Abstract
Background
We sought to describe the proportion of acute ischemic stroke admissions for very old patients (≥85 years), compare the characteristics of very old versus younger patients and identify factors among very old patients associated with adverse outcomes.
Methods
The 2000 Healthcare Cost and Utilization Project data included acute ischemic stroke hospitalizations for patients ≥45 years. The combined outcome was in-hospital mortality or discharge to a long-term care facility.
Results
Among 15,020 stroke hospitalizations, 20.4% were for very old patients. The outcome rate was higher in hospitalizations for very old patients (2,176/3,058, 71.2%; versus 5,748/11,962, 48%; p < 0.0001). More hospitalizations for very old patients were for women (73.5 versus 55.1%; p < 0.0001), fewer for Blacks (6.1 versus 12.3%; p < 0.0001) and fewer at teaching hospitals (30.4 versus 36.2%; p < 0.0001). Among very old patients, factors that were independently associated with the outcome included: age [years; adjusted OR = 1.02 (95% CI = 1.000–1.05)], female gender [1.4 (1.18–1.68)], atrial fibrillation [1.37 (1.15–1.63)], acute myocardial infarction [1.68 (1.20–2.35)], respiratory failure [3.59 (1.60–8.05)] and teaching hospital admission [0.82 (0.69–0.98)]. Similar results were observed in the hospitalizations for younger patients. The adjusted OR for the outcome displayed geographic disparities in both age groups, but the pattern of the geographic variation was not similar between the two age groups.
Conclusions
The very old constitute a substantial proportion of stroke hospitalizations. Hospitalizations for very old patients are more likely to end in death or discharge to a long-term care facility than hospitalizations for younger patients. The pattern of geographic disparity in poststroke adverse outcomes differs between younger and very old patients.
Key Words: Brain ischemia; Brain infarction; Aged, 80 and over
Introduction
The incidence of stroke is strongly associated with increasing age [1]. Although the very old (age ≥85 years) are the fastest growing segment of our population [1], the epidemiology of ischemic stroke in the very old in the USA has not been well described.
Prior studies of stroke patients across the age spectrum have demonstrated that stroke prevalence and mortality increase with advancing age, as does poststroke discharge to a skilled nursing facility [2,3,4]. Studies that have focused upon very old patients with stroke are based on data from Europe and the West Indies [5,6,7,8,9]; these investigations suggest that very old stroke patients have higher inpatient mortality rates [6] and longer lengths of stay [6], and they receive less aggressive care than younger patients [7]. To our knowledge, this is the first study to focus upon very old patients with acute ischemic stroke hospitalized at a US medical center.
Our research objectives were to: (1) describe the proportion of acute ischemic stroke admissions that were for very old patients; (2) present the clinical and hospital characteristics of very old patients compared with younger patients who were hospitalized with an acute ischemic stroke; (3) to identify factors associated with mortality and discharge disposition to a skilled nursing facility among very old patients hospitalized for an ischemic stroke, and (4) to compare the factors related to mortality and discharge to a skilled nursing facility among very old versus younger patients.
Methods
We used data from the 2000 Nationwide Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP). HCUP is maintained by the Agency for Healthcare Research and Quality and is the largest collection of hospital care data in the USA; including discharge level information from community medical centers [10]. The 2000 NIS comprised a 20% random sample of 994 US hospitals in 28 states consisting of approximately 7.4 million hospital discharges. A 20% sample of the NIS, including 1.49 million discharges, was used in this study. The inpatient stay records consisted of clinical and resource use information obtained from hospital administrative discharge data. The unit of analysis for this study was an individual hospitalization.
Definitions
We dichotomized patient age using age demarcations employed in other studies of very old patents: very old age was defined as ≥85 years and younger age as <85 years [1,6,7,8].
