June 2009
Reviewer: Jeff C. Huffman, MD
Healthcare costs associated with depression in medically ill fee-for-service Medicare participants
Unützer J, Schoenbaum M, Katon WJ, et al
J Am Geriatr Soc 2009; 57(3):506-510
Background: Depression in patients with medical disease is common and has been associated with negative medical outcomes. In addition, depression in older adults has also been independently associated with increased healthcare costs in some, but not all, studies. No prior studies have evaluated the links between depression and healthcare costs in Medicare recipients with chronic illnesses such as diabetes or CHF.
Methods: The study sample was composed of 14,902 Medicare beneficiaries participating in a pilot Medicare care management program in 9 counties in Florida. This pilot program was designed for patients with significant chronic medical illness (diabetes and CHF) who had high healthcare costs under traditional Medicare. The authors evaluated these patients’ depression status and healthcare costs/Medicare claims over the 12 months before they were enrolled in the pilot program (this spanned November 2004-August 2006). Regarding depression status, the authors split the patients into those with depression (claim for ICD-9 diagnosis of depression), probable depression (positive depression screen using the PHQ-2 or use of antidepressant), or no depression.
Results: Participants with depression (N=2108) had higher healthcare costs than those with probable depression (N=1081) or no depression (N=11,713), and those with probable depression had higher costs than those with no depression. After adjustments for differences in demographic variables and medical illness (e.g., severity of medical illness using a comorbidity index), these significant differences in healthcare costs remained both statistically significant and substantial, with increases in overall healthcare costs of roughly 60-80% (depressed vs. nondepressed) across different levels of medical comorbidity, for an average of approximately $4000 greater costs per year in this high-utilizing population.
Discussion: This report confirms the association of depression with increased healthcare costs, independent of demographic factors or an index of medical illness severity, in older adults with CHF or diabetes. The study had limitations in its population (only high utilizers), depression diagnostic methods (claims data or PHQ-2 results), and potential for regional bias (only in Florida). In addition, there may have been other important factors (for which the authors did not account) that could have led to the apparent association between depression and costs. However, this was a study of a large and important population, and the authors did control for many key variables. Of import, only a small portion of the healthcare cost differences were attributable to mental health issues.
These results—when combined with extant data about the impact of depression on health-related quality of life, hospitalizations, and mortality in medically-ill patients—should prompt a continued push for both a clinical and research focus on the management of depression among patients with medical illness. Certainly depression screening and management protocols in medical settings are not easy to implement and have their own costs, but there is increasing evidence that ‘collaborative care’-type programs may serve as a very useful model. These programs utilize nurse care managers to assist with depression screening, patient education, and implementation of depression treatment, and these programs have led to improvements in depression outcomes in a cost-effective manner in outpatient medical centers—and in some cases have led to cost savings. Further study of these programs to determine their clinical feasibility may prove fruitful in reducing mood symptoms, healthcare costs, and even mortality rates among patients with depression and medical illness.