Medicaid and Access to Care Among Persons With Systemic Lupus Erythematosus

Abstract

Objective

To evaluate the associations between Medicaid insurance and distance traveled by patients to treating physicians and health care utilization for patients with systemic lupus erythematosus (SLE).

Methods

A total of 982 adults with SLE were recruited between 2002 and 2004. We calculated the distance between patient homes and physicians using Mapquest, an Internet mapping program. We then assessed the association between Medicaid status and distance traveled to the primary SLE provider, presence of ≥1 physician visits, and the number of all physician visits, with and without adjustment for demographic and medical covariates.

Results

On an unadjusted basis, Medicaid patients traveled longer distances to see their primary SLE provider. This effect was pronounced for patients under the care of a rheumatologist. Adjustment reduced, but did not eliminate, these differences. With adjustment for covariates, Medicaid patients were equally as likely to see a rheumatologist as non-Medicaid patients. However, Medicaid patients were more likely to be seen by a general practitioner or in the emergency room for their SLE, and reported more visits to general practitioners and the emergency room for SLE.

Conclusion

Medicaid patients with SLE traveled longer distances to see an SLE physician, especially rheumatologists. They also reported a different pattern of health care utilization. These results suggest that Medicaid patients may face barriers in obtaining comprehensive medical services in proximity to their residences.

INTRODUCTION

Medicaid, one of the nation’s largest insurance programs, covers approximately 55 million persons at an annual cost of ~$300 billion. Established in 1965, Medicaid aims to provide equal access to health care for those in poverty. However, some studies suggest declining physician participation in the program, thereby limiting access to enrolled patients, especially for those requiring specialists (1). In 2003, Bindman et al demonstrated that in California there were 9 specialists per 100,000 Medicaid patients as compared with 25 specialists per 100,000 patients with other types of insurance (2). Similarly, a 2001 California report showed that only 55% of medical specialists reported treating Medicaid patients in their practice and less than half stated that they would accept a new patient with Medicaid insurance (3). The effects of Medicaid status on access may be particularly pronounced for those with chronic illnesses who depend on the care of a specialist.

Four studies have examined access to specialty care in the Medicaid population and all suggest a disadvantage for those with Medicaid coverage. In 2001, a study by Skaggs et al conducted in Los Angeles County found that a timely orthopedic appointment was provided to 100% of children with private insurance and only 2% of those with Medicaid insurance (4). In 2006, a nationwide version of the study by Skaggs et al found that 38% of orthopedic practices provided limited or no access for Medicaid patients (5). These authors also noted that lack of access correlated with low physician reimbursement (5). In 2004, Wang et al (6) surveyed 100 California otolaryngologists in a clinical scenario and found that 97% would offer an appointment to a child with private insurance compared with 27% if the child was covered by Medicaid. Reasons provided for not offering an appointment included excessive paperwork and/or administrative burdens (97%) and low monetary reimbursement for the office visit (87%). Finally, Resneck et al (7), in a national sample, found that Medicaid patients were less likely to receive a dermatology appointment (32%) compared with patients with Medicare (85%) or other insurance carriers (87%). Resneck et al also noted that wait time until scheduled appointments was longer (50 days versus 37).

Systemic lupus erythematosus (SLE) is a chronic auto-immune condition of unknown etiology that primarily affects women of childbearing age. No cure exists for SLE, and patients face lifelong periods of disease flares with resultant inflammation of numerous organ systems. Although somewhat uncommon in the general population (1 in 2,000 persons in the US), SLE is more common in minority populations such as African Americans and Hispanics. These ethnic groups are disproportionally represented among those eligible for Medicaid on the basis of income even in the absence of active disease. In addition, many patients with SLE can meet the income eligibility criterion by “spending down” their resources as a result of high medical costs.