HCUP used the American Hospital Association definitions of hospital type including: nonfederal, short-term, general and other specialty hospitals. Specialized hospitals were obstetrics-gynecology, ear-nose-throat, short-term rehabilitation, orthopedic and pediatric facilities. The 2000 NIS covered the following states: Arizona, California, Colorado, Connecticut, Florida, Georgia, Hawaii, Illinois, Iowa, Kansas, Kentucky, Maryland, Massachusetts, Maine, Montana, North Carolina, New Jersey, New York, Oregon, Pennsylvania, South Carolina, Tennessee, Texas, Utah, Virginia, Washington, Wisconsin and West Virginia. The hospital regions were then categorized by HCUP as Northeast, South, Midwest or West. Other hospital characteristics included teaching status, urban or rural location and bed size (small, medium and large).
Outcome Measure
The combined outcome measure of in-hospital mortality and discharge to a long-term care facility was used in this study. Death and institutionalization are often combined as a single endpoint as a common and undesirable stroke outcome for patients [11,12,13,14,15,16,17].
Stroke Hospitalization Selection
Acute ischemic stroke hospitalizations were defined as those with a principal discharge diagnosis of occlusion of cerebral arteries [International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) code 434] or acute but ill- defined cerebrovascular disease (ICD-9-CM code 436) [18, 19]. We included hospitalizations for patients ≥45 years old and with a length of stay of at least 24 h. We excluded hospitalizations for patients transferred from other acute care facilities [20].
Comorbidity and Procedures
Comorbid conditions, identified using ICD-9-CM codes, were: atrial fibrillation, ischemic heart disease, acute myocardial infarction, respiratory failure, hypertension, congestive heart failure, peripheral vascular disease, transient ischemic attack and diabetes mellitus. In addition to these prespecified comorbid conditions, we also evaluated the total number of all diagnoses per hospitalization. ICD-9-CM procedure codes were used to determine the number of procedures for each hospitalization (up to a maximum of 15 procedure codes per hospitalization).
Cost
Cost estimates were made with the cost-to-charge ratios provided by HCUP for the year 2000. These ratios are created using both operating and capital-related costs to help generate accurate inpatient stay costs [21].
Statistical Analyses
Univariate and Bivariate
Descriptive statistics (such as ranges, means with standard deviations and medians with interquartile ranges) were applied to summarize the baseline characteristics. χ2 tests were used to compare binary and ordinal variables, and Student's t tests were employed to compare dimensional variables for the 2 age groups. Statistical significance was accepted when the p value was <0.05. No imputations were made for missing values.
Multivariable
Restricting our analysis to hospitalizations for very old patients, we developed a multivariable logistic regression model to identify factors associated with the dependent variable (the combined outcome of inpatient mortality and discharge to a long-term care facility). Potential independent variables were identified based on a priori clinical judgment as well as statistical significance from the bivariate analysis (p < 0.05). Acute myocardial infarction and respiratory failure were the only 2 variables forced into the model, on the basis of a priori clinical judgment. Using a backward selection process, independent factors were identified with a p value of <0.05. These independent factors were then evaluated in a full logistic regression model. An event-per-variable ratio of >10:1 was maintained during the multivariable modeling [22, 23]. All computations were made using PC-SAS version 6.12 for Windows (SAS Institute Inc.; Cary, N.C.).
Results
Our dataset included 17,706 hospitalizations for acute ischemic stroke during the year 2000; 15,020 hospitalizations met our inclusion criteria, and 3,058/15,020 (20.4%) involved very old patients (table 1).
Table 1.
Stroke hospitalization frequency
| Specific age groups | Stroke hospitalizations (n = 15,020) | Summary age groups | Stroke hospitalizations (n = 15,020) |
|---|---|---|---|
| ≥45 to <55 years | 1,094 (7.3) | <85 years | 11,962 (79.6) |
| ≥55 to <65 years | 1,961 (13.1) | ||
| ≥65 to <75 years | 3,705 (24.7) | ||
| ≥75 to <85 years | 5,202 (34.6) | ||
| ≥85 to <95 years | 2,817 (18.7) | ≥85 years | 3,058 (20.4) |
| ≥95 years | 241 (1.6) | ||
Figures in parentheses are percentages.
