Published online before print January 18, 2007, doi: 10.1161/01.STR.0000257312.12989.af
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Stroke. 2007;38:1010.)
© 2007 American Heart Association, Inc.
Original Contributions |
Disparities in Outcomes Among Patients With Stroke Associated With Insurance Status
From the Department of Health Care Administration and Policy (J.J.S.), School of Public Health, University of Nevada at Las Vegas, Las Vegas, Nev; and Aunt Martha’s Youth Service Center, Inc. (E.L.W.), Chicago Heights, Ill.
Correspondence to Jay J. Shen, PhD, Department of Health Care Administration and Policy, School of Public Health, University of Nevada at Las Vegas, 4505 Maryland Parkway, Las Vegas, NV 89154-3023. E-mail jie.shen{at}unlv.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background and Purpose— Despite well-documented discrepancies in many clinical conditions across insurance groups, limited research has examined insurance-related disparities for patients with stroke. This study examined the relationship between insurance status and hospital care for patients with stroke.
Methods— Discharges with intracerebral hemorrhage and acute ischemic stroke were abstracted from the 2002 National Inpatient Sample. Neurologic impairment status and mortality were examined.
Results— Compared with privately insured patients, uninsured patients had a higher level of neurologic impairment, a longer average length of hospital stay, and higher mortality risk. For patients with intracerebral hemorrhage and acute ischemic stroke, mortality risk of uninsured patients was approximately 24% and 56% higher, respectively, than that of their privately insured peers.
Conclusions— Policy should promote access to outpatient and preventive care for uninsured patients so risk factors such as hypertension can be detected and treated during early, asymptomatic stages. Further research is needed to evaluate the extent to which differences in outcomes are attributable to differences in severity level on admission.
Key Words: outcome • severity of illness • stroke • uninsured
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Lack of health insurance has increasingly become problematic for Americans.1 The percentage of uninsured, nonelderly Americans has grown steadily from 16.1% (39.6 million) in 2000 to 17.8% (46.5 million) in 2004.1 When one also considers those who are underinsured, 35% of adults (61 million) between the ages of 19 to 64 are either uninsured or underinsured.2 Given the magnitude of this problem, it is no surprise that lack of health insurance is now the top health policy story.3
Health insurance coverage serves as an enabling indicator reflecting the degree of financial and social barriers that impede access to health services and positive outcomes.4,5 Discrepancies in hospital care across insurance subgroups have been widely documented demonstrating that unfavorable health insurance status contributes to reduced access to high-quality care and increased likelihood of poor outcomes.6,7 Substantial empiric evidence indicates that uninsured patients are less likely to receive needed care and experience poorer outcomes than other patients.6,8–10
Despite widely documented discrepancies associated with many clinical conditions across insurance subgroups, few studies have examined insurance-related disparities in outcomes of patients with stroke. Stroke is the third leading cause of death in the United States accounting for 162 672 deaths in 2002 and the leading cause of long-term disability. The cost of treating stroke was estimated at $56.8 billion in 2002, of which approximately $35.0 billion was directly related to medical treatment.11 Recent national data show there are approximately 960 000 hospital discharges related to hospital admissions for stroke, among which uninsured patients experience higher in-hospital mortality than patients with insurance.12 We examined disparities in neurologic impairment status and outcomes of patients with stroke across different insurance groups with a special focus on comparisons between the uninsured and those covered by private insurance. Both patterns of care for hemorrhagic stroke and patterns of care for acute ischemic stroke were analyzed to provide a comprehensive picture of stroke care patterns and outcomes across different insurance groups.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Data
We abstracted adult discharges from the 2002 National Inpatient Sample representing 20% of hospital stays from community hospitals in the nation. Based on principal diagnosis International Classification of Diseases, 9th Revision codes, we selected 2 broadly defined types of acute stroke, hemorrhagic stroke and ischemic stroke. We also evaluated care patterns for those admitted with carotid artery occlusion representing a diagnosis not necessarily indicative of acute stroke. For analysis of hemorrhagic stroke, we used code 431 representing intracerebral hemorrhage as the principle diagnosis code. To analyze patterns of care for acute ischemic stroke, we included all patients with a principle diagnosis of code 434 representing occlusion of the cerebral arteries or code 436 representing acute but ill-defined cerebrovascular disease because literature shows that codes 434 and 436 provide reasonably accurate representations of acute stroke.13,14 In addition, we conducted a separate analysis looking at code 433.1 to determine the likelihood of receipt of carotid endarterectomy when carotid artery occlusion was the presenting diagnosis.
