Project final report. Baltimore, MD, USA: Johns Hopkins University and Aetna Health Plans.
Johns Hopkins University and Aetna Health Plans, USA
As the US health care system reshapes itself, new demands are being placed on insurance companies to directly manage the delivery of care to their beneficiaries. Increasingly, data systems embodying clinical information are required for this purpose.
A new methodology for deriving health status information was developed by a team of researchers at the Johns Hopkins University. This state-of-the-art technique is known as the Ambulatory Care Group (ACG) case-mix adjustment system. The overall theoretical goal of ACG assignment is to categorize similar conditions based on their expected impact on health services resource consumption. ACGs incorporate an approach for clustering ICD-9-CM diagnostic codes to derive health status information from existing insurance claim data sources on the premise that a measure of a population’s “illness burden” can help explain variation in health care consumption. Individuals are categorized based on their age, gender, and diagnoses assigned by their providers during contact with the delivery system over a specified period of time, such as a year.
Potentially, ACGs have numerous applications within managed care health plans. The Ambulatory Care Case-Mix Development Project – reported on here – was a collaborative effort between Aetna Health Plans (AHP) and the Johns Hopkins University (JHU) to apply and evaluate the Johns Hopkins Ambulatory Care Group case-mix system within Aetna. During this project, claims data were obtained from three AHP HMO sites (Atlanta, Cleveland, and Minneapolis) and from Aetna’s national indemnity book-of-business. The main objective of the project was to evaluate how ACGs could be applied to provider profiling, setting premium rates, and adjusting capitation payments. Specifically, two key questions were answered:
(1) Does ACG adjustment improve provider (e.g., primary care physician) profiling relative to age/gender adjustment?
(2) How accurate are ACG-based actuarial analyses, compared with demographic and experience rating?