Abstract
Purpose: To improve a case management (CM) program using the principles and tools of epidemiology. Specifically, to use epidemiology to describe the population being managed, to analyze factors influencing outcomes, to assess the degree to which the CM process (or intervention) is related to those outcomes, to utilize the findings in order to make recommendations (to take action) for both better evaluation and improved and more efficient CM process, to provide an estimate of the impact of the CM program based on the comparison of pre-CM and post-CM interventions, and finally to discuss the caveat that pre-CM period probably does not provide a prediction of patterns to be expected in the post-CM period had CM not been present (an ideal, but difficult-to-find referent population for this kind of effectiveness analysis).
Primary practice setting(s): A Medicare Advantage health plan.
Findings: There were a total of 12,185 individuals who met the continuous enrollment requirement of 6 months (28 days) prior to initial contact with the CM department and 6 months after the contact date: 53% were female; the average age was 73.9 (standard deviation = +/-9.5). There was a linear relationship between the average "dose" of CM-as measured by the number of times a case manager had contact with a case-and the risk profile of the case-as measured by a standard risk assessment tool provided by the Centers for Medicare & Medicaid Services. The month-to-month costs for the study population prior to CM showed a sharp rise in the month prior to the initial contact with CM and a sharp decline beginning before the contact. This pattern was consistent across different risk profiles and our operational definition of CM "dose."
The average costs 6 months prior to CM were higher than the average costs 6 months after the CM. The difference in cost varied by "dose" category. When coupled with the number of cases per dosage category, the greatest value for the CM program was in the management of moderate risk cases called two to four times.
However, some of the overall decline that begins prior to the initial CM contact is likely attributed to "regression to the mean" (i.e., costs may have shown a decline in spite of CM) but not all. Subsequent studies will be designed to assess the degree to which this is the case by including an equivalent referent; ideally, one that has not experienced CM or, in the absence of that, one that experienced a different kind of CM so that a valid "comparative effectiveness" study can be conducted.
Implications for CM practice: The "dose" of CM to its cases was in synch with an independent assessment of risk of the cases. This implies that case managers were directing their resources to those in need. However, case managers and CM processes are a limited resource and it is of interest for management to allocate those CM resources in the most efficient way possible. Methods of assessment based on individual experiences of case managers can be improved with structured, population-based assessment. These population-based tools, according to the principles of epidemiology, will be used to better allocate CM resources for optimized impact on patient populations in the future.