We applied several exclusion criteria. First, we excluded transient cerebral ischemia from our analysis because by definition, a transient cerebral ischemia resolves within 24 hours and therefore does not lend itself to a meaningful evaluation of differences in outcomes. Then, we excluded discharges that were transferred to another hospital to avoid double counts in the National Inpatient Sample data. Finally, we excluded discharges with any secondary diagnostic codes that were likely to be associated with potentially confounding clinical conditions (see Table 1
). Those conditions may represent unique circumstances that are not typical of stroke, including conditions related to vasculitis, trauma, and other relatively rare secondary causes of stroke or stroke-like symptoms. To obtain the national total estimate, original case weights in the National Inpatient Sample data set were reweighted by multiplying an adjuster equal to the total weight of original cases divided by total weights after the exclusions.15 A final sample consisted of 9137 adult discharges (age 
18) with intracerebral hemorrhagic stroke and 63 500 adult discharges with occlusion of cerebral arteries or ill-defined stroke both representing acute stroke and 18 050 discharges with carotid artery occlusion, whereas the weighted numbers for the three types were 64 152, 421 882, and 148 989, respectively.
|
|
Measures
We categorized insurance status as Medicaid, uninsured, privately insured including HMO/prepaid health plans, Medicare, and other insurance categories (ie, Worker’s Compensation, CHAMPUS, CHAMPVA, Title V, and other government programs). To apply the national sample weight in the analysis, Medicare discharges and discharges with other insurance categories were not excluded in the analysis, but only results of Medicaid, uninsured, and privately insured patients are reported.
Response variables included neurologic impairment status and in-hospital mortality, all of which were dichotomous variables. Using paralysis and coma as increasingly severe and hierarchical indicators of neurologic impairment, we compared the level of neurologic impairment as a severity measure. If a discharge had coma recorded, it was considered the most severe degree of neurologic impairment and therefore inclusive of paralysis. Less severe forms of neurologic impairment, aphasia and lethargy, were eliminated from the analysis because few patients presented with these findings in the absence of more severe impairment. We also examined hospital mortality because stroke has relatively high in-hospital mortality.
Statistical Analysis
We conducted multiple logistic regression. For all response variables, if a variable was not a "common event" (frequency 
10%), we used OR to approximate risk ratio because the 2 ratios are close. If a variable was a common event (frequency above 10%) in which the OR can be sizably larger than the risk ratio, we converted the OR to risk ratio.16 A set of covariates were controlled through multivariate analysis. First, we controlled for comorbid conditions by including Agency for Healthcare Research and Quality comorbidities (eg, congestive heart failure, depression, diabetes, and hypertension) to reflect nonneurologic clinical conditions.17 Because hypertension is a key risk factor associated with treatment and outcome of stroke,18 it was included in the multivariate models regardless of its statistical significance.
We controlled for sociodemographics, including age, sex, race, and median income level. For the purpose of expanding age to a meaningful interval, we divided age into 5 categories: 18 to 44, 45 to 54, 55 to 64, 65 to 74, 75 to 84, and 85 or older. Race/ethnicity was categorized as white, black, Hispanic, Asian/Pacific Islander, other racial categories, and race recorded as missing value (when race was not reported). The median income by zip code of patient’s residence was divided into 4 levels: $25 000, $25 000 to $34 999; $35 000 to $44 999, and $45 000.
Finally, we controlled for hospital characteristics such as hospital stroke volume (the number of stroke discharges in 2002), hospital bed size (small, medium, and large), as defined by the Healthcare Cost and Utilization Project,19 teaching hospital status, urban/rural location, as defined by the American Hospital Association (metropolitan area means urban and nonmetropolitan area means rural), and geographic region (Northeast, Midwest, South, and West).
