Multivariate Analysis of Patterns of Informal and Formal Caregiving among Privately Insured and Non-Privately Insured Disabled Elders Living in the Community. C. The Relationship between Insurance Status and Undermet ADL Neet


Much of the current research on long-term care is focused on characterizing the disabled population, identifying the type and level of services needed and used, and estimating public and private expenditures on care. Given the development of both longitudinal and cross-sectional databases, researchers have also tracked trends in disability rates and service use over time. Yet, there has been almost no examination of how well disabled individuals actually function in the community. Although the number of dollars spent on home care has been increasing, this does not address the question of whether, at a given point in time, needs are fully served. We do not know, for example whether the service delivery system is adequately responding to the care needs of disabled elders living at home.

In the section that follows, we examine the relationship between insurance status and the perception of the performance of the formal and informal service system. More specifically, we focus on whether the service system is leaving people with a feeling that their ADL needs are being met. To measure this, we focus on the concept of Undermet Need. An undermet need is present when an individual indicates that he or she could use more help or believes that he/she had to wait too long to receive the help. The presence of an undermet need in ADLs may reflect the fact that caregivers are not available in a timely manner. This can result from scheduling difficulties or because, once in the home, caregivers are not providing the care when it is needed.

Figure 3 shows that the rate of reported undermet need is about 1.4 times greater among privately insured claimants than among non-insured disabled elders. This difference is statistically significant at the .05 level.

  FIGURE 3: Proportion Reporting Undermet ADL Needs by Insurance Status  
Bar Chart: Privately Insured Claimants (24.7%); Non-Privately Insured Disabled (17.1%).
SOURCE: 1999 National Claimant Study (n= 595 claimants); 1994 NLTCS (n=1,348).

Given the differences in the characteristics and service use patterns of the two populations, until we control for relevant variables it is difficult to isolate the relationship between insurance status and the higher rate of reported undermet ADL needs. As shown in Table 6, there are significant differences in relevant characteristics between the two groups on almost all variables that one might expect to be related to reported levels of undermet ADL needs. For example, individuals in the insured sample reporting undermet ADL needs are more disabled (i.e. greater ADL and IADL impairments), have more ADL helpers, receive more hours of formal care, and have higher income levels than do those in the non-insured sample. Individuals in the non-privately insured sample reporting undermet ADL needs have more medical diagnoses, are more cognitively impaired, are more likely to perceive their health status to be poor, and are more likely to have informal caregivers and receive more hours of informal care.

  TABLE 6. Characteristics Associated with Reported Undermet ADL Needs by Insurance Status  
Characteristics   Privately Insured  
  Non-Privately Insured  
Average Age
     Percent over age 80  
78 years
80 years
Percent Female 68% 69%
Number of ADL Limitations 4.1*** 3.8
IADL Limitations 7.4* 7.1
Perceived Health to be Poor 29%*** 53%
Cognitive Impairment 42%*** 69%
Number of Medical Diagnoses .6*** 1.2
Presence of Informal Caregiver   85%*** 97%
Number of ADL Helpers 2.3*** 1.7
Weekly Hours of Formal Care 32.0*** 15.3
Weekly Hours of Informal Care 38.9*** 53.9
Average Income $36,918*** $14,741
Income Greater than $30,000 47%*** 6%
ANOVA F-statistic significance level: *** .01; ** .05; * .10.

Undermet ADL need is a binary variable for the existence of any reported undermet need. Table 7 reports the results of a logistic regression relating undermet need to a variety of independent variables. While each variable showing a significant difference in Table 6 was initially included in the equation, some were dropped because of the high degree of inter-variable correlation. Also, to aid in the interpretation of results, certain variables like age were transformed from continuous to categorical variables.