Prior research in rheumatoid arthritis demonstrates that care by a rheumatologist may result in reduced disability, less pain, and better overall functional status compared with care by generalists (8,9). Although similar studies have not been completed in SLE, the complexity of the disease may necessitate care by a rheumatologist or by another health care professional experienced in dealing with the complex array of manifestations.

Patients with SLE who have Medicaid coverage may face additional barriers to access if they live in a nonurban location. Prior literature suggests that patients in rural areas may be less likely to be seen by a rheumatologist for their musculoskeletal symptoms (10). In one Veteran’s Administration study not specific to SLE (11), longer distance to care was correlated with increased utilization of primary care and emergency services, perhaps indicating that lack of geographic access may promote substitution of available care for specialty care.

Few studies have examined the effects of insurance status on access to appropriate care for rheumatology patients (10,12–14), and, to our knowledge, no studies have examined access to care for patients with SLE. In our study, we calculated the distance to primary lupus providers in a large established cohort of patients with SLE and examined the association of Medicaid insurance with distance traveled to care. We hypothesized that the low physician enrollment in Medicaid may require Medicaid patients with lupus to travel longer distances to obtain appropriate specialty medical care. We also compared the likelihood of an outpatient visit within the past year and the average number of outpatient visits between participants with Medicaid insurance alone and those with other types of coverage to determine how distance to care may affect patterns of health care utilization.

PATIENTS AND METHODS

Patients

For this analysis, we used data from the first year of the University of California San Francisco (UCSF) Lupus Outcomes Study (LOS), an ongoing longitudinal survey of 982 English-speaking patients with SLE. Details regarding creation of the cohort are described elsewhere (15,16); however, relevant aspects are summarized here. Approximately two-thirds of the initial LOS cohort were derived from nonclinical sources such as lupus support groups and conferences, newsletters, Web sites, and other forms of publicity. Twenty-two percent of patients were recruited from academic rheumatology offices, and 11% from community rheumatology offices. Two-thirds of the participants reside in California, whereas the remainder reside in 40 states in the US. All LOS participants were confirmed to have SLE after chart review completed by a rheumatologist or a registered nurse working under a rheumatologist’s supervision. The UCSF Committee on Human Research approved the study protocol. All participants provided informed consent prior to the interview.

Data

The principal source of data for the LOS is a structured 1-hour telephone survey conducted by trained interviewers. The survey includes well-validated items covering the following domains: demographics and socioeconomic status, status of SLE, disability, general health status and social functioning, employment status, psychological and cognitive status, health care utilization, medications, and health insurance coverage (15).

Measures

Distance to primary SLE provider

Our primary outcome in this analysis was the number of miles traveled to see primary SLE-treating physicians. Patients identified their primary SLE provider and indicated the specialty of the physician. Patient mailing addresses as well as physician addresses were provided. For patient addresses with post office boxes, actual residence addresses were obtained for the purposes of geocoding and mapping. We verified the address and specialty of each primary SLE physician using the American College of Rheumatology directory, Internet resources, and telephone calls to physician practices. The distances between patient homes and primary SLE physician were then calculated using Mapquest, an Internet-based mapping tool that provides the most reasonable travel route (www.mapquest.com). We excluded 47 persons from our analysis due to inability to map this distance: no lupus provider/physician could be identified for 24 patients, patient or physician address could not be found for 13, and patient and physician were located in noncontiguous states for 10. For 5 individuals with difficult to use addresses (Mapquest could not locate the street address), zip code centroids were used.

To control for distance to SLE provider in the utilization analysis, we used the calculated distance variable. Because of the skewed nature of this variable, we truncated the variable at the third standard deviation above the mean. This limited the mileage for 21 patients (2.0%) whose physicians were >150 miles away from their residences. Distances were adjusted for 7 patients with Medicaid only and 14 participants in the comparison group.

Demographic/socioeconomic variables

Predictors of interest included socioeconomic and demographic characteristics, urban versus nonurban status, and SLE disease status. The socioeconomic and demographic characteristics included age at interview (in years), sex, race/ethnicity (white as referent, Hispanic, African American, and other), and education (high school education or less as referent, some college, college graduate, and postgraduate degree).