Comparison of Characteristics among Very Old versus Younger Stroke Patients
Table 2 provides the data for the comparison of characteristics for hospitalizations for very old versus younger patients. A smaller proportion of hospitalizations for very old patients was for Blacks (6.1 vs. 12.3%; p < 0.0001) and a greater fraction for women (73.5 vs. 55.1%; p < 0.0001). The admission source for stroke hospitalizations was most commonly via the emergency department in both age groups. Older patients were twice as likely to be admitted from long-term care facilities (2.8 vs. 1.4%; p = 0.0001), although this source at admission was uncommon even among hospitalizations for the very old. The mean number of diagnoses was higher among hospitalizations for the very old (mean ± standard deviation: 7.2 ± 2.6 vs. 6.9 ± 2.6; p < 0.0001). Hospitalizations for the very old included fewer procedures than for the younger patients (0.7 ± 1.4 vs. 1.0 ± 1.7; p < 0.0001).
Table 2.
Comparison of characteristics and outcomes among younger versus very old stroke patients
| Characteristic | Age groups |
p value | |
|---|---|---|---|
| <85 years (n = 11,962) | ≥85 years (n = 3,058) | ||
| Age, years | 70.7 ± 10.0 | 88.9 ± 3.4 | − |
| Female gender | 6,597 (55.1) | 2,247 (73.5) | <0.0001 |
| Race | 2,442 (20.4) | 371 (12.1) | <0.0001 |
| White | 6,670 (55.8) | 1,963 (64.2) | − |
| Black | 1,468 (12.3) | 186 (6.1) | − |
| Hispanic | 563 (4.7) | 102 (3.3) | − |
| Asian/Pacific Islander | 232 (1.9) | 34 (1.1) | − |
| Native American | 22 (0.2) | 6 (0.2) | − |
| Other | 157 (1.3) | 43 (1.4) | − |
| Missing | 2,850 (23.8) | 724 (23.7) | − |
| Admission source | − | − | <0.0001 |
| Emergency department | 9,498 (79.4) | 2,387 (78.1) | − |
| Long-term care facilities | 166 (1.4) | 87 (2.8) | − |
| Other | 2,298 (19.2) | 584 (19.1) | − |
| Number of diagnoses | 6.9 ± 2.6 | 7.2 ± 2.6 | <0.0001 |
| Comorbid conditions | − | − | − |
| Hypertension | 8,323 (69.6) | 1,885 (61.6) | <0.0001 |
| Diabetes mellitus | 3,924 (32.8) | 488 (16.0) | <0.0001 |
| Ischemic heart disease | 2,949 (24.7) | 767 (25.1) | 0.6 |
| Atrial fibrillation | 2,148 (18.0) | 1,043 (34.1) | <0.0001 |
| Congestive heart failure | 1,395 (11.7) | 652 (21.3) | <0.0001 |
| Valvular disease | 698 (5.8) | 245 (8.0) | <0.0001 |
| Peripheral vascular disease | 572 (4.8) | 164 (5.4) | 0.2 |
| Respiratory failure | 322 (2.7) | 70 (2.3) | 0.2 |
| Myocardial infarction | 177 (1.5) | 62 (2.0) | 0.03 |
| Facility characteristics | − | − | − |
| Small hospital size | 1,428 (11.9) | 415 (13.6) | 0.01 |
| Rural location | 2,115 (17.7) | 685 (22.4) | <0.0001 |
| Teaching status | 4,332 (36.2) | 931 (30.4) | <0.0001 |
| Southern region | 5,283 (44.2) | 1,218 (39.8) | <0.0001 |
| Discharge disposition | − | − | − |
| Death or long-term care facility | 5,748 (48.0) | 2,176 (71.2) | <0.0001 |
| Long-term care facility | 5,113 (42.7) | 1,868 (61.1) | <0.0001 |
| Died in hospital | 635 (5.3) | 308 (10.1) | <0.0001 |
| Home | 6,214 (52.0) | 882 (28.8) | <0.0001 |
| Cost, USD | 11,351 ± 14,594 | 10,367 ± 17,323 | 0.1 |
| Length of stay, days | 6.4 ± 7.2 | 6.7 ± 11.7 | 0.1 |
| Procedures | 1.0 ± 1.7 | 0.7 ± 1.4 | <0.0001 |
Figures are means ± standard deviation or numbers of subjects with percentages in parentheses.
