Background: Traditional diagnostic criteria for depression and anxiety fail to account for symptom severity. We previously evaluated a severity-based classification system of mood and anxiety symptoms. This study examines whether those severity groups are predictive of differences in health care utilization.
Methods: We used a cohort design to compare the health care utilization of 1232 subjects classified into 4 groups according to symptom severity. Health care billing data were evaluated for each subject for a 15-month period around the index visit. Multiple linear regression models were used to examine relative contributions of individual variables to differences in health care utilization. Analysis of variance procedures were used to compare charges among the severity groups after adjusting for demographic and medical comorbidity variables.
Results: After adjustment, significant differences in health care utilization between groups were seen in all but 3 of the 15 months studied. Also, after adjustment, the presence of a mood or anxiety disorder influenced utilization for only a 6-month period. At 9 to 12 months, subjects in the high-severity group showed a more than twofold difference in adjusted charges compared with the low-severity group ($225.36 vs $94.37).
Conclusions: Our severity-based classification predicts statistically and clinically significant differences in health care utilization over most of a 15-month period. Differences in utilization persist even after adjustment for medical comorbidity and significant demographic covariates. Our work lends additional evidence that beyond screening for the presence of mood and anxiety disorders, it is important to assess symptom severity in primary care patients. Further study directed toward developing effective methods of identifying patients with high levels of mood and anxiety symptom severity could result in significant cost savings.