Differential Impacts Among Subgroups of Early Channeling Enrollees Six Months After Randomization

U.S. Department of Health and Human Services

Differential Impacts Among Subgroups of Early Channeling Enrollees Six Months After Randomization

Executive Summary

Thomas W. Grannemann, Randall S. Brown and Shari Miller Dunstan

Mathematica Policy Research, Inc.

July 1984

This report was prepared under contract #HHS-100-80-0157 between the U.S. Department of Health and Human Services (HHS), Office of Social Services Policy (now the Office of Disability, Aging and Long-Term Care Policy) and Mathematica Policy Research, Inc. For additional information about the study, you may visit the DALTCP home page at http://aspe.hhs.gov/daltcp/home.htm or contact the office at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, SW, Washington, DC 20201. The e-mail address is: webmaster.DALTCP@hhs.gov. The DALTCP Project Officer was Robert Clark.

This report, which serves as a supplement to Channeling Effects for an Early Sample at 6-Month Followup, Kemper et al. (1984), addresses the issue of variation in channeling impacts among subgroups of sample members. The objective of this investigation is to determine whether estimated overall channeling impacts mask important differences in impacts among subgroups. The results indicate the types of clients for whom channeling impacts were greatest and can help identify groups that might benefit most from future channeling type programs.

The sample and outcome measures used in this report are the same as those used in Kemper et al. (1984). As such, the outcome measures come from tracking data and 6-month followup interviews with early channeling sample members or proxy respondents. Impact estimates were calculated using multiple regression methods to control for the effects of other determinants of outcomes, including the effects of other subgroup variables. Subgroups investigated are defined by level of impairment, incontinence, income, insurance coverage, and living arrangement.

Because each subgroup contains a smaller number of observations than the full sample, our ability to detect subgroup impacts is less than for overall impacts. For this reason failure to observe a statistically significant difference among subgroups for an outcome measure should not be interpreted as conclusive evidence that no difference exists. On the other hand because the number of statistical tests involving subgroups is very large, it is likely that some effects that meet conventional criteria of statistical significance will, in fact, be chance occurrences. Given these limitations, the subgroup results should be viewed as very tentative evidence, useful primarily for raising issues for further investigation when additional data become available.

The overall finding of this investigation is that, where channeling impacts exist, there is not strong evidence that those impacts differ markedly among various types of clients. We have little evidence that outcomes produced by channeling are distinctly different for any identifiable subgroup of clients. For the most part, differences were not highly significant statistically and did not produce consistent patterns of differences across outcome measures and subgroups. There are, however, a number of outcome measures for which we found differences in impacts among subgroups. Among the findings, to which the above mentioned caveats apply, are:

• There is some limited evidence that some channeling impacts differ systematically among ADL impairment subgroups. In the basic model, channeling had a differential impact on hospital admissions, with reductions for the less impaired and increases for the extremely severely impaired. This difference was not observed in the financial control model.

• In the financial control model, there appears to be a correspondence between ADL impairment and the type of in-home formal services for which impacts were observed. In-home medical care hours were increased for the very severely impaired; personal care hours were increase for the very severely impaired; personal care hours were increased for those mildly impaired; and housekeeper/other hours were increased for those with no ADL impairment. No such correspondence is apparent for the basic model.

• In the financial control model we observed a reduction in total hospital and nursing home days for the Medicare-nonMedicaid subgroup, an effect that is not observed for Medicaid eligibles.

• There is evidence that the financial control model effect of increasing use of community based services did not extend to those with income over $1,000 per month.

• There is evidence that the basic case management model reduced use of some community based services by clients with income over $1,000 per month.

• There is some evidence that impacts differed between the continent and incontinent subgroups, with suggestive evidence that the incontinent experienced somewhat more beneficial impacts on well-being measures.

• There is suggestive evidence that the financial control model impacts (increases) on use of in-home formal services were smaller for those living alone that for those living with a spouse, child, or others.