reports

ACGs: an ambulatory care case-mix measure. Health Quality section update

Center for Health Quality Outcomes and Economic Research. Bedford, MA, USA: CHQOER Quarterly, ENRM VA Hospital.

Center for Health Quality Outcomes and Economic Research, Bedford, MA, USA

The Quality Section at CHQOER was recently awarded a HSR&D grant to evaluate the feasibility of using existing ambulatory care case-mix measures in the VA to predict resource utilization. Given the recent emphasis on managed care in the VA, the need to assess health care utilization and manage resources effectively has become critical.
The measures that will be tested are Ambulatory Care Groups (ACGs), a methodology developed by health services researchers at Johns Hopkins University in the late 1980’s, and Diagnostic Cost Groups (DCGs) developed collaboratively by researchers from Boston University and Brandeis University. This article focuses on one of these measures. A recent conference hosted by the developers of ACGs highlighted the growing interest among researchers and the private sector in using the measure as a risk adjustment tool particularly for reimbursement and profiling.
ACGs are based on the premise that a measure of illness burden, or case-mix, can help explain variation in health care resource consumption (i.e., costs and visits). ACGs represent a population-based approach to case-mix measurement. This approach focuses on the patient as the unit of analysis, rather than the visit or episode. A single ACG is assigned to an individual based on their age, gender, and their constellation of diagnosis codes. ACGs are formed using 4 steps: first, each of the patient’s ICD-9-CM codes during a defined period of time are placed into one of 34 clusters, called ADGs, based on expected resource utilization. Second, similar ADGs are further collapsed into 12 ADGs called CADGs. Third, based on the individuals CADGs, he/she is placed into 1 of 25 mutually exclusive major ambulatory categories (MACs). Finally, based on age, gender, presence or absence of certain individual ADGs, and number of ADGs, individuals are categorized into 1 of 51 mutually exclusive ACGs. Version 4.0 of ACGs is currently available and contains several revisions that are designed to reduce heterogeneity within both the ADG and ACG categories.
ACGs have many diverse applications, including use as a capitation or financial risk adjustment system, use as a utilization classification system, use as a proxy health status measure for quality or outcomes research, and use in physician profiling. ACGs were developed using data from three HMOs and a Medicaid program, so that both the ACG groupings and case-mix weights (derived for each ACG) are based on these original populations. In adapting this measure to the VA, it is expected that the new case-mix weights and further partitioning of individual ADGs will be necessary in order to derive a case-mix measure that accurately reflects the morbidity of the VA population. In fact, preliminary analyses using version 3.0 show that VA patients are heavily weighted towards the more complex ACGs, each of which are composed of four or more separate ADGs. This work was presented in June 1997 at the annual meeting of the Association for Health Services Research.