poorer self-care in patients with Type 2 diabetes. Diabetic Medicine 25(9):1102–1107. DOI: 10.1111/j.1464-5491.2008.02535.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Black SA, Markides KS, Ray LA (2003) Depression predicts increased incidence of adverse health outcomes in older Mexican Americans with type 2 diabetes. Diabetes Care 26(10):2822–8. DOI: 10.2337/diacare.26.10.2822. [DOI] [PubMed] [Google Scholar]
- 23.Polonsky WH, Anderson BJ, Lohrer PA, Welch G, Jacobson AM, Aponte JE, Schwartz CE (1995) Assessment of diabetes-related distress. Diabetes Care 18(6):754–60. DOI: 10.2337/diacare.18.6.754. [DOI] [PubMed] [Google Scholar]
- 24.Polonsky WH, Fisher L, Earles J, Dudl RJ, Lees J, Mullan J, Jackson RA (2005) Assessing psychosocial distress in diabetes: development of the diabetes distress scale. Diabetes Care 28(3):626–31. DOI: 10.2337/DIACARE.28.3.626. [DOI] [PubMed] [Google Scholar]
- 25.Perrin NE, Davies MJ, Robertson N, Snoek FJ, & Khunti K (2017) The prevalence of diabetes‐specific emotional distress in people with Type 2 diabetes: a systematic review and meta-analysis. Diabetic Medicine 34(11), 1508–1520.DOI: 10.1111/dme.13448. [DOI] [PubMed] [Google Scholar]
- 26.Aikens JE. (2012) Prospective associations between emotional distress and poor outcomes in type 2 diabetes. Diabetes Care 35(12):2472–8. DOI: 10.2337/DC12-0181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Gonzalez JS, Delahanty LM, Safren SA, Meigs JB, Grant RW. (2008) Differentiating symptoms of depression from diabetes-specific distress: relationships with self-care in type 2 diabetes. Diabetologia 51(10):1822–5. DOI: 10.1007/s00125-008-1113-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gonzalez JS, Shreck E, Psaros C, Safren SA. (2015) Distress and type 2 diabetes-treatment adherence: A mediating role for perceived control. Health Psychology 34(5):505. DOI: 10.1037/hea0000131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Fisher L, Mullan JT, Arean P, Glasgow RE, Hessler D, Masharani U. (2010) Diabetes distress but not clinical depression or depressive symptoms is associated with glycemic control in both cross-sectional and longitudinal analyses. Diabetes Care 33(1):23–8. DOI: 10.2337/dc09-1238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.The GRADE Study Research Group. (2022) Glycemia Reduction in Type 2 Diabetes — Glycemic Outcomes. New England Journal of Medicine 387(12):1063–1074. DOI: 10.1056/NEJMoa2200433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Nathan DM, Buse JB, Kahn SE, Krause-Steinrauf H, Larkin ME, Staten M, et al. (2013) Rationale and design of the glycemia reduction approaches in diabetes: a comparative effectiveness study (GRADE). Diabetes Care 2254–2261. DOI: 10.2337/dc13-0356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Cherrington AL, Krause-Steinrauf H, Bebu I, Naik AD, Walker E, Golden SH, et al. (2021) Study of emotional distress in a comparative effectiveness trial of diabetes treatments: Rationale and design. Contemporary Clinical Trials 107, 106366. DOI: 10.1016/j.cct.2021.106366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kroenke K, Strine TW, Spitzer RL, Williams JB, Berry JT, Mokdad AH. (2008) The PHQ-8 as a measure of current depression in the general population. Journal of Affective Disorders 114(1–3):163–173. DOI: 10.1016/j.jad.2008.06.026. [DOI] [PubMed] [Google Scholar]
- 34.Fisher L, Hessler DM, Polonsky WH, Mullan J. (2012) When is diabetes distress clinically meaningful?: establishing cut points for the Diabetes Distress Scale. Diabetes Care 35(2):259–64. DOI: 10.2337/DC11-1572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wilson IB, Fowler FJ Jr., Cosenza CA, et al. (2014) Cognitive and field testing of a new set of medication adherence self-report items for HIV care. AIDS Behav. 18(12):2349–58. DOI: 10.1007/s10461-013-0610-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Wilson IB, Lee Y, Michaud J, Fowler FJ Jr., Rogers WH. (2016) Validation of a New Three-Item Self-Report Measure for Medication Adherence. AIDS Behav. 20(11):2700–2708. DOI: 10.1007/s10461-016-1406-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Gonzalez JS, Schneider HE, Wexler DJ, Psaros C, Delahanty LM, Cagliero E, & Safren SA (2013) Validity of Medication Adherence Self-Reports in Adults With Type 2 Diabetes. Diabetes Care 36(4), 831–837. DOI: 10.2337/dc12-0410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Heisler M, Smith DM, Hayward RA, Krein SL, Kerr EA. (2003) How well do patients’ assessments of their diabetes self-management correlate with actual glycemic control and receipt of recommended diabetes services?. Diabetes Care 26(3):738–743. DOI: 10.2337/diacare.26.3.738. [DOI] [PubMed] [Google Scholar]
- 39.Markus R, Peritz E, Gabriel KR. (1976) On closed testing procedures with special reference to ordered analyses of variance. Biometrika 63(3), 655–60. DOI: 10.1093/biomet/63.3.655 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE) is being conducted with funding from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). This manuscript is based on the baseline (pre-treatment) data from the 5047 participants enrolled into the study. This baseline data will be archived with the NIDDK data repository and will be available for sharing with other investigators in 2022.