Conclusion

The measures examined in this study get at different aspects of access to care, including financial and nonfinancial barriers, and some of the steps involved in the health care-seeking process. These measures were largely consistent in showing that access to care for uninsured persons is generally worse than for insured persons, and that there is substantial variation in access for uninsured persons across communities. However, the measures did not show a consistent pattern of variation across communities for uninsured persons. Thus, communities that had a relatively high percentage of uninsured reporting difficulty getting care were not necessarily the same communities where uninsured persons were less likely to have a usual source of care, had longer waiting and travel times to the physician’s office and had less health care use overall.

What accounts for the lack of consistency in the pattern of variation across communities? Part of the reason may be that the community-specific sample sizes for uninsured persons are relatively small, especially for those communities that are not among the 12 case study sites. This results in some fairly high confidence intervals around individual estimates, which may attenuate the correlation across communities between measures.

Aside from methodological issues, the lack of correlation between measures demonstrates that access to care does not consist of a single concept (which can be tapped into by a variety of measures), but that a more appropriate view of access is that it represents a process or series of steps in care-seeking. According to this view, problems encountered at any specific stage of care-seeking do not necessarily imply that problems will be encountered at prior or subsequent stages (although they often are). For example, uninsured individuals living in poor urban areas may find that public hospitals and community health centers are convenient and accessible places to obtain medical care. However, excessively long waiting times to see a physician and the inability to obtain certain specialty services may result if services for medically indigent persons are rationed in these facilities because of resource constraints. Thus, having a close and identifiable source of care does not necessarily guarantee that all necessary services can be obtained easily.

This is particularly a problem for uninsured persons, since their underlying lack of financial access to care may outweigh or negate some of the other nonfinancial aspects of care-seeking. For uninsured persons, cost or lack of insurance is by far the most frequently cited reason given for delaying or not getting needed medical care (90 percent), and these financial barriers may occur at any step during the care-seeking process, from trying to get into the system (e.g., for a examination and diagnosis) to obtaining highly specialized services or procedures, once they are in the system. Thus, the measure of reported difficulty in getting care (Table 1) essentially reflects financial barriers to care that were experienced, which may or may not be related to measures that primarily reflect nonfinancial barriers to care, such as the usual source of care and the convenience of getting care. In fact, having a usual source of care or a short travel time to get to the physician may matter very little if an individual has difficulty paying for needed services.

In addition, the lack of consistency may also be related to the fact that the pattern of variation across communities on some measures reflects more general problems of access to care in the community (i.e., not just the uninsured population), while access problems based on other measures are more specific to the uninsured population. For example, there was no correlation in the proportion reporting difficulty between the uninsured and privately insured across all 60 communities (r = .04). The correlation between the uninsured and privately insured was higher on other measures, including the proportion with no usual source of care (r = .82) and travel time to the last physician visit (r = .42). Thus, the specific experiences that uninsured persons have in not being able to obtain needed care appear to be unique to them in a given community, while variations in usual source of care and travel time reflect in part more generic differences among communities.

Nevertheless, the findings showed a modest correlation across communities between reported difficulty for uninsured persons and their rate of ED use, which has implications for both policy and care delivery. Much of the concern about the use of hospital EDs as a source of primary care has focused on the higher costs to the health care system compared with nonurgent care provided in community health centers or private physicians’ offices. Since hospital ED use is associated with greater difficulty in obtaining care by uninsured persons, these findings suggest that reliance on hospital-based facilities-particularly hospital EDs-as sources of care for medically indigent persons results not only in higher costs but lower access.

Although the pattern of variation across communities is different, depending on the specific measure of access used, the implications of these findings concerning the expansion of health insurance coverage are similar. In the article that appeared in JAMA, simulations were conducted that showed that expanding private or public health insurance coverage would not only reduce substantially the amount of difficulty that uninsured persons experience in getting medical care, but also would result in greater uniformity across communities in the level of difficulty that uninsured persons currently experience. Similar simulations were also conducted for other selected measures (e.g., percent with no usual source of care, number of ambulatory visits), and the conclusions from these results were virtually identical to those stated above. Furthermore, simulation results for all of the measures consistently showed that access to care for uninsured persons could be improved more effectively by expanding health insurance coverage rather than by improving access to a level that is similar to communities that currently provide the so called best access to care for uninsured persons.

In sum, improving access to care for uninsured persons through the current system of direct delivery of services by safety net providers is complicated by the fact that most communities are neither consistently poor or consistently good on all aspects of access to care for uninsured persons. This not only makes it difficult to rate communities in terms of how well they provide services to the medically indigent, but finding a specific mechanism by which access for the uninsured can be improved is not straightforward. Expanding insurance coverage, on the other hand, will lead to improved access on virtually all measures, and will also reduce the importance of community as a major factor in determining access to care among uninsured persons.