SLE-specific variables

SLE-specific variables included duration of disease (in years), age at diagnosis (in years), self-reported SLE flare in the last 3 months, and measure of SLE activity (scale from 1 to 10, with 10 representing more severe activity). We also created another measure of severity from questionnaire items. For this measure, disease was considered severe if patients required major immunosuppressive medications (mycophenolate mofetil, cyclophosphamide, methotrexate, intravenous steroids, chlorambucil, or azathioprine) or experienced major end-organ manifestations within the 2 years prior to the interview (kidney disease requiring biopsy, dialysis, or transplantation; bronchial or open-lung biopsy; hemoptysis or venous thromboembolism).

Health care utilization

The health care utilization section of the questionnaire asks participants about their medical care over the past 12 months. It included an enumeration of all health care practitioner visits by specialty and whether the visits were for SLE or another indication. Similarly, the section also included information about emergency room use and notes if the visit was for SLE or for another indication. To address the skewed distribution of these variables, we truncated the values for each variable at the third standard deviation above the mean, resetting the upper limit of the variable, rather than dropping these potential outliers from the analysis. To ensure correct interpretation, we also completed these analyses using 1 + log of response and found the results to be unchanged.

Health insurance

The insurance section, derived from the Medical Expenditures Panel Survey (17), included items regarding the type of health plan (health maintenance organization versus fee-for-service) and source of coverage (employment based, individually purchased plan, or public program). We chose to initially examine the characteristics of 3 insurance groups: Medicaid only, dual coverage (Medicare plus Medicaid), and other insurance. However, to best investigate access for individuals with only Medicaid insurance, for our analysis we chose to dichotomize the health insurance variable to represent those with only Medicaid insurance versus all other forms of insurance (Medicare, Medicare plus Medicaid, employer based, independent, or Veteran’s Affairs) because there is evidence that physicians accept patients with Medicare coverage in a fashion similar to privately insured patients (7). We excluded a small number of patients (n = 15) who reported having no health insurance and 1 individual who did not respond to this item because of inadequate power to draw reliable conclusions from this group.

All patient addresses were geocoded by Sonoma Technology Inc. (Petaluma, CA), assigning each participant to a census block. Consequently, information about neighborhood characteristics could be obtained and used for analysis. Using the density of the census block group, we designated each LOS participant’s residence as urban (higher local density) or nonurban (lower local density).

Statistical analysis

Distance to care

We used linear regression to assess the association between Medicaid status and distance traveled to the principal SLE provider, with and without adjustment for age, ethnicity, education, disease severity, and urban/nonurban locale. The same model was completed for the subset of patients (~75%) with a rheumatologist as their primary provider.

Health care utilization

We used univariate and multivariate linear and logistic regression, where appropriate, to assess the association between Medicaid status and various health care utilization measures, including whether visits to a specific type of practitioner had occurred within the last year and the number of visits within the last year, with and without adjustment for the same covariates as were used in distance to care analysis. To determine if the effect of Medicaid only status differed by poverty status, educational level, African American race, and Hispanic ethnicity, we created interaction terms for Medicaid and each of these characteristics. STATA software, version 8.0 (StataCorp, College Station, TX) was used for all statistical analyses.

RESULTS

A total of 58 patients (5.9%) reported Medicaid as the only source of insurance coverage compared with 84 persons (9.1%) with dual coverage (Medicare plus Medicaid) and 778 (85%) with other types of coverage (Table 1). The majority of individuals identified a rheumatologist as the primary SLE physician (74% of those with Medicaid insurance only, 68% with dual, and 81% with other insurance). As expected, because of Medicare coverage of end-stage renal disease, patients with dual coverage were much more likely to report a history of dialysis or transplant, and consequently these patients were more likely than the other groups to report a nephrologist as the primary SLE physician (11% versus 7% with Medicaid only and 4% with other insurance; P < 0.05).

