Participants (n = 2759) were compared to those not completing the survey (n = 1017), not allowing use of automated data (n = 139) or missing the depression questionnaire at either baseline or 5-year follow-up (n = 11) (). Participants included in this study did not differ from nonparticipants on gender (participants 49.2% female vs. nonparticipants 49.3% female) or baseline PHQ-9 score (participants 5.6 ± 5.5, nonparticipants 5.6 ± 5.7), but did significantly differ on age (participants 61.3 ± 12.4, nonparticipants 59.5 ± 13.6, t(3924) = 2.19, p < .03), years with diabetes (participants 9.0 ± 9.0, nonparticipants 8.2 ± 8.2, t(3924) = 2.60, p < .009), and race (participants 80.9% Caucasian, nonparticipants 67.0% Caucasian, X2(1) = 87.41, p < .001). The large sample size magnifies the small clinical differences found between participants and nonparticipants. Because the larger difference between participants and nonparticipants on race was more meaningful, race is included as a covariate in subsequent analysis.
describes the initial PHQ-9 scores and 5-year follow-up PHQ-9 scores across the four subgroups of patients. We also analyzed the number of patients with a ≥10 at baseline who had a ≥5 point increase in their PHQ-9 score; there were only 16 patients meeting this criteria and they were included in the persistent depression group.
Mean and SD PHQ-9 scale scores at baseline and 5-year assessments
describes that the four depression groups differed globally on all measures except for race/ethnicity, Rxrisk, microvascular events/procedures post-baseline, and macrovascular events pre-baseline. In general, the no depression group differed significantly from the improved depression and persistent depression groups, whereas the no depression and the worsening depression groups were generally similar in baseline characteristics. In comparison to those with no depression those with either improved depression or persistent depression were more likely to be women, married, younger, have Type II diabetes, have higher HbA1c levels, have more diabetes complications, have a higher likelihood of being treated with insulin, and had more microvascular events/procedures before baseline.
Baseline or pre-baseline characteristics of the depression groups
Based on the ANCOVAs, demonstrates that the four depression groups had significant differences in the number of days with a healthy diet, number of days eating ≥5 servings of fruits and vegetables, number of days with ≥30 min of physical activity and ≥30 min of specific exercise at both baseline and at the 5-year follow-up after adjusting for the covariates.
Unadjusted means and standard deviations for the health behaviors at each assessment, and tests of difference in adjusted means
At baseline, compared to the no depression group the improved depression group had significantly fewer mean days in the prior week of each of the four measures of healthy eating and exercise, significantly more days of eating high fat foods, and a significantly greater proportion of smokers (). At baseline, compared to the no depression group, the worsening depression group did not differ on any self-care measure and the persistent depression group had significantly fewer days in the prior week on the four measures of healthy eating and healthy exercise, but did not differ on the number of days of eating high fat food or the proportion of smokers.
At 5-year assessment, compared to the no depression group, those with improved depressive symptoms did not differ on any of the five dietary or exercise symptoms; those with worsening depressive symptoms had significantly fewer days on the four healthy eating and exercise measures and significantly more days of eating high fat foods; those with persistent symptoms had significantly fewer days eating a healthy diet and fewer days on the two exercise measures, but not on the number of days eating fruits and vegetables of at least 5 servings or on the number of days eating high fat foods (). No differences in the percent of patients who smoke were seen at the 5-year follow-up across the four depression groups.
shows that patients with either worsening or persistent depressive symptoms were significantly more likely than the no depressive symptom group over the 5-year period to have a decrease from 2–7 days to 0–1 day in the prior week of: eating a healthy diet, doing ≥30 min of a specific physical continuous activity or ≥30 min of a specific exercise session (see ). Compared to the no depression group, the subgroup with worsening depressive symptoms was also significantly more likely to have a decrease from 2–7 days to 0–1 day in the prior week of eating fruits and vegetables ≥5 times a day.
Relative risk (RR) ratios and 95% CI for the depression groups based on changes in health behaviorsa. All diet and exercise questions were dichotomized into poor (0–1 day) and good (2–7 days)
also describes that in patients who had very low initial levels of self-care, the subgroup with clinically important worsening depressive symptoms was significantly more likely than the subgroup with no depression to have remained at 0–1 day of eating at least 5 servings a day of fruits and vegetables as well as doing ≥30 min of a specific physical activity. The subgroup with persistently high levels of depressive symptoms was significantly more likely than the no depression group to have remained at 0–1 day of ≥30 min of either any physical activity or a specific physical exercise.
For the subgroup with improved depressive symptoms the only differences from the no depression group is that these patients were significantly more likely to have decreases in ≥30 min of physical activity and ≥30 min of specific exercise at 5-year follow-up.
Initiation of smoking was significantly related to worsening depression. There were no other significant changes in smoking associated with differences in course of depressive symptoms.