Vernon L. Greene

Substitution between formally and informally provided care for the impaired elderly in the community.

This article assesses the extent to which formally provided comprehensive community care tends to substitute for informal care provided by family and friends to impaired elderly persons living in the community. Using simultaneous equations causal modeling techniques to control for selective targeting and other intervening factors, results indicate a substantial tendency for formally provided care to be substituted for informal care. Other results indicate that unmet need appears to be the major variable predicting both informal and formal support levels, with informal care providers appearing to be somewhat more precise in conditioning support levels on need than formal care providers. Policy implications of these findings are discussed.

Do Community-Based, Long-Term-Care Services Reduce Nursing Home Use? A Transition Probability Analysis

This study offers logit estimates of the probability of transition from the community to a nursing home based upon data from the National Long-Term-Care Demonstration. It is found that nurses deter entry by those using a wheelchair while home-health aides deter entry for those with cognitive impairments. Personal-care aides and housekeepers reduce admission risk for those with severe functional disabilities. These findings suggest that appropriate targeting of community-based services would improve the degree to which they offset nursing home expenditures.

Prospective Budgeting for Home Care: Making Titration Work

Objective: To provide a model for prospective budgeting for home care that is plausible, coherent, flexible, and sufficiently tractable that it can serve as a template for practical decision making and to clarify what would be the data requirements and statistical framework to calibrate the model.Methods: Methods used are standard risk-neutral expected value theory, cost benefit analysis, and the conditional logistic probability model.Results: A simple but effective prospective budgeting model that provides analytic scaffolding for a practical decision support system for home care case managers, consultants, and program evaluators that can improve program equity, efficiency, and effectiveness.Discussion: The author criticizes the well-known Titration Budgeting Model of Weissert, Chernew, and Hirth in terms of its logical and operational problems but then goes on to develop a framework within which the goals of the titration model can be met and home care resources can be more efficiently allocated.

Inconsistency in level of care assignment decisions in skilled nursing facilities.

In Arizona, a non-Medicaid state, we investigated the extent to which unregulated level of care assignments in Skilled Nursing Facilities consistently reflect level and nature of patient impairment. Using Multiple Discriminant Analysis (MDA) to develop optimal prediction functions, approximately 70 per cent of patients could be correctly classified. Factors identified by MDA as discriminating among patients at different levels of care are an Activities of Daily Living (ADL) impairment factor, and a factor defined by confusion, transitory contact with the social environment, and propensity to wander. Results are compared with those of studies using MDA to replicate Multidisciplinary Review Team (MRT) or other expert level of care assignments intended to develop patient classification functions for clinical use. MRT assignments appear to reflect patient impairment characteristics only slightly better than do unregulated institutional assignments, suggesting that such utilization review efforts may result in minimal net gains in appropriateness of placement.

Kinship Caregivers: Health and Burden

Grandparents and other relatives are raising more than seven million children in kinship care households in the United States. Kinship caregivers are increasingly sought out to provide care for children as a mechanism for preventing children from entering the formal foster care system. Kinship caregivers who were participants in the KinNET program were surveyed to assess their level of perceived burden, health, and experience in support groups. Satisfaction was high for the support groups and facilitators. The independent variables in the model account for 43% of the variability in the caregivers burden score. The effect of the number of hours of care provided is negligible in magnitude and statistically insignificant.

Simultaneous optimization of factor assessibility and representativeness

P. M. Bentler has shown that Raos canonical factor analysis is in effect a psychometric analysis, leading to factors that are maximally assessible from the data. He contrasts this with Kaiser and Caffreys alpha factor analysis that leads to factors that maximally represent the true factors in the content domain. Noting the problems associated with factors that may be highly assessible, but not very representative, or vice versa, Bentler suggests the need for a technique that would, insofar as possible, be optimal with respect to both criteria. Such a technique is presented here, and is shown to resolve into a traditional scaling method, which in turn acquires a richer psychometric interpretation.

