Objective: To investigate whether the severity of depression symptoms is associated with poor long-term control of plasma glucose levels in individuals with type 2 diabetes. Methods: Electronic health record [EHR] data of 2842 individuals with mental illness enrolled in the Penn State Clinical Assessment and Rating Evaluation System [PCARES] registry were used. Demographics, body mass index [BMI], baseline type 2 diabetes mellitus [T2DM] status and all available glucose labs were extracted from the EHR. The nine-item patient health questionnaire [PHQ-9] was used to determine baseline depression symptoms. PHQ-9 scores greater than/equal to 10 indicated moderate-to-severe depression symptoms, whereas scores less than 10 indicated none-to-mild depression symptoms. While the baseline glucose measurement had to be within ± 90 days of the baseline PHQ-9 date, longitudinal glucose measurements had to be on or after this date and within one year of the second follow-up glucose lab test date. Each glucose lab test had to be within one year of the preceding lab test. There were 917 individuals that met the criteria for baseline and follow-up glucose measurements and contributed to the effective study sample. Linear mixed-effects models were used to assess the association between baseline depression and changes in glucose levels, with a focus on persons with baseline diabetes. Results are reported as beta-coefficients (standard errors [SE]) and P -values. Results: The study sample included 917 individuals, with 65% females (596 of 917) and 85% (780 of 917) Non-Hispanic Caucasians. The mean (SD) age, PHQ-9 score, BMI, and glucose were 47.7 (16.9) years, 12.0 (7.1), 31.6 (8.6) Kg/m 2, and 115.9 (48.8) mg/dl, respectively. At baseline, 62.0% had moderate-to-severe depression (569 of 917), and 37.1% of persons had T2DM (341 of 917). Among individuals without T2DM, there was no association between the severity of depression symptoms and follow-up glucose levels with a beta (SE) of 1.1 (1.1) and P =0.32. Among persons with T2DM (N=341), there was an average increase in glucose levels by 2.8 (2.1) mg/dl, P =0.17, per year of follow-up. When stratified by the severity of depression symptoms, individuals with moderate-to-severe depression symptoms had a significant increase in glucose levels at 6.2 (2.7) mg/dl, P =0.02, per year of follow-up, indicative of poor control of blood glucose levels. Whereas, among persons with none-to-mild depression symptoms, blood glucose levels showed a non-significant decline at 2.6 (3.4) mg/dl, P =0.45, per year of follow-up. Conclusions: In this clinic-based sample of persons with mental illness, moderate-to-severe depression symptoms were associated with significantly increasing blood glucose levels among persons with co-morbid T2DM. Our findings underscore integrated physical and mental healthcare services and routine depression screening among persons with diabetes.