Effect of collaborative depression treatment on risk for diabetes: A 9-year follow-up of the IMPACT randomized controlled trial

Author/Creator ORCID





Citation of Original Publication

Khambaty T, Callahan CM, Stewart JC (2018) Effect of collaborative depression treatment on risk for diabetes: A 9-year follow-up of the IMPACT randomized controlled trial. PLoS ONE 13(8): e0200248. https://doi.org/10.1371/journal.pone.0200248


This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please contact the author.
Attribution 4.0 International (CC BY 4.0)


Considerable epidemiologic evidence and plausible biobehavioral mechanisms suggest that depression is an independent risk factor for diabetes. Moreover, reducing the elevated diabetes risk of depressed individuals is imperative given that both conditions are leading causes of death and disability. However, because no prior study has examined clinical diabetes outcomes among depressed patients at risk for diabetes, the question of whether depression treatment prevents or delays diabetes onset remains unanswered. Accordingly, we examined the effect of a 12-month collaborative care program for late-life depression on 9-year diabetes incidence among depressed, older adults initially free of diabetes. Participants were 119 primary care patients [M (SD) age: 67.2 (6.9) years, 41% African American] with a depressive disorder but without diabetes enrolled at the Indiana sites of the Improving Mood-Promoting Access to Collaborative Treatment (IMPACT) trial. Incident diabetes cases were defined as diabetes diagnoses, positive laboratory values, or diabetes medication prescription, and were identified using electronic medical record and Medicare/Medicaid data. Surprisingly, the rate of incident diabetes in the collaborative care group was 37% (22/59) versus 28% (17/60) in the usual care group. Even though the collaborative care group exhibited greater reductions in depressive symptom severity (p = .024), unadjusted (HR = 1.29, 95% CI: 0.69–2.43, p = .428) and adjusted (HR = 1.18, 95% CI: 0.61–2.29, p = .616) Cox proportional hazards models indicated that the risk of incident diabetes did not differ between the treatment groups. Our novel preliminary findings raise the possibility that depression treatment alone may be insufficient to reduce the excess diabetes risk of depressed, older adults.