A larger proportion of hospitalizations for very old patients, compared with younger ones, occurred at small community medical centers (13.6 vs. 11.9%; p = 0.01), rural hospitals (22.4 vs. 17.7%; p < 0.0001) and non-southern hospitals (60.2 vs. 55.8%; p < 0.0001). A smaller fraction of hospitalizations for very old patients occurred at teaching hospitals (30.4 vs. 36.2%; p < 0.0001).
The combined outcome of in-hospital mortality or discharge to a long-term care facility was much higher in the hospitalizations for very old stroke patients (71.2 vs. 48%; p < 0.0001). Similarly, the in-hospital mortality was twice as high among hospitalizations for the very old (10.1 vs. 5.3%; p <0.0001). The discharge disposition to a long-term care facility was also more frequent among hospitalizations for the very old (61.1 vs. 42.7%; p < 0.0001).
Multivariable Analysis
The multivariable results are presented in table 3. Among hospitalizations for the very old, several factors were independently associated with the combined outcome of mortality or discharge to a long-term care facility including: age, female gender, atrial fibrillation, acute myocardial infarction and respiratory failure. Admission to a teaching hospital was related to a decrease in the combined outcome. Geographic disparities were observed such that hospitalizations in the following states had lower odds of the combined outcome among the very old: Maryland, New York, Texas and Virgina (fig. 1).
Table 3.
Factors associated with in-hospital mortality or discharge to a long-term care facility
| Variable | Adjusted OR |
|
|---|---|---|
| <85 years | ≥85 years | |
| Age, years | 1.05 (1.04–1.05) | 1.02 (1.00–1.05) |
| Female gender | 1.12 (1.03–1.20) | 1.40 (1.18–1.68) |
| Atrial fibrillation | 1.14 (1.03–1.26) | 1.37 (1.15–1.63) |
| Acute myocardial infarction | 1.68 (1.20–2.35) | 1.68 (1.20–2.35) |
| Respiratory failure | 3.58 (2.67–4.78) | 3.59 (1.60–8.05) |
| Teaching hospital | 0.92 (0.85–1.00) | 0.82 (0.69–0.98) |
| Florida | 0.84 (0.74–0.96) | 0.79 (0.60–1.03) |
| Maryland | 0.82 (0.66–1.03) | 0.44 (0.27–0.72) |
| North Carolina | 1.15 (0.98–1.35) | 0.70 (0.48–1.01) |
| New Jersey | 0.74 (0.60–0.92) | 0.69 (0.46–1.05) |
| New York | 0.77 (0.66–0.90) | 0.61 (0.45–0.84) |
| South Carolina | 0.76 (0.60–0.97) | 0.96 (0.50–1.83) |
| Tennessee | 0.82 (0.66–1.02) | 1.12 (0.70–1.79) |
| Texas | 1.07 (0.94–1.23) | 0.72 (0.54–0.96) |
| Virginia | 0.75 (0.60–0.95) | 0.31 (0.19–0.51) |
Figures in parentheses are 95% CI.
Fig. 1.