Table 7 shows that, of ten variables tested, nine were statistically significant at the .10 level or greater. Individuals over age 80 are 2.74 times more likely to report undermet ADL needs than are those under age 80. Moreover, ADL dependency is also positively related to the probability of reporting undermet ADL needs. An interesting result, however, is that when age and disability status is interacted, the effect on reported undermet need is negative. This suggests that while the probability of reporting undermet need increases with both main effects, age and disability, this is moderated by the interaction term. Put another way, the probability of reporting undermet need increases with age and disability, but at a decreasing rate. Those reporting the presence of cognitive impairment are also less likely to report undermet ADL needs than are their cognitively intact counterparts.18

TABLE 7. Logistic Regression Model for Estimating the Probability of Reporting Undermet ADL Needs
Variable   Coefficients     Standard  
  Exp (B)  
Percentage over age 80 1.0087** .3925 2.74
Number of ADLs .2698*** .0828 1.31
ADLs interacted with Age -.2935*** .1008 .75
Self-assessed health
(0=excellent, good or fair)
.3425** .1786 1.41
Cognitive Impairment
-.3121* .1820 .73
Number of ADL Helpers .5666*** .0967 1.76
Number of Hours of Informal Care   .0061*** .0024 1.01
Number of Hours of Formal Care -.0046* .0028 .99
Has LTC Insurance .2303 .2316 1.26
Income -.0156** .0051 .98
Constant1 -2.5525*** .3329 -------
Significance Level * p=.10; ** p=.05; *** p=.01. (n=717).
Nagelkerke R2 = 22.4%.
  1. The constant is adjusted to reflect the underlying probability of reporting Undermet ADL need among all sample respondents -- 19.5%.

Another surprising finding is that, as the number of ADL helpers increase, the probability of reporting undermet ADL needs also increases. After age, this variable has the largest impact on the dependent variable.

This suggests that, when multiple caregivers are involved in caring for an individual, certain needs may "fall through the cracks". In the absence of clearly delineated lines of responsibility and accountability across caregivers, there is a much greater chance that individual ADL needs will not be attended to properly. Alternatively, it may be that individuals with more needs have more helpers, and there is a greater probability of having an undermet need as one's needs increase.

Whether care is provided formally or informally is also an important determinant of whether someone is likely to report undermet ADL needs. As shown in Table 7, as the number of hours of informal care increase, the probability of reporting undermet ADL needs also increases. In contrast, however, as the number of hours of formal care increase, the probability of reporting undermet ADL needs decreases. This suggests that when relatives or friends provide informal care, it may be perceived as inadequate. If this perception is valid, it would follow that training informal caregivers would better prepare them to respond to the multiple needs of their disabled care receivers.

Finally, increasing income levels are associated with lower levels of reported undermet ADL needs, and having LTC insurance is not related to the dependent variable at all. Even though the level of reported undermet need is higher among the privately insured sample, after allowing for the influence of age, disability status, service utilization patterns and income among the two groups, having LTC insurance in and of itself does not influence whether or not one's ADL needs are met.19


Clearly, long-term care insurance is succeeding in bringing formal caregivers into the homes of disabled elders, allowing many to remain in the community. The majority of claimants do not report undermet needs. Yet, somewhat surprisingly, individuals with multiple caregivers and those who receive a great deal of care informally are most likely to report undermet ADL needs. It is reasonable to assume that those who receive the most care also need it the most. For them, anything less than round-the-clock care may be perceived as a shortfall. Given that most insurance policies have daily limits on the amount of care that can be reimbursed, it may be that such policy limits lead to undermet needs for the most seriously impaired.

Another issue relates to care coordination. Findings presented here suggest that clear lines of responsibility vis-à-vis coordination of care may not always be present. Also, informal caregivers may not be adequately trained or responsive to meeting the needs of their disabled relatives. Both of these findings imply a need for greater coordination of care and for more training of informal caregivers. Similarly, although greater levels of formal care reduce the probability of reporting an undermet ADL need, problems with service availability and scheduling persist, and this contributes to undermet ADL need.

While both age and disability status are positively associated with the probability of reporting undermet need, the effects moderate somewhat among the oldest and most disabled individuals. This suggests that the insurance is helping to address the needs of the group most "at risk" for potentially having to access more costly institutional care. Finally, the presence of private LTC insurance does not influence whether or not someone reports undermet ADL needs. Even though the probability of reporting undermet ADL needs is greater among the privately insured, when the influence of other variables is taken into account, insurance is not a relevant factor. This supports the proposition that service delivery issues rather than insurance policy design are the most important factors underlying individuals' sense that their ADL needs are not being adequately met.

View full report


"multanal.pdf" (pdf, 597.33Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®