Background: New patient-centered models of ambulatory care aim to substitute better primary care for preventable acute care within existing primary care practices. This study aims to identify whether mental illness and other characteristics of primary care patients are related to risk for an acute event for an ambulatory care-sensitive condition (ACSC).
Methods: We conducted a 2-year, longitudinal analysis comparing ambulatory care-sensitive admissions and emergency department (ED) visits for a cohort of 18,526 primary care patients followed in 5 veterans affairs (VA) primary care sites. We compared rates, risks, and costs of ACSC-related acute events during a follow-up year for patients with and without mental illness seen during the previous year in primary care.
Results: The 12-month rate of ACSC admissions was 31.7 admissions per 1000 patients with mental health diagnoses compared with 21.0 admissions per 1000 patients without (P=0.0009). The ACSC-associated ED visit rate was also significantly higher (P<0.0001). In adjusted analyses controlling for demographics, chronic disease, illness severity, and prior ambulatory care, those with depression or drug use disorders had higher odds of receiving ACSC-related acute care (odds ratio=1.10, 95% confidence interval: 1.03, 1.17 for depression; odds ratio=1.48, 95% confidence interval: 1.05, 1.99 for drug use disorders). Costs per admission and ED visit were similar across patient groups. Higher medication use and lower medication regimen complexity were significantly associated with decreased risk for ACSC events.
Conclusions: Prior mental health diagnoses and medication use were independent risk factors for ACSC-related acute care. These risk factors require focused attention if the full benefits of new primary care models are to be achieved.
*Health Economics Resource Center, VA Palo Alto Health Care System, Menlo Park
†Center of Excellence for the Study of Healthcare Provider Behavior, VA Greater Los Angeles, North Hills
‡UCLA School of Public Health, Los Angeles
§Department of Ambulatory Care, VA Greater Los Angeles, North Hills
∥UCLA School of Medicine, Los Angeles
¶RAND Corp, Santa Monica
#Health Research and Policy Department, Stanford University, Stanford, CA
The authors declare no conflict of interest. Funding for this study was provided by the Department of Veteran Affairs, Veterans Health Administration, Patient Care Services, the VA Improvement and Assessment Laboratory (Project XVA 65-018).
Reprints: Jean Yoon, PhD, Health Economics Resource Center, VA Palo Alto Health Care System, 795 Willow Rd, 152 MPD, Menlo Park, CA 94025. E-mail: email@example.com.