Adjusted geographic variation in in-hospital mortality or discharge to a long-term care facility. The shaded areas represent the states that are independently associated with a reduced odds of the combined outcome (in-hospital mortality or discharge to a long-term care facility). a Adjusted outcome rates among hospitalizations for very old patients (≥85 years). b Adjusted outcome rates among hospitalizations for younger patients (< 85 years).
The factors that were independently associated with the combined outcome in the very old patients were similarly related to the combined outcome in younger patients with the exception of teaching status. The pattern of geographic disparities differed in the younger patients such that hospitalizations in the following states had lower odds of the combined outcome: Florida, New Jersey, New York, South Carolina and Virgina (fig. 1).
Discussion
Limitations
We used the HCUP dataset because it is a large state-based inpatient database, containing hospitalization data from a representative sample of US community hospitals from all regions of the country. The HCUP data are subject to the limitations, however, specifically the lack of detailed clinical information about individual patients. Therefore, we could not differentiate between patients experiencing their first versus recurrent strokes. The HCUP data also do not contain information about patients’ functional status, stroke subtype or stroke severity, items that are known to be associated with the outcomes of mortality and discharge disposition. Because we were not able to include them in our multivariable modeling, the results of the logistic regression should be examined in the context of this limitation. The HCUP data are cross-sectional in nature, limiting any causal inferences that can be made.
Also, we used the combined endpoint of in-hospital mortality and discharge to a skilled nursing facility as the outcome in the multivariable modeling. The variables associated with the combined endpoint might not be the same as those for death alone or discharge to a skilled nursing facility alone.
Finally, we used the HCUP dataset, which includes community hospitals, because the majority of ischemic stroke care in the USA takes place in community hospitals. Future studies should be conducted to determine if similar results are observed at tertiary care centers.
Despite these limitations, we have found that the very old account for 20% of the hospitalizations for ischemic stroke in community medical centers in the USA. This proportion is higher than expected from estimates of the US population who are aged ≥85 years. Specifically, in the USA in the year 2000, although 4% of the population ≥45 years was aged ≥85 years [24], 13% of the community hospitalizations for patients ≥45 years were for patients aged ≥85 years [10].
Our observed proportion of stroke in the very old is higher than reported rates in the Netherlands (10%) [25], Spain (13.1–16.5%) [26], Denmark (16%) [8], the French West Indies (17%) [7] or Sweden (18.8%) [27] but is lower than that in the European Union (30%) [5]. International differences in hospitalization rates for nonstroke medical and surgical indications have been demonstrated previously [28].
We also found that almost three fourths of the hospitalizations for the very old result in either death or discharge to a long-term care facility after stroke. The previous studies regarding stroke hospitalizations in very old patients do not report data for this combined outcome. Rather, they found rates of death or discharge to a skilled nursing facility ranging from 5 to 70%, depending on stroke subtype, stroke severity, patient characteristics and follow-up period [13,15,16,17].
As expected, some comorbid conditions were seen more commonly in very old patients (e.g. atrial fibrillation), and others were more common in younger ones (e.g. diabetes). These results were similar to prior studies [7, 8, 29].
Our results indicate that stroke hospitalizations for the very old are less likely to occur at teaching centers. This finding, plus our observation that hospitalizations for very old patients were associated with fewer procedures, may suggest that very old stroke patients receive less aggressive care compared with younger ones. Although the HCUP data may underreport procedures overall, it is unlikely that differential procedure reporting would occur by age; it rather varies among hospitals. Prior investigations have described less intensive care for very old stroke patients [5, 7, 30]. For example, in a study of 100 black patients ≥85 years and 480 patients <85 years from Martinique, the very old were less likely to be admitted to a neurology ward (8 vs. 24%, p < 0.001) or to receive brain imaging (82 vs. 94%, p < 0.0001) [7].
We performed multivariable modeling to identify factors that were associated with poor outcomes for stroke patients (death or institutionalization). Among the very old, we found that age was independently related to in-hospital mortality or discharge to a long-term care facility. Although age has been identified as an independent predictor of death and institutionalization in numerous studies, to our knowledge, none has examined this issue restricted to very old patients. Future research should examine the role of age in prognostication for very old stroke patients.