Effectiveness of an abstinence-only intervention sited in neighborhood community centers

Choices Enhanced is a community-based demonstration and evaluation project targeting inner-city youth, ages nine to 15 at pretest, in Syracuse, New York. The projects evaluation design is a randomized field trial, and the intervention uses an education and youth development curriculum designed to deter unmarried teen pregnancy. Findings reported here are from early in the third year of the project. The intervention was designed to provide each treated subject 24 hours of intervention contact. The projects central goal is to delay the onset of sexual intercourse among unmarried-youth subjects. This was to be achieved through: (1) strengthening commitments to the values of abstinence and the importance of marriage; (2) informing youth more fully with respect to the risks and consequences of youth sexual intercourse outside marriage; (3) strengthening the ability to resist social pressures to have intercourse; and (4) encouraging restraint with regard to future behavior. Hypotheses tested specifically address the goals listed above, using linear and logistic regression methods. The magnitude of estimated treatment effects was modest, and change was sometimes not in the expected direction. At a 0.05 significance level, none of the estimated intervention effects were found to be statistically significant.

Kinship care and issues in permanency planning

ABSTRACT Kinship caregivers are a child-care resource for families experiencing stress or temporary parenting due to illness, incarceration, or death of a parent. This article examines whether and how felt caregiver burden influences the reported propensity of caregivers to want to adopt the children in their care. Kinship caregivers who were enrolled in KinNET completed the survey (N = 102) and the data were entered anonymously into SPSS for analysis. Their mean age was 57.51 years (SD = 10.13), 95% were female (SD = .19), and two-thirds were non-white (SD = .73). Using “likelihood of adopting the child in my care,” as the outcome variable in the linear regression analysis, caregiver’s age, monthly income, and total hours employed were significant predictors. Total pressures, family service needs, and physical problems scales were not statistically significant predictors. The adjusted R square was .439 and significant (.006). Understanding the factors that are predictive of adopting children in kinship care will help programs target services more effectively. Helping kinship caregivers and the children in their care is also important in promoting their health and social well-being.

Response by Greene

I very much appreciate the thoughtful attention that my colleagues have given to my work on titration budgeting, and I think their critiques and admonishments serve to further clarify and, in some cases, correctively qualify a number of the points made in the article. In what follows, I respond to each set of comments separately. Although Hirth, Chernew, and Weissert (2005) acknowledge the value of my extension of the titration model, for which I am grateful, they have chosen not to respond substantively to my critique of their implementation of it—indeed, they say that the problems I raise were already “recognized and acknowledged” (p. 431) and refer to my argument, perhaps a bit dismissively, as being “overstated for effect” (p. 431). I take this remark to mean that they feel I am overstating the substantive significance of problems, or that I dwell on them at excessive length, or both. Let me address these possible interpretations of their observation in turn. My point in highlighting their “constant effectiveness” (Hirth et al., 2005, p. 431) assumption in particular was not simply that it is unrealistic—that much was clear and had indeed been recognized and acknowledged by them. Rather, the problem I point out is that in fact they apparently did not recognize, and certainly fail to acknowledge, its full logical consequences. Among these consequences, as I show in the article, is that the computed titration budgets that they have offered in several published articles on the topic, which require this assumption for their calculation, in fact do not have the key properties they claim for them. In particular, these budgets have no rational basis as prescribed spending targets, and they cannot be shown to be costeffective precisely because they confuse the cost of producing an effect with the monetized value of the effect. With all due respect, as these properties provide the main rationale and motivation for these budgets, this cannot be dismissed as a minor matter that has been overstated for effect.

A mixed estimator interpretation of ridge regression

Abstract It is shown that a formal isomorphism between the ridge estimator and the homogeneous case of the Theil-Goldberger mixed estimator leads to a general interpretation of ridge regression as ordinary least squares estimation subject to a prior stochastic constraint that all slope coefficients in the model are zero. Users of ridge regression are cautioned to consider this implied specification in their use and interpretation of the technique.