To account for the clustering of hospital discharges that the National Inpatient Sample contains, we used the SAS-9 SURVEYFREQ, SURVEYMEANS, and SURVEYLOGIST procedures to perform data analyses. When the analyses were conducted, the sample weights were further adjusted for the exclusions mentioned earlier. Therefore, all estimates presented in the text and tables were weighted to represent national population totals.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Table 2 compared unadjusted sociodemographic and hospitalization characteristics of Medicaid, uninsured, and privately insured patients. For the 3 groups, uninsured patients were the youngest followed by Medicaid patients, whereas privately insured patients were the oldest. Although younger, both Medicaid patients and uninsured patients showed more severe neurologic impairment status than their privately insured counterparts. Among patients with intracerebral hemorrhage, percentages of patients with coma were 8.1%, 8.4%, and 7.3% for Medicaid, the uninsured, and the privately insured, respectively. Furthermore, unadjusted mortality was the highest among uninsured patients (34.7%) followed by Medicaid patients (28.0%) and the privately insured (28.2%). Among patients with acute ischemic stroke, mortality was 6.0%, 5.3%, and 4.4% for uninsured patients, Medicaid patients, and privately insured patients, respectively.
|
Table 3 shows adjusted results. For intracerebral hemorrhage, Medicaid and privately insured patients experienced comparable mortality risk (relative risk [RR] [CI], 1.13 [0.95 to 1.32], but uninsured patients experienced higher mortality risk (RR [CI], 1.24 [1.07 to 1.43]).
|
For acute ischemic stroke, as compared with privately insured patients, patients with Medicaid had comparable risk of coma (OR [CI], 1.19 [0.90 to 1.58]) but higher risk of paralysis (RR [CI], 1.21 [1.14 to 1.27]) and mortality (OR [CI], 1.24 [1.02 to 1.52]); whereas uninsured patients had a higher risk of paralysis (RR[CI], 1.13 [1.06 to 1.21]), coma (OR [CI], 1.64 [1.14 to 2.35]), and mortality (OR[CI], 1.56 [1.24 to 1.97]).
When we compared frequencies of receipt of carotid endarterectomy among patients with carotid artery occlusion, patients with private insurance were more likely to receive the procedure than the uninsured (RR [CI]; 0.41 [0.31 to 0.53]) and patients with Medicaid (RR [CI], 0.84 [0.74 to 0.95]).
To account for potential differences associated with a higher likelihood of privately insured patients receiving the procedure on an elective basis, we evaluated patients with the primary diagnosis code of 434 or 436 indicating acute stroke combined with code 433.1 representing stenosis or occlusion of the carotid artery and the procedure code for carotid endarterectomy. As Table 4 shows, we found that very few patients received carotid endarterectomy (less than 3% of the total patients with both codes received carotid endarterectomy), although a much higher percentage of those with only code 433.1 received the procedure as shown in Table 2.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This nationally representative study demonstrated that uninsured patients had the highest mortality risk for both types of acute stroke (ischemic and hemorrhagic) and were the youngest among the 3 insurance groups. The mortality risk of uninsured patients was 24% to 56% higher than that of their privately insured peers for acute hemorrhagic and acute ischemic stroke, respectively. In addition, uninsured patients with acute ischemic stroke had the highest risk for the most severe neurologic impairment status, coma, among the 3 insurance groups. Differences in both age and severity levels seem to be an indication of inadequate access to preventive services, creating greater vulnerability among the uninsured population, which is consistent with other findings that lack of health insurance is associated with unrecognized risk factors for stroke and increased overall risk for decline in health status.20–22
Greater levels of neurologic impairment among the uninsured may also be caused or exacerbated by lack of access to preventive ambulatory care during which chronic risk factors for stroke (eg, hypertension and hypercholesterolemia) may be detected and controlled.23,24 Furthermore, being uninsured may create stress or other hazards that make patients less likely to adhere to medication and lifestyle modification (eg, smoking cessation, low-sodium diet, and physical activities) to control chronic risk factors for stroke.22,25,26
Probably because of the fact that primary intracerebral hemorrhage is associated with high mortality,27 we did not observe differences in severe neurologic impairment status (eg, coma) across health insurance status. Nevertheless, with its high case mortality rate among patients (32.3% in our sample), the relative risk of mortality for the uninsured was approximately one fourth higher than that of the privately insured. Because clinical interventions for hemorrhagic stroke are rather limited, higher mortality among the uninsured might be the result of lack of control of comorbidities and risk factors in ambulatory settings before the onset of stroke.25 Patients who are able to recognize stroke symptoms as serious tend to arrive at the hospital more promptly, and enhanced access to ambulatory care tends to be associated with increased awareness of health risk factors.20,26,28 The finding of higher frequencies of receipt of carotid endarterectomy in the absence of acute stroke symptoms supports the idea that privately insured patients are more likely to undergo primary or secondary prevention in the absence of acute symptoms. Further studies are warranted to investigate whether uninsured patients with stroke have prolonged delays before hospital arrival and/or greater illness severity on arrival than their peers.
