Patterns of Care

As health care analysts evaluate the quality of care provided to health care recipients, the continuum of care for a clinical condition can be described as a series of processes and subprocesses that correspond to specific events, such as a hospitalization or course of treatment for a respiratory infection in an ambulatory patient. Efforts to improve the quality of care examine the variation in these processes, then define changes in the process that improve quality and lower cost of care. As the processes of care are examined over large numbers of providers, similarities in the processes define models that can be analyzed for variation. These models are termed “patterns of care.”

The approach to medical care that evaluates patterns of care has been applied in a number of situations. For example, the National Cancer Institute's Surveillance, Epidemiology, and End Results program, initiated in 1973, provides evaluations and data for cancer diagnosis, treatment, and survival in the United States. The SEER program has developed several registries around the United States that collect and publish cancer incidence and survival data from 11 population-based cancer registries and three supplemental registries covering approximately 26% of the population. The database contains information on more than 3 million in situ and invasive cancer cases, with the addition of nearly 170,000 new cases each year. The database includes information on patient demographics, primary tumor site, morphology, stage at diagnosis, first course of treatment, and follow-up for vital status, and the data are used for a number of reports on patterns of care, mortality, and distribution of cancer in the areas in which the SEER project operates. The longevity of the project has produced exceedingly useful information on patterns of care for cancer in the United States, and the database is used by clinicians and health planners alike.

Patterns of care often help in evaluating the cost of care, by codifying processes of care and then dividing them into subprocesses that can have costs assigned to each step. For example, in a study of the treatment of asthma, a group of researchers in Switzerland (Szucs, Anderhub, & Rutishauser, 2000) compared patterns of care and associated costs between general practitioners and specialists. The technique is used frequently in research studies, but it has also been increasingly applied by organizations that pay for health care services and insurance companies to perform cost effectiveness studies and quality assessments. Using appropriate statistical approaches, patterns of care can be discerned and stratified in a number of ways, such as by provider, by institution, by diagnosis-related group (DRG), by level of disease severity, and many other variables. This type of information can provide support for decisions on which patterns of care provide the greatest value. The development of data mining and statistical pattern recognition techniques over the past decade has increased the utility of large databases in identifying patterns of care.

Data sources for patterns of care analysis are diverse. Insurers usually use claims databases that are created from transactions submitted by practitioners for payment. Some of the more sophisticated payers create data warehouses that combine several sources of data into a very large database, adding content and value to the care patterns analysis. Although transaction data are readily available, they often lack sufficient clinical depth to perform the types of analysis necessary to determine optimum patterns of care. Thus these studies often provide suggestions for further analysis of care that requires abstraction of data from clinical records at the site of care. If an electronic medical record is not available, then paper chart reviews must be conducted by experienced chart abstractors.

...