Table 1.

Demographic and medical characteristics of study participants by insurance status^*

Characteristic	Medicaid only (n = 58, 5.9%)	Medicare plus Medicaid (n = 84, 9.1%)	Other insurance (n = 778, 85%)†
Demographics
Age, mean ± SD years‡	40.7 ± 11.0	46.4 ± 12.0	47.6 ± 13.4
Female sex	87.9	95.2	90.6
Working‡	20.6	11.9	51.9
Ethnicity
White‡	46.6	52.4	68.9
Hispanic	12.1	6.0	10.2
African American	17.2	28.6	5.5
Asian	8.6	8.3	10.0
Native American	1.7	1.2	0.4
Household income
<$20,000‡	62.0	64.4	8.5
$20,000–$40,000	30.0	27.4	19.9
$40,000–$60,000	8.0	6.9	21.9
$60,000–$80,000	–	1.4	16.8
$80,000–$100,000	–	–	12.7
>$100,000	–	–	20.2
Education
High school or less‡	53.5	27.4	18.1
Vocational/trade/some college	37.9	52.4	40.5
College	6.9	14.3	25.7
Postgraduate	1.7	6.0	15.7
Medical characteristics
Disease duration, mean years	10.7	13.3	13.0
Flare in last 3 months‡	62.1	50.0	44.8
SLE activity in last 3 months (1–10)‡	6.5	4.9	4.2
Severe disease	51.7	51.2	40.6
Taking oral steroids	50.0	54.8	44.5
Taking IV steroids in past year‡	18.5	6.2	7.8
IV cyclophosphamide in past year	5.4	3.8	1.9
Taking mycophenolate mofetil	5.4	11.0	8.5
Taking azathioprine	7.3	13.1	8.8
Organ manifestations
Kidney problems‡	50.0	51.3	38.5
Dialysis (ever)‡	17.2	40.9	16.0
Transplant‡	3.5	25.0	11.4
Lung problems (ever)	43.9	46.4	40.5
Urban locale	70.7	83.3	82.9
Primary SLE physician is rheumatologist‡	73.7	67.5	80.6

^*Values are the percentage unless otherwise indicated. SLE = systemic lupus erythematosus; IV = intravenous.

^†Includes employer based, independent, Medicare, and Veteran’s Affairs.

^‡P < 0.05.

Patients with only Medicaid insurance (hereafter referred to as Medicaid for simplicity) tended to be younger than patients with dual coverage or with other insurance (mean age 40.7 years, 46.7 years, and 46.4 years, respectively) and were more likely to live in a nonurban locale as categorized by census data. As might be predicted, patients with Medicaid coverage and dual coverage were less likely to be working and reported lower incomes than the group with other insurance. In general, the Medicaid group reported less formal education than the dual coverage group or the group with other insurance, although 37.9% of the Medicaid group reported attending trade school, vocational school, or some college.

In unadjusted analysis, the Medicaid and dual coverage groups reported higher SLE disease activity as measured by flare in the last 3 months and higher patient global assessments. Medicaid patients were more likely to have received intravenous steroids within the past year, but the 3 groups were equally likely to be taking oral steroids, cyclophosphamide, mycophenolate mofetil, or azathioprine.

In unadjusted analysis (Table 2), Medicaid patients reported traveling longer distances to see their primary SLE provider than those with Medicare, Medicare plus Medicaid, or other types of insurance coverage (mean distance 41.9 miles, 23.8 miles, 20.3 miles, and 24.3 miles, respectively; P < 0.05). For those who identified their primary SLE provider as a rheumatologist, Medicaid patients traveled longer distances than those with only Medicare, Medicare plus Medicaid, or other insurance (54.1 miles, 28.3 miles, 18.5 miles, and 26.9 miles, respectively; P < 0.05). Although the group with dual coverage more closely resembled the Medicaid group with respect to sociodemographic characteristics, for our final analyses we merged the dual coverage group and other insurance group to specifically examine access issues for the group with Medicaid as their sole insurance.

