Episode-based payment: evaluating the impact on chronic conditions.

BACKGROUND: Policy makers are interested in aggregating fee-for-service reimbursement into episode-based bundle payments, hoping it will lead to greater efficiency in the provision of care. The focus of bundled payment initiatives has been upon surgical or discrete procedures. Relatively little is known about calculating and implementing episode-based payments for chronic conditions. OBJECTIVE: Compare the differences in two different episode-creation algorithms for two common chronic conditions: diabetes and coronary artery disease (CAD). STUDY DESIGN: We conducted a retrospective evaluation using enrollees with continuous coverage in a self-funded plan from 2003 to 2006, meeting Healthcare Effectiveness Data and Information Set (HEDIS) criteria for diabetes or CAD. For each condition, an annual episode-based payment was assessed using two algorithms: Episode Treatment Groups (ETGs) and the Prometheus model. PRINCIPAL FINDINGS: We began with 1,580 diabetes patients with a 4-year total payment mean of $67,280. ETGs identified 1,447 (92%) as having diabetes with 4-year episode-based mean payments of $12,731; while the Prometheus model identified 1,512 (96%) as having diabetes, but included only 1,195 of them in the Prometheus model with mean diabetes payments of $23,250. Beginning with 1,644 CAD patients with a 4-year total payment mean of $65,661, ETGs identified 983 patients (60%) with a 4-year episode-based mean of $24,362. The Prometheus model identified 1,135 (69%) as CAD patients with 948 CAD patients having a mean of $26,536. CONCLUSIONS: The two episode-based methods identify different patients with these two chronic conditions. In addition, there are significant differences in the episode-based payment estimates for diabetes, but similar estimates for CAD. Implementing episode-based payments for chronic conditions is challenging, and thoughtful discussions are needed to determine appropriate payments.

Effect of multiple chronic conditions among working-age adults.

OBJECTIVE: To determine the longitudinal effect on healthcare costs of multiple chronic conditions among adults aged 18 to 64 years. STUDY DESIGN: Retrospective cohort assessment of working-age employees and their dependents with continuous coverage in a self-funded health plan from January 1, 2004, to December 31, 2007. Data were obtained from health benefit enrollment files and from medical and pharmacy claims. METHODS: Individuals were defined as having chronic conditions based on modification of a published method. The mean annual healthcare costs were estimated for individuals with 0, 1, 2, 3, 4, or 5 or more chronic conditions. The probability of persistence in high-cost categories across years was estimated for individuals in each group. RESULTS: Overall, 75.3% of working-age adult enrollees had at least 1 chronic condition, 54.3% had multiple chronic conditions, and 16.5% had 5 or more chronic conditions. The cost of healthcare was higher among individuals with more chronic conditions for all ages. The mean medical cost per year for an individual with no chronic conditions was $2137, while that for an individual with 5 or more chronic conditions was $21,183. Enrollees with more chronic conditions had higher persistence in high-cost categories between years and persisted at these high costs for more years. CONCLUSIONS: While multiple chronic conditions are common in the population 65 years and older, they are also of great concern for the working-age population. Understanding how to effectively manage individuals with multiple chronic conditions is an important challenge. Effective care management focused on managing the patient as opposed to a condition has the potential to significantly affect healthcare costs.