It is not difficult to imagine that certain groups of consumers might adapt to consumer direction more quickly or be more satisfied with it than others, although there is little theory or empirical evidence upon which to base such predictions. For example, consumer direction might be too burdensome for beneficiaries (or their representatives) if the beneficiary is very ill or has severe cognitive or functional limitations. On the other hand, those with an ample network of unpaid helpers might find it easier to hire workers. It is also possible that consumers getting a relatively more generous allowance would be more satisfied with the program. In addition, as program staff gain more experience and the program matures, it may run more efficiently or its rules may change, thus affecting consumers' satisfaction with the program and care arrangements.
To explore possible differences in the way groups of consumers experienced IndependentChoices, we conducted regression analysis on a few key outcomes: (1) whether the consumer started receiving the allowance within three months of enrolling,27 (2) whether the consumer found it difficult to hire a worker, (3) whether the consumer found program spending rules restrictive, (4) how satisfied the consumer was with the program, (5) how satisfied the consumer was with overall care arrangements, (6) whether the consumer had an unmet need for personal care, and (7) whether the consumer disenrolled voluntarily from the program (Appendix Table A.24, Table A.25, Table A.26, Table A.27, and Table A.28).
Only a few consistent patterns emerged. Consumers with one or more unpaid caregivers at enrollment tended to be more satisfied with IndependentChoices and with their overall care after nine months and were less likely to disenroll voluntarily after a year, perhaps because it was easier for them to hire workers. Such consumers were also less likely to report that the program spending rules were restrictive. Consumers with more than six hours of planned care each week at enrollment, and thus, relatively larger allowances, were more likely to start receiving the allowance within three months, were more satisfied with IndependentChoices, were less likely to report an unmet need for personal care nine months after enrollment, and were less likely to disenroll from IndependentChoices voluntarily. (The more generous allowance may have made it easier to attract and hire a worker and thus start receiving the allowance relatively quickly.) On the other hand, consumers for whom a proxy respondent answered most questions in the baseline interview, and who thus might have been in poorer health or might have had more cognitive limitations, were less likely to start receiving the allowance within three months, more likely to report unmet need for personal care at nine months, and more likely to disenroll voluntarily. Finally, consumers who enrolled later in the demonstration (in 2000 or 2001) were more likely to start receiving the allowance within three months of enrolling. However, the later enrollees were also less likely to be highly satisfied with IndependentChoices. Those who came forward for the program in its earlier days may have been more dissatisfied with agency care and may have been in situations that better supported consumer-directed care than those who came forward later.
A few characteristics also seemed to be strongly related to starting on the allowance quickly and remaining in the program, perhaps suggesting a strong interest in consumer direction on the part of consumers and caregivers. Consumers who did not have publicly funded home care at enrollment were more likely to start receiving the allowance within three months and were less likely to disenroll. Consumers dissatisfied with their paid care at enrollment were also less likely to disenroll. Those consumers who had an informal caregiver at baseline who was interested in becoming a paid worker were more likely to start receiving the allowance within three months and were less likely to disenroll voluntarily.
We also used regression analysis to investigate the independent effect of hiring a family member on the seven key outcomes described above. For five outcomes hiring a family member had no effect; however, those who hired family were less likely to have found hiring hard and more likely to be very satisfied with overall care arrangements (not shown). The lack of effect for most of the outcomes analyzed may reflect the fact that of consumers who did not hire family, most hired acquaintances, many of whom had been providing care before they were hired.