Table 2.

Average distance in miles traveled to systemic lupus erythematosus (SLE) physician by Medicaid status^*

	Any SLE physician†		Primary SLE physician is rheumatologist
	No. of patients	Miles traveled	No. of patients	Miles traveled
Medicaid only	58	41.9 ± 61.7‡	42	54.1 ± 68.2‡
Medicare plus Medicaid	84	20.3 ± 40.1	56	18.5 ± 19.4
Medicare only	130	23.8 ± 44.0	94	28.3 ± 50.7
Other insurance	648	24.3 ± 37.7	529	26.9 ± 40.0

^*Values are the mean ± SD unless otherwise indicated. Analysis of variance (ANOVA) for Medicare only, Medicaid and Medicare, and other insurance: P = 0.68. ANOVA for Medicare only, Medicaid and Medicare, and other insurance for those with primary physician as rheumatologist: P = 0.30.

^†Regardless of specialty (includes general practitioners).

^‡P < 0.05.

Adjustment for covariates such as SLE severity, age, ethnicity, urban status, and education reduced, but did not eliminate, the difference in distances traveled (Table 3). For all SLE providers, Medicaid patients traveled 11.5 more miles than those with other insurance, and for those seeing a rheumatologist, patients traveled 19.8 more miles (P < 0.05).

Table 3.

Multivariate linear analysis of distance (in excess miles traveled) to primary systemic lupus erythematosus (SLE) provider by Medicaid status^*

	Unadjusted	Adjusted† (main effect only)	Adjusted (with interaction for education and Medicaid status)
All SLE providers
Medicaid	18.1 (7.13, 29.0)‡	11.5 (0.63, 22.3)‡
Medicaid/education interaction
>High school education/other insurance			Referent
>High school education/Medicaid			25.1 (10.0, 40.1)‡
<High school education/other insurance			2.72 (−4.01, 9.44)
<High school education/Medicaid			0.79 (−13.4, 15.1)
Rheumatologist as primary SLE physician
Medicaid	27.6 (14.4, 40.9)‡	19.8 (6.92, 32.7)‡
Medicaid/education interaction
>High school education/other insurance			Referent
>High school education/Medicaid			31.6 (14.3, 48.9)
<High school education/other insurance			1.40 (−6.33, 9.11)
<High school education/Medicaid			8.37 (−9.26, 26.0)

^*Values are the number of excess miles traveled (95% confidence interval).

^†Adjusted for SLE disease activity, age, urban/nonurban locale, race/ethnicity, and education.

^‡P < 0.05.

We used flare in the last 3 months as a surrogate for SLE illness activity in our final model (Table 3). However, to determine if the results were sensitive to the choice of severity measure, we also modeled the results using our constructed measure of disease severity and patient global assessment (scale of 1 to 10) without significant alteration of results (data not shown).

A significant interaction between Medicaid only status, education, and distance to primary SLE provider was detected (P = 0.03). Specifically, Medicaid patients with higher than a high school education traveled longer distances to receive SLE care than those of similar educational background without Medicaid status and those with less than a high school education regardless of Medicaid status. This interaction was also seen in our subset group analysis of patients who reported a rheumatologist as their primary provider (P = 0.07) (Table 3).

For health care utilization, Medicaid individuals were equally as likely to have seen a rheumatologist within the past year (Table 4) as those with other types of insurance, and no significant difference existed in the number of rheumatology visits between the 2 groups (mean excess visits = 0.95 for Medicaid; 95% confidence interval [95% CI] 0.02, 1.88). There was also no difference between groups with regard to utilization of other health care practitioners such as nephrologists, gynecologists, and pulmonologists (data not shown). Although both groups were equally as likely to have had ≥1 visits to a general practitioner in the past year, Medicaid participants were more likely to see a general practitioner for SLE-related symptoms (odds ratio 2.63; 95% CI 1.50, 4.61) and reported a mean of ~5 more visits to the general practitioner within the past year for SLE-related issues (P < 0.05).

