Table 1 repeats the findings from the JAMA article. Consistent with previous studies, these findings show that uninsured persons were about twice as likely to report some difficulty getting care (30.8 percent) as privately insured persons (14.6 percent) and persons with Medicaid or other state health insurance coverage (16.3 percent), and more than three times as likely to report difficulty as Medicare beneficiaries (9.5 percent).
However, there was variation across the 12 case study communities in the proportion of uninsured persons who reported difficulty obtaining health care, from a high of about 36 percent in Cleveland to a low of around 22 percent in Newark and Orange County. While these site-specific estimates are adjusted to control for other individual and family characteristics that are also correlated with access to care, the amount of variation across all 60 communities in the study actually changed very little even after controlling for these factors.
Since about 90 percent of uninsured persons who reported difficulty obtaining health care cited the cost of care or lack of insurance as the main reason they experienced difficulty, the measure essentially reflects "financial" barriers to care for uninsured persons. Access to care is also influenced by other considerations that may not be directly related to economic factors, such as whether an individual has an identifiable source of primary care, the proximity of health care providers to an individual’s place of residence or work, the ease and convenience of seeing health care providers and the appropriateness of the care setting where services are received. The rest of this report examines how communities vary on other measures of access to care for uninsured persons.
Having a particular place to go to when a person is sick or needs advice about his or her health can facilitate entry into the health care system and has also been shown to be associated with better preventive care, more timely and appropriate use of care, greater satisfaction and lower cost6. Consistent with other studies, data from the Community Tracking Study Household Survey show that uninsured persons nationally are much more likely to be without a usual source of care (32.1 percent) than persons with insurance coverage, including those with Medicaid and other public insurance (Table 2). In addition, uninsured persons are only about half as likely as privately insured and Medicare beneficiaries to use a physician’s office as their usual source of care, and-along with Medicaid beneficiaries-are more likely to use hospital-based facilities and other clinics.
There is also considerable variation across communities in whether uninsured persons have a usual source of care. For example, about 46 percent of uninsured persons in Phoenix do not have a usual source of care, compared with less than one-fourth of uninsured persons in Cleveland, Indianapolis and Greenville. The type of place that uninsured persons use for health care also varies. For example, uninsured persons in Boston and Little Rock are more likely to use hospital-based facilities (e.g. emergency departments and outpatient clinics) as their source of care. By contrast, uninsured persons in Lansing, Indianapolis, Greenville, and Newark are more likely to use physicians’ offices, while the uninsured in Syracuse and Seattle are more likely to use clinics and health centers, such as neighborhood or community health centers.
Despite this considerable variation, there is no clear correlation between the level of difficulty that uninsured persons report in getting health care (as reported in Table 1) and the type of place they go primarily for their care, if any. Communities in Table 2 appear in the same order that they appeared in Table 1 (i.e., ordered from high to low in terms of the level of reported difficulty), and a visual scan of the results shows that sites with the highest levels of reported difficulty in getting care (e.g., Cleveland, Lansing) are generally not the same sites with the highest percentage of uninsured persons who do not have a usual source of care. Only Miami had both a relatively high percentage of uninsured persons reporting difficulty getting care and a high percentage of uninsured persons without a usual source of care. Although the correlation across all 60 communities between the proportion of uninsured persons reporting difficulty and the proportion without a usual source of care was positive, it was small (r = .10) and not statistically significant.
There was a slight correlation across sites between reported difficulty and the proportion who used physicians offices as their usual source of care (r = -.22). Thus, reported difficulty tends to be slightly less in sites where a high proportion of uninsured persons use physicians offices as their usual source of care, although this correlation was only marginally significant (p < .10). Correlations across the 60 sites between reported difficulty and other sources of care were small and not statistically significant.
Regardless of where people go for their care, access can be more difficult to the extent that appointments are harder to make, it takes a longer time to get to the physician, and people have to spend longer times in the waiting room before seeing him or her. When uninsured persons see a physician, they are less likely than insured persons to make appointments in advance and more likely to walk in without an appointment (Table 3).7,8 This reflects the fact that they are more likely to use hospital-based facilities and other clinics that more typically see patients on a walk-in basis.
While walking in does not necessarily imply worse access relative to having an appointment ahead of time, some may be deterred from seeking care if it means having to wait a long time in a physician’s office or clinic. In fact, uninsured persons spend a longer time waiting in physicians’ offices on average, compared with Medicare and privately insured persons (35 minutes vs. 23 minutes), which in part reflects the fact that they are less likely to have scheduled appointments. On the other hand, uninsured persons have only slightly longer travel times than insured persons in getting to a physician.
There is little discernible variation across sites in whether uninsured persons make appointments when seeing a physician. Time spent getting to the physician varies somewhat more, with uninsured persons in Lansing averaging about 17 minutes, and uninsured persons in Miami averaging about 25 minutes. The average waiting time in a physician’s office for uninsured persons also varies, from 26 minutes on average in Phoenix to almost 50 minutes in Greenville.