The multivariable analysis results also indicated that very old patients who were admitted to a teaching hospital had better outcomes than very old subjects admitted to non-teaching centers. Future studies should be directed at confirming this finding and identifying the differences in patient characteristics or stroke care that can account for the discrepancies in patient outcomes.
Geographic variation in poststroke outcomes, especially in poststroke mortality, has been demonstrated previously [31]. The southern part of the USA has the highest rate of poststroke mortality and is often referred to as the Stroke Belt. Recently, studies have indicated that the absolute differences in mortality rates may be varying and the specific states with the highest poststroke mortality may be changing with time [31]. Our findings differ from the previous literature in terms of the precise states with the lowest mortality. Also, little of the previous literature has focused on the differences in geographic variation (e.g. the specific states where the outcomes are either worse or better) between very old and younger stroke patients. Howard et al. [32] compared poststroke mortality in the Stroke Belt to the non-Stroke-Belt states by age and race/ethnicity groups, finding that the mortality differences were pronounced for very old Whites but with less geographic variation in outcome for very old Blacks.
Conclusions
Our research findings suggest that the very old constitute a large proportion of ischemic stroke hospitalizations and that they are more likely to suffer adverse outcomes after stroke than younger patients. Future studies of ischemic stroke should examine very old patients and should be directed toward reducing age-based disparities in outcomes.
Acknowledgment
D.M.B. was supported by an Advanced Career Development Award from the Department of Veteran Affairs Health Services Research and Development Service and the Robert Wood Johnson Generalist Physician Faculty Scholars Award Program. L.M.B. was supported, in part, by awards from the National Institute of Aging (RO3 AG022075-01) and the National Institute of Neurological Disorders and Stroke (R01 NS043322-01 A1).
References
- 1.Campion EW. The oldest old. N EngJ Med. 1994;330:1819–1820. doi: 10.1056/NEJM199406233302509. [DOI] [PubMed] [Google Scholar]
- 2.Brown RD, Whisnant JP, Sicks JD, O'Fallon WM, Wiebers DO. Stroke incidence, prevalence, and survival: secular trends in Rochester, Minnesota, through 1989. Stroke. 1996;27:373–380. [PubMed] [Google Scholar]
- 3.Wolf PA, D'Agostino RB, O'Neal MA, Sytkowski P, Kase CS, Belanger AJ, Kannel WB. Secular trends in stroke incidence and mortality: The Framingham Study. Stroke. 1992;23:1551–1555. doi: 10.1161/01.str.23.11.1551. [DOI] [PubMed] [Google Scholar]
- 4.Bravata DM, Shih-Yieh H, Brass L, Concato J, Scinto J, Meehan T. Long-term mortality in cerebrovascular disease. Stroke. 2003;34:699–704. doi: 10.1161/01.STR.0000057578.26828.78. [DOI] [PubMed] [Google Scholar]
- 5.Di Carlo A, Lamassa M, Pracucci G, Basile AM, Trefoloni G, Vanni P, Wolfe CD, Tilling K, Ebrahim S, Inzitari D, European Biomed Study of Stroke Care Group Stroke in the very old: clinical presentation and determinants of 3-month functional outcome – a European perspective. Stroke. 1999;30:2313–2319. doi: 10.1161/01.str.30.11.2313. [DOI] [PubMed] [Google Scholar]