Table 4.

Health care utilization differences between persons with and without Medicaid coverage^*

	No. of visits, mean (range)	≥1 visits OR (95% CI)		Mean difference in no. of visits (95% CI)
		Unadjusted	Adjusted†	Unadjusted	Adjusted†
Rheumatologist	3.64 (0–16)	1.14 (0.58, 2.24)	0.71 (0.39, 1.49)	0.95 (0.02, 1.88)	0.36 (−0.58, 1.31)
Generalist for any indication	4.98 (0–30)	1.56 (0.75, 3.25)	1.60 (0.74, 3.49)	6.13 (4.39, 7.86)‡	5.54 (3.75, 7.43)‡
Generalist for SLE	3.65 (0–25)	2.63 (1.50, 4.61)‡	3.78 (1.50, 5.17)‡	5.01 (3.34, 6.67)‡	4.91 (3.20, 6.63)‡
Emergency room visits for any indication	1.19 (0–7)	3.65 (1.98, 6.71)‡	3.43 (1.80, 6.53)‡	1.22 (0.73, 1.71)‡	1.09 (0.58, 1.60)‡
Emergency room visits for SLE	0.74 (0–5)	3.00 (1.77, 5.11)‡	2.40 (1.36, 4.24)‡	0.89 (0.50, 1.30)‡	0.79 (0.35, 1.17)‡

^*OR = odds ratio; 95% CI = 95% confidence interval; SLE = systemic lupus erythematosus.

^†Adjusted for distance (in miles) traveled to SLE physician, race/ethnicity, education, urban/non-urban locale, age, and SLE disease activity

^‡P < 0.05.

Similarly, Medicaid patients were more likely to have had ≥1 visits to the emergency room for SLE within the last year and reported more visits to the emergency room for SLE (mean 0.89 more visits; P < 0.05). Our findings with respect to general physician and emergency room visits persisted when modeled for the subset of patients primarily cared for by a rheumatologist (data not shown). In multivariate analysis, controlling for SLE disease activity, age, education, race/ethnicity, urban status, and distance to SLE provider, the utilization patterns remained significant (Table 4).

DISCUSSION

Equal access to health care for Americans remains an implicit goal of public policy (18). Although prior studies have established that Medicaid patients may face barriers in obtaining health care, few studies have considered insurance status as a predictor for access to rheumatology subspecialty services and none have done so for persons with SLE. The present study was designed to assess whether persons with Medicaid travel longer distances to obtain care and whether they have different utilization patterns than those with other forms of coverage.

We found that patients who reported Medicaid insurance traveled longer distances to obtain SLE care by any health professional and especially when the primary SLE physician was a rheumatologist. Even though geocoding of patient addresses demonstrated that Medicaid patients were more likely to live in a nonurban locale, our results persisted when controlling for this potential confounder. Although patients with dual coverage (Medicare plus Medicaid) more closely resembled the Medicaid group demographically, the addition of Medicare seemed to predict shorter distances to care. This finding helps highlight the possible effect of insurance coverage as a predictor above and beyond the effect of sociodemographic status. Only a few (n = 15) individuals in our study were uninsured. Because the number was small, meaningful analysis could not be performed in this group alone. However, it is likely that this group also faces difficulty in obtaining care.

Although Medicaid patients were as likely to have seen a rheumatologist and reported a similar number of visits per year as those with other insurance, Medicaid patients reported more visits to general practitioners and to the emergency room for SLE than other participants. However, in multivariate analysis, distance to care did not fully explain this finding. There are several possible explanations for this observed trend. For example, the Medicaid group may in fact have less controlled disease that was not fully accounted for in our analysis, or a larger burden of comorbid illness. However, we believe a more complex explanation may be needed. Perhaps Medicaid patients, due to lack of social support or greater disease activity, require increased access to rheumatologists to adequately care for their SLE. Indeed, the increased utilization of available services, such as the emergency room, may reflect this lack of ready access to specialty care.