As with usual source of care, there was little apparent correlation across communities between levels of reported difficulty obtaining care and the convenience of care. With the exception of Miami, travel times and office waiting times were not any longer in the communities with high levels of reported difficulty (which are ordered first in the table), nor were they necessarily much lower in communities with low levels of reported difficulty (which are ordered last in the table).
One important caveat to note from this table is that the results are reported only for persons who had a physician visit in the previous year. Since uninsured persons are less likely to visit physicians than insured persons-which may be due to their having less access to physicians-it is likely that the differences with insured persons on these measures are understated. In addition, the variation across communities on these measures might also be understated to the extent that access to physicians also varies across communities for uninsured persons.
Differences in the use of health care services among population subgroups have also been used to draw inferences about access. Uninsured persons have consistently been found to use fewer services of all types relative to insured persons, which is almost always interpreted as indicating lower access to care.9 In terms of ambulatory care visits, uninsured persons average fewer visits in a given year, compared with insured persons (Table 4). However, there is a considerable amount of variation across communities in ambulatory care use by the uninsured, from two visits a year in Miami and Orange County to well over three visits a year in Syracuse and Greenville.
The extent to which the level of reported difficulty in getting health care should correlate with ambulatory health care use is unclear. Since fewer visits imply less access, one might expect higher levels of difficulty to be reported in communities where there were fewer visits. On the other hand, it could also be argued that more frequent contact with the health care system will induce greater unmet medical needs for uninsured persons, as they find it difficult to obtain the medications, treatments, specialized services and high-tech procedures prescribed by physicians (i.e., services they may not otherwise have been aware they needed). Thus, higher levels of reported difficulty in getting care in a community might actually be correlated with greater numbers of ambulatory care visits.
In fact, our findings provide little evidence to support either hypothesis. There is very little correlation between the level of ambulatory care use in the community and the level of difficulty in getting care. Low levels of ambulatory care use by the uninsured can be observed both in communities that had low levels of reported difficulty in obtaining care (e.g., Orange County) and communities with relatively high levels of reported difficulty (e.g., Miami). Given the lack of correlation between utilization and reported difficulty, it is likely that difficulty getting care is experienced not only when trying to gain entry into the health care system, but also when trying to obtain medicines, treatments and procedures and follow-up visits that are prescribed over the course of an episode of care. Simple counts of the number of visits may obscure difficulties encountered in obtaining these more specific types of services.
The pattern of health care use-especially the extent to which hospital emergency departments (EDs) are used for nonurgent health problems-may also have important implications for problems that uninsured persons encounter in obtaining health care. The proportion of ambulatory visits at hospital EDs is considerably higher for uninsured persons (about 17 percent) than for the privately insured (8 percent) and Medicare beneficiaries (6 percent). There is some variation across communities in the proportion of ambulatory visits to hospital EDs by uninsured persons, from a low of about 14 percent in Orange County to a high of 26 percent in Boston.
We also find that there is a correlation across the 60 communities between the level of reported difficulty and visits to hospital emergency departments. That is, the proportion of ambulatory visits to hospital EDs tends to be higher in communities where reported difficulty is also higher (r = .33, p < .05). Since hospital EDs are often viewed as the providers of last resort in a community, this suggests that uninsured persons end up in the hospital ED more frequently when they encounter difficulty getting care in other primary care settings. Conversely, uninsured persons who rely more heavily on hospital EDs for their health care may find it more difficult to obtain other necessary care and services that are not provided directly in the hospital ED.
The frequency with which uninsured persons use the hospital ED for their care is closely related to the notion of usual source of care. As one would expect, use of EDs for ambulatory care tends to be higher in communities that rely more heavily on hospital-based facilities to serve as a usual source of care for the uninsured (r = .29, p < .05). By contrast, communities with higher percentages of uninsured who use health centers and clinics as their usual source of care tend to have lower levels of hospital ED use (r = -.28, p < .05). In sum, the types of places that serve as sources of primary care for uninsured persons in a community are related to the rate at which hospital EDs are used for ambulatory care, and this may have consequences for the level of difficulty that uninsured persons report in obtaining health care.
6 Starfield B. Primary Care: Concept, Evaluation, and Policy. New York: Oxford University Press, 1992.
7 All of the estimates reported in Table 3 refer to the "last physician visit" if it occurred within the 12 months prior to the interview. Averaged across all individuals within a population or within a population subgroup, responses to questions based on the last physician visit are assumed to be representative of individuals’ encounters with physicians.
8 The survey also included questions on how long individuals had to wait for an appointment if an appointment was made. However, sample sizes for uninsured persons were too small (and therefore not statistically reliable) to permit reporting for individual communities.
9 Andersen RM, Davidson PL. "Measuring Access and Trends" in Changing the U.S. Health Care System, Andersen RM, Rice TH, Kominski GF, eds. San Francisco: Jossey-Bass, 1995. Kasper JD. "Asking About Access: Challenges for Surveys in a Changing Healthcare Environment." Health Services Research, Vol. 33, No. 3, Part II: 715-740 (1998).