- 6.Arboix A, Garcia-Eroles L, Massons J, Oliveres M, Targa C. Acute stroke in very old people: clinical features and predictors of in-hospital mortality. J Am Geriatr Soc. 2000;48:36–41. doi: 10.1111/j.1532-5415.2000.tb03026.x. [DOI] [PubMed] [Google Scholar]
- 7.Olindo S, Cabre P, Deschamps R, Chatot-Henry C, Rene-Corail P, Fournerie P, Saint-Vil M, May F, Smadja D. Acute stroke in the very elderly: epidemiological features, stroke subtypes, management, and outcome in Martinique, French West Indies. Stroke. 2003;34:1593–1597. doi: 10.1161/01.STR.0000077924.71088.02. [DOI] [PubMed] [Google Scholar]
- 8.Kammersgaard LP, Jorgensen HS, Reith J, Nakayama H, Pedersen PM, Olsen TS, Copenhagen Stroke Study Short-and long-term prognosis for very old stroke patients: The Copenhagen Stroke Study. Age Ageing. 2004;33:149–154. doi: 10.1093/ageing/afh052. [DOI] [PubMed] [Google Scholar]
- 9.Marini C, Baldassarre M, Russo T, De Santis F, Sacco S, Ciancarelli I, Carolei A. Burden of first-ever ischemic stroke in the oldest old: evidence from a population-based study. Neurology. 2004;62:77–81. doi: 10.1212/01.wnl.0000101461.61501.65. [DOI] [PubMed] [Google Scholar]
- 10.Agency for Healthcare Research and Quality: Healthcare Cost and Utilization Project (HCUP), 1988–2000. http://www.ahrq.gov/data/hcup/hcup-pkt.htm [PubMed]
- 11.Kalra L, Eade J. Role of stroke rehabilitation units in managing severe disability after stroke. Stroke. 1995;26:2031–2034. doi: 10.1161/01.str.26.11.2031. [DOI] [PubMed] [Google Scholar]
- 12.Fagerberg B, Claesson L, Gosman-Hedstrom G, Blomstrand C. Effect of acute stroke unit care integrated with care continuum versus conventional treatment: a randomized 1-year study of elderly patients: the Goteborg 70+ Stroke Study. Stroke. 2000;31:2578–2584. doi: 10.1161/01.str.31.11.2578. [DOI] [PubMed] [Google Scholar]
- 13.Evans A, Harraf F, Donaldson N, Kalra L. Randomized controlled study of stroke unit care versus stroke team care in different stroke subtypes. Stroke. 2002;33:449–455. doi: 10.1161/hs0202.102364. [DOI] [PubMed] [Google Scholar]
- 14.Indredavik B, Bakke F, Slordahl SA, Rokseth R, Haheim LL. Stroke unit treatment: 10-year follow-up. Stroke. 1999;30:1524–1527. doi: 10.1161/01.str.30.8.1524. [DOI] [PubMed] [Google Scholar]
- 15.Kalra L, Evans A, Perez I, Knapp M, Donaldson N, Swift CG. Alternative strategies for stroke care: a prospective randomised controlled trial. Lancet. 2000;356:894–899. doi: 10.1016/S0140-6736(00)02679-9. [DOI] [PubMed] [Google Scholar]
- 16.Hankey GJ, Jamrozik K, Broadhurst RJ, Forbes S, Anderson CS. Long-term disability after first-ever stroke and related prognostic factors in the Perth Community Stroke Study, 1989–1990. Stroke. 2002;33:1034–1040. doi: 10.1161/01.str.0000012515.66889.24. [DOI] [PubMed] [Google Scholar]
- 17.Stroke Unit Trialists' Collaboration How do stroke units improve patient outcomes? A collaborative systematic review of the randomized trials. Stroke. 1997;28:2139–2144. doi: 10.1161/01.str.28.11.2139. [DOI] [PubMed] [Google Scholar]
- 18.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. doi: 10.1161/01.str.29.8.1602. [DOI] [PubMed] [Google Scholar]