The low frequency of receipt of endarterectomy among patients with acute stroke symptoms as opposed to those without such symptoms suggests this procedure was much more likely to be performed electively among the privately insured patients and supports the concept that access to preventive care (eg, primary or secondary prevention related to performance of carotid endarterectomy in the absence of acute symptoms) is higher among the privately insured as compared with the uninsured.
Disparities in neurologic impairment status and mortality between patients with Medicaid and privately insured patients existed in some cases but to a lesser degree than disparities between uninsured patients and privately insured patients. Our findings demonstrate that the uninsured group was the most vulnerable of the 3.
This study had limitations. First, the National Inpatient Sample data did not allow us to examine postdischarge outcomes (eg, 30-day, 60-day, or 180-day mortality). In addition, clinical information such as vital signs and initial severity to hospital arrival was not available. Despite these limitations, our findings consistently indicate that the uninsured had significantly worse clinical outcomes than the privately insured for both hemorrhage stroke and acute ischemic stroke.
Public policy should promote access to outpatient and preventive care for uninsured patients so that risk factors such as hypertension and hyperlipidemia can be detected and treated during early, asymptomatic stages. These risk factors are known to contribute to racial disparities in outcomes and are likely to be operative in insurance-related disparities. Further research is needed to assess initial severity of stroke on hospital arrival across health insurance status as well as the potential for delayed presentation to the hospital arrival after onset of stroke symptoms for the uninsured.
|
Acknowledgments |
|---|
We thank the 3 anonymous reviewers’ comments on the earlier drafts of this paper.
Disclosures
None.
Received June 15, 2006; revision received August 18, 2006; accepted October 3, 2006.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
- Kaiser Commission on Medicaid and the Uninsured. Health insurance coverage in America, 2004 data update. Available at: www.kff.org/uninsured/upload/Health-Coverage-in-America-2004-Data-Update-Report.pdf. Accessed November 2005.
- Schoen C, Doty MM, Collins SR, Holmgren AL. Insured but not protected: how many adults are underinsured? Health Aff. 2005; 24: 289–302.
[Free Full Text] - Commonwealth Fund. Top health policy stories of 2005—Commonwealth Fund/Health Affairs Online Survey Results. Available at: www.cmwf.org/General/General_show.htm?doc_id=329686&#doc329686. Accessed January 5, 2006.
- Mueller KJ, Patil K, Boilesen E The role of uninsured and race in healthcare utilization by rural minorities. Health Serv Res. 1998; 33: 597–610.[Medline] [Order article via Infotrieve]
- Schoen C, DesRoches C. Uninsured and unstably insured: the importance of continuous insurance coverage. Health Serv Res. 2000; 35: 187–206.[Medline] [Order article via Infotrieve]
- Institute of Medicine. Coverage Matters: Insurance and Health Care. Washington, DC: National Academies Press; 2002.
- Shi L. The convergence of vulnerable characteristics and health insurance in the US. Soc Sci Med. 2001; 53: 519–529.[CrossRef][Medline] [Order article via Infotrieve]
- Bradbury RC, Golec JH, Steen PM. Comparing uninsured and privately insured hospital patients: admission severity, health outcomes and resource use. Health Serv Manage Res. 2001; 14: 203–210.
