Identification of Potentially Avoidable Emergency Department Visits Using Claims Data

In recent years, some states have established all-payer claims databases to collect health care claims data from all insurers across an entire state. These databases provide a rich source of data for analyzing utilization, cost, and quality of health care and for evaluating the effectiveness of delivery system reforms to reduce cost and improve quality. It is widely understood that the reduction of potentially avoidable ED visits represents an important cost-savings opportunity. However, unlike the standard methods for identifying inpatient ambulatory care sensitive conditions (ACS), there is currently no nationally recognized standard method for identifying potentially avoidable outpatient ED visits.

We have developed a set of diagnostic categories for potentially avoidable ED utilization (ED visits for conditions that are non-emergent and could be more cost effectively treated in other settings) and have applied this method to analysis of claims data. Using the all-payer claims databases in northern New England states, we have identified specific ICD-9 codes to identify conditions, such as sore throat, strep, viral infection, anxiety, conjunctivitis, ear infection, upper respiratory infection, bronchitis, asthma, dermatitis and rash, joint pain, lower back pain, muscle and soft tissue pain, fatigue, and headache. This presentation will describe the development of these methods. It will also provide examples of applications of these methods to provide policy-relevant information in northern New England states.