- 19.Tirschwell DL, Longstreth WT., Jr Validating administrative data in stroke research. Stroke. 2002;33:2465–2470. doi: 10.1161/01.str.0000032240.28636.bd. [DOI] [PubMed] [Google Scholar]
- 20.Marciniak TA, Ellerbeck EF, Radford MJ, Kresowik TF, Gold JA, Krumholz HM, Kiefe CI, Allman RM, Vogel RA, Jencks SF. Improving the quality of care for Medicare patients with acute myocardial infarction: results from the Cooperative Cardiovascular Project. JAMA. 1998;279:1351–1357. doi: 10.1001/jama.279.17.1351. [DOI] [PubMed] [Google Scholar]
- 21.Friedman B, De La Mare J, Andrews R, McKenzie DH. Practical options for estimating cost of hospital inpatient stays. J Health Care Finance. 2002;29:1–13. [PubMed] [Google Scholar]
- 22.Concato J Peduzzi P, Holford T, Feinstein A. Importance of events per independent variable in proportional hazards analysis. I. Background, goals, and general strategy. J Clin Epideminol. 1995:1495–1501. doi: 10.1016/0895-4356(95)00510-2. [DOI] [PubMed] [Google Scholar]
- 23.Peduzzi P, Concato J, Feinstein A, Holford T. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epideminol. 1995:1503–1510. doi: 10.1016/0895-4356(95)00048-8. [DOI] [PubMed] [Google Scholar]
- 24.US Census Bureau Population Division Information and Research Services Staff: Annual Estimates of the Population by Sex and Five-Year Age Groups for the United States: April 1, 2000 to July 1, 2004. http://www.census.gov/popest/national/asrh/NC-EST2004/NC-EST2004.01.xls
- 25.Bots M, Looman S, Koudstaal P, Hofman A, Hoes A, Grobbee D. Prevalence of stroke in the general population: the Rotterdam Study. Stroke. 1996;27:1499–1501. doi: 10.1161/01.str.27.9.1499. [DOI] [PubMed] [Google Scholar]
- 26.Arboix A, Miguel M, Ciscar E, Garcia-Eroles L, Massons J, Balcells M. Cardiovascular risk factors in patients aged 85 or older with ischemic stroke. Clin Neurol Neurosurg. 2006;108:638–643. doi: 10.1016/j.clineuro.2005.10.010. [DOI] [PubMed] [Google Scholar]
- 27.Liebetrau M, Steen B, Skoog I. Stroke in 85-year-olds: prevalence, incidence, risk factors, and relation to mortality and dementia. Stroke. 2003;34:2617–2622. doi: 10.1161/01.STR.0000094420.80781.A9. [DOI] [PubMed] [Google Scholar]
- 28.McPherson K. International differences in medical care practices. Health Care Financ Rev. 1989;(spec No):9–20. [PMC free article] [PubMed] [Google Scholar]
- 29.Rastas S, Verkkoniemi A, Polvikoski T, Juva K, Niinisto L, Mattila K, Lansimies E, Pirttila T, Sulkava R. Atrial fibrillation, stroke, and cognition: a longitudinal population-based study of people aged 85 and older. Stroke. 2007;38:1454–1460. doi: 10.1161/STROKEAHA.106.477299. [DOI] [PubMed] [Google Scholar]
- 30.Perls TT, Wood ER. Acute care costs of the oldest old: they cost less, their care intensity is less, and they go to nonteaching hospitals. Arch Intern Med. 1996;156:754–760. doi: 10.1001/archinte.156.7.754. [DOI] [PubMed] [Google Scholar]
- 31.Howard G, Howard VJ, Katholi C, Oli MK, Huston S. Decline in US stroke mortality: an analysis of temporal patterns by sex, race, and geographic region. Stroke. 2001;32:2213–2220. doi: 10.1161/hs1001.096047. [DOI] [PubMed] [Google Scholar]
- 32.Howard G, Evans G, Pearce K, Howard V, Bell R, Mayer E, Burke G. Is the Stroke Belt disappearing? An analysis of racial, temporal, and age effects. Stroke. 1995;26:1153–1158. doi: 10.1161/01.str.26.7.1153. [DOI] [PubMed] [Google Scholar]