Monitoring and Evaluating Medicaid Fee-for-Service Care Management Programs: User's Guide
Appendix 2. Glossary of Evaluation Terms
Baseline data: Starting data that can be used for comparing data collected after the intervention begins.
Comparable data: Data from another population or location that contain similar characteristics to those of the group undergoing the intervention. These data are used when it is not possible to select a reference group from within the study.
Cohort evaluation: Follows the reference and intervention group over time.
Confounding variable: A factor that cannot be controlled or measured that may influence the outcome of the evaluation.
Control group: Also known as a reference group. A group not receiving the care management intervention or services. This population is used as a comparison group with those receiving the intervention. Selection of the control group depends on the design of the care management program.
Cross-sectional evaluation: Conducted at a single point in time; after the program has been implemented.
Longitudinal evaluation: Examines the differences in intervention outcomes at several points in time both before and after the implementation of a care management program.
Matched control: A population that is similar to the intervention group chosen to serve as a reference group.
Pre/post design: The population receiving the intervention is used as its own reference group. Characteristics of the group before and after the intervention are compared.
Randomized control trial: Participants are arbitrarily assigned to the reference or treatment group. Randomized control trials analyze program effectiveness and are considered the scientific gold standard because they control for any potential biases in the groups that may unfairly influence the evaluation.
Regression to the mean: A phenomenon that is similar to the "law of averages." In care management, for example, a group of people who have high medical costs in one year will tend to have average costs that are considerably lower in the following year.
Risk stratification: A statistical process in which the care management and reference groups may be divided according to disease status (or case mix) that allows for separate comparisons to be made for people in different disease states.
Staged implementation: When a care management program is rolled out in several phases. Staged implementation can be selected to designate a reference group by using those who have not yet received services as the reference (or control) population.