[Abstract/Free Full Text] - Hadley J, Steinberg EP, Feder J. Comparison of uninsured and privately insured hospital patients. Condition on Admission, Resource Use, and Outcome. JAMA. 1991; 265: 374–379.[Abstract]
- Shen JJ, Wan TTH, Perlin JB. An exploration of the complex relationship of socioecological factors in the treatment and outcomes of acute myocardial infarction in disadvantaged populations. Health Serv Res. 2001; 36: 711–732.[Medline] [Order article via Infotrieve]
- Agency for Healthcare Research and Quality. National and regional statistics from the NIS. Available at: http://hcup.ahrq.gov/HCUPnet.asp. Accessed December 7, 2005.
- Heart Disease and Stroke Statistics—2005 Update. Dallas: American Heart Association; 2005.
- Benesch C, Witter DM Jr, Wilder AL, Duncan PW, Samsa GP, Matchar DB. Inaccuracy of the International Classification of Diseases (ICD-9-CM) in identifying the diagnosis of ischemic cerebrovascular disease. Neurology. 1997; 49: 660–664.
[Abstract/Free Full Text] - Goldstein LB. Accuracy of ICD-9-CM coding for the identification of patients with acute ischemic stroke: effect of modifier codes. Stroke. 1998; 29: 1602–1604.
[Abstract/Free Full Text] - Cost-to-Charge Ratio Files: 2002 National Inpatient Sample (NIS) User Guide. Rockville, MD: Agency for Healthcare Research and Quality; 2005.
- Zhang J, Yu KF. What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA. 1998; 280: 1690–1691.
[Abstract/Free Full Text] - Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998; 36: 8–27.[CrossRef][Medline] [Order article via Infotrieve]
- Iezzoni LI, Shwartz M, Ash AS, Mackiernan YD. Predicting in-hospital mortality for stroke patients: results differ across severity-measurement methods. Med Decis Making. 1996; 16: 348–356.
[Abstract/Free Full Text] - Agency for Healthcare Research and Quality. NIS database documentation. Available at: www.hcup-us.ahrq.gov/db/nation/nis/nisdbdocumentation.jsp. Accessed March 1, 2006.
- Ayanian JZ, Zaslavsky AM, Weissman JS, Schneider EC, Ginsburg JA. Undiagnosed hypertension and hypercholesterolemia among uninsured and insured adults in the Third National Health and Nutrition Examination Survey. Am J Public Health. 2003; 93: 2051–2054.
[Free Full Text] - Baker DW, Sudano JJ, Albert JM, Borawski EA, Dor A. Lack of health insurance and decline in overall health in late middle age. N Engl J Med. 2001; 345: 1106–1112.
[Abstract/Free Full Text] - Centers for Disease Control and Prevention (CDC). Self-assessed health status and selected behavioral risk factors among persons with and without health-care coverage—United States, 1994–1995. MMWR Morb Mortal Wkly Rep. 1998; 47: 176–180.[Medline] [Order article via Infotrieve]
- Gandelman G, Aronow WS, Varma R. Prevalence of adequate blood pressure control in self-pay or Medicare patients versus Medicaid or private insurance patients with systemic hypertension followed in a university cardiology or general medicine clinic. Am J Cardiol. 2004; 94: 815–816.[CrossRef][Medline] [Order article via Infotrieve]
- Moy E, Bartman BA, Weir MR. Access to hypertensive care. Effects of income, insurance, and source of care. Arch Intern Med. 1995; 155: 1497–1502.[Abstract]
- Ayanian JZ, Weissman JS, Schneider EC, Ginsburg JA, Zaslavsky AM. Unmet health needs of uninsured adults in the United States. JAMA. 2000; 284: 2061–2069.
[Abstract/Free Full Text] - Rice T, Lavarreda SA, Ponce NA, Brown ER. The impact of private and public health insurance on medication use for adults with chronic diseases. Med Care Res Rev. 2005; 62: 231–249.[Abstract]
- Barber M, Roditi G, Stott DJ, Langhorne P. Poor outcome in primary intracerebral haemorrhage: results of a matched comparison. Postgrad Med J. 2004; 80: 89–92.
[Abstract/Free Full Text] - Barr J, McKinley S, O’Brien E, Herkes G. Patient recognition of and response to symptoms of TIA or stroke. Neuroepidemiology. 2006; 26: 168–175.[CrossRef][Medline] [Order article via Infotrieve]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||