Discussion
A hallmark of California's IHSSs program is participants' freedom to employ individual personal assistance workers of their choosing. Because it may be difficult for program participants and families to find nonrelative providers on their own, worker registries are maintained at the county level as a resource for IHSS beneficiaries seeking nonrelative providers. Workers advertising their availability for hire via the registries have all passed criminal background checks. Nevertheless, many IHSS program participants, regardless of their age, tend to choose relatives rather than nonrelatives as paid caregivers. This tendency is especially notable among Asians and Hispanics. This may be related to "familist" values being more prevalent among these ethnic groups (about which we have no direct information), but it may also be associated with differences in household composition. Our data show that members of these ethnic groups are more likely to live in larger households and less likely to live alone.
The first study questions concerned the availability of and the decision to use a spouse as a paid personal assistance provider. Exercising a preference for relatives, including a spouse, as paid caregivers depends on family members' availability and ability to do the job. The use of nonrelative caregivers was more prevalent among program participants who lived alone. Only a minority of IHSS program participants had the opportunity to "choose" a spouse provider because most were currently unmarried (83% nonaged, 70% aged) or their spouses had been determined by IHSS social workers as not "able." The spouses present were often either IHSS recipients themselves or otherwise not physically/mentally able to be paid caregivers.
The other study questions investigated whether IHSS program recipients having a spouse as a paid personal assistance provider obtained at least comparable outcomes compared with recipients having other paid providers. The outcomes given focus were whether IHSS recipients remained at home and at comparable costs to the Medicaid program in terms of hospital admissions, nursing home use, and total Medicaid expenditures.
Provider Effectiveness
It is appropriate for policymakers to condition decisions about whether to honor individual and family preferences regarding the choice of paid caregivers on the impact of such choices on health outcomes. In general, recipients of the respective age groups with paid spouse providers had comparable or more limitations in ADL and cognitive functioning compared with those with other providers. There were comparable numbers of chronic health conditions within recipient age cohorts among most provider groups.
This research identified no program performance or outcome results that argue against honoring a program participant's selection of relatives, including spouses and parents, as their paid caregivers. In contrast, there were several reasons found in favor of honoring this choice. In the service use measures of greatest interest and relevance (i.e., rates of ambulatory care-sensitive hospital admissions and Medicaid-covered nursing home placements), IHSS recipients' choice of spouse, parent, other relative, or nonrelative caregivers either made no difference or the differences were statistically significant in the direction of more desirable outcomes for those who chose relatives as paid caregivers. Among nonelderly adults, those with parent paid caregivers and among persons aged 65 years and older, those with spouse paid caregivers had significantly fewer ACS hospital admissions and fewer nursing home placements. In no comparisons did those with spouse providers have worse outcomes than those with nonrelative providers. These comparisons were adjusted to control for severity of disability and medical conditions and other risk factors.
Consistent with these utilization findings, average monthly Medicaid expenditures (for all services) were lower for IHSS recipients with family providers (either a paid spouse or other relative) compared with those with non-relative providers. Medicaid expenditures differences were not statistically significant for IHSS recipients with paid parent providers compared with those with nonrelative providers. These comparisons include but are not solely the result of lower IHSS expenditures for those with spouse providers.
The effects on Medicaid expenditures suggest possible parallel effects on Medicare expenditures. Further research (requiring data unavailable for this study) would be necessary to determine whether Medicare costs were lower as a result of less medical care use, particularly lower ACS hospital admissions among dually eligible IHSS recipients who had paid parent providers (adults age 18–64 years) or paid spouse providers (among those aged 65 years and older).
Other areas where further research could be informative for practice and policy are the interrelationships between household composition, shifts in caregiving responsibility among family members, and the effects of paid care on unpaid caregiving. We found all these measures to be confounded with provider type. Persons with nonrelative providers, much more frequently than those with other provider types, live alone. Hispanics and Asians are proportionately much more likely to use other relatives and spouse providers than are Whites and Blacks. Further complicating matters is that when someone who lives alone becomes disabled or as they become more severely disabled, they may be less likely to live alone (unless they go to a nursing home). For example, if a spouse caregiver dies or begins to also need care, there is a likelihood, especially among low-income families, that household composition will change. One option is that an adult child will bring the disabled parent(s) into his or her home or will move into the parental home to provide care. We adjusted for provider type and household composition in the multivariate models using baseline. Future studies measuring changes in these attributes over time will refine understanding of the dynamic pattern of assistance and whether it is enhanced by the use of paid providers, particularly spouse and other relative providers.
Conclusions
The findings from this research can provide useful information to policymakers in states where there is still ongoing debate about whether to authorize payments to family caregivers, especially spouses (Simon-Rusinowitz, Martinez Garcia, et al., 2010). The low number of program recipients likely to have a spouse who is able and available to be a paid provider should assuage worries about stimulating induced demand. However, the likelihood of healthy spouses becoming paid caregivers is further reduced by Medicaid income and asset eligibility criteria. In California, state plan personal care service policy does not allow a couple to separate their income and assets in determining Medicaid eligibility for home care. This contrasts with the separation of income and assets that are available for couples in many of HCBS waivers and when one partner is a nursing homes resident. Relaxing this eligibility restriction could possibly expand the participation of spouses as paid home care providers. The upper limit on this is constrained by the number of recipients with available and able spouses.
Although there is some elasticity in the potential participation of spouses as paid providers, the finding of lower Medicaid expenditures for those with spouse providers suggests that substitution of paid for unpaid family caregiving and other cost shifting can be minimized. IHSS assessment and benefit allocation processes demonstrate how to support and reward the continued involvement of family caregivers. This is done by adjusting the amount of publicly funded help to take into account household maintenance tasks that relatives residing with the public program participant should be expected to do without pay. This practice has produced a system where personal assistance costs, holding recipient needs levels constant, tend to be lower among those with spouses and other relatives as paid providers and where recipients experience at least comparable and sometimes better outcomes than recipients with nonrelative providers.
In summary, these analyses found no financial disadvantage to Medicaid and some quality and cost advantages (particularly the much lower rate of preventable hospital stays among elderly IHSS recipients) from allowing spouses to be paid providers. This argues in favor of honoring recipient and family preferences for paid spouse providers.