We observed an interaction between Medicaid status and education when examining distance traveled to care; those with Medicaid coverage and more than a high school education traveled the longest distances to see any type of SLE provider. Perhaps education afforded these individuals the ability to find housing outside of low-income urban areas, but Medicaid status limited them to urban health care providers such as public clinics or tertiary care facilities. Alternatively, Medicaid persons with SLE who have higher levels of education may have greater ability to access specialty health care, but at the price of having to travel longer distances.

Although the LOS provides extensive information on one of the largest SLE cohorts, the nature of this data source limits the generalizability of the results. Participants in the LOS are English speaking, generally insured (only 15 patients lacked insurance), and may underrepresent certain ethnic minority groups, such as African Americans and Hispanics (8.5% and 10% in the studied cohort, respectively), thereby limiting our ability to detect differential access based on ethnicity alone. Perhaps due to these limitations in our cohort, the number of Medicaid patients was likely lower than that within the true population of patients with SLE. Furthermore, although we attempted to adjust for disease status using several patient self-reported measures, this is likely an imperfect surrogate for detailed clinical information.

In addition, although recruitment for the LOS drew from both clinical and nonclinical sources, Medicaid patients were more likely to be recruited from academic practices (47% versus 32% for patients with Medicare plus Medicaid and 20% for those with other insurance; P < 0.05). This observation has several implications for the findings in our study. Patients recruited from university settings may be more likely to report specialist care, which may limit our ability to detect differential access to specialty care for Medicaid patients. In regards to distance traveled, patients at these settings may choose to seek care from a well-known institution or from a noted individual at the expense of longer travel distance. However, we believe that the most likely explanation of this finding is that university centers may be more willing to accept patients with Medicaid, and that patients face barriers to access within local community settings.

The influence of variable referral patterns and patient choice may also contribute to the findings in our study. In 2004, Losina et al (19) found that Medicare patients of low socioeconomic status as well as suburban or rural residence were more likely to use low-volume hospitals for elective total knee replacement surgery, even though high-volume centers were associated with better outcomes. In their study, low socioeconomic status may have reflected differential referral patterns for surgical care, and for individuals outside of urban areas choice of a low-volume center may have been influenced by shorter travel distances; however, neither fully explained patient choice.

Although patient choice may contribute to the findings in our study, our clinical experience supports the fact that a limited number of rheumatologists routinely accept Medicaid. However, we believe that studying this phenomenon in a rigorous fashion would be difficult because some rheumatologists may only accept current patients who become Medicaid enrollees or may only accept a new Medicaid patient under certain circumstances.

In this study, we used a novel method to assess access to care for rheumatology patients. Extensive data from our large SLE cohort also allowed us to comment on the possible implications of differential access on health care utilization. We found that Medicaid patients appear to travel longer distances to seek SLE care than patients with other types of insurance. This difference is particularly pronounced when patients report a rheumatologist as their primary SLE provider. The distances likely reflect a lack of SLE practitioners accepting Medicaid insurance, requiring individuals to look beyond local communities. We also found that Medicaid patients reported higher rates of utilization of general practitioners and emergency rooms, especially for SLE-related concerns. Although this was not fully explained by distance to care, we believe it may represent the downstream effects of differential access to care.

Our results generate multiple hypotheses surrounding access to care for Medicaid patients that can be approached in future studies. Although we demonstrated that Medicaid patients in the LOS travel longer distances for SLE care, future studies should investigate the potential downstream effects of differential access to care on health care quality and clinical outcomes. The results of such a study would enable further exploration of the relationship between socioeconomic status and clinical outcomes, and perhaps influence the development of public policy promoting access to care for less advantaged populations.