Costs of Acute Care of First-Ever Ischemic Stroke in Taiwan
Background and Purpose— We sought to investigate the direct costs of acute hospitalization for patients with first-ever ischemic stroke in Taiwan.
Methods— Data were prospectively collected from 360 first-ever ischemic stroke patients. Hospital charges were used for analysis. Multiple linear regression analysis was used to identify the main factors influencing costs.
Results— Mean age was 64.9 years (median, 67.0 years), and 58% were male. Mean National Institutes of Health Stroke Scale (NIHSS) score at admission was 9.4 (median, 6.0). Mean initial score of modified Barthel Index was 10.7 (median, 12.0). Median length of stay was 7 days (range, 1 to 122 days). In-hospital mortality was 8%. Overall, median cost per patient was 26 326 New Taiwan dollars (NTD) (original currency) or $841; median cost per day was 3777 NTD or $121. Median costs for patients with initial NIHSS score 0 to 6, 7 to 15, and 16 to 38 were 20 365 NTD ($650), 31 954 NTD ($1020), and 62 653 NTD ($2000), respectively. Daily component (physician and ward charges) accounted for approximately 38% of total costs. Initial NIHSS score, small-vessel occlusion, admission to intensive care unit, sex, and smoking had significant impacts on costs.
Conclusions— Apart from providing cost estimates, we note that stroke severity strongly affects costs.
Stroke is the second most common cause of mortality in Taiwan. In 2001, among the 22 million inhabitants, >13 000 residents died of cerebrovascular disease. In 2000, there were approximately 102 000 inpatients with cerebrovascular disease, and the associated costs claimed for inpatient care exceeded US $148 million. In facing shrinkage of resources for healthcare and the potential use of thrombolytic agents in ischemic stroke, cost estimation is urgently needed. This report presents cost estimates of acute hospitalization for patients with first-ever ischemic stroke in Taiwan and the main determinants of such costs.
Subjects and Methods
This prospective study included 360 patients admitted with first-ever ischemic stroke within 48 hours of onset.1 Patients who received acute thrombolytic therapy were excluded. In this study costs were hospital charges because real costs were not available. The charges were the reimbursement claims made to the Bureau of National Health Insurance in Taiwan. The discharge date was recorded as the date of the patient died or was discharged to any place other than the Neurology Department of the study hospital.
Costs were subdivided into daily component and ancillary component.2 Daily component included physician charges and ward charges. The rest, including charges for emergency services, laboratory workup, at least 1 brain image, and pharmaceuticals, were classified as ancillary component. Costs estimates are presented in original currency (New Taiwan dollars [NTD]) and US dollars (US $1 equals approximately 31.32 NTD).
Charge data came from the discharge database of the hospital; others were prospectively collected at admission. Prespecified independent variables included age (≤65 versus >65 years), sex, comorbidity, smoking, congestive heart failure, valvular heart disease, atrial fibrillation, history of cardiac disease,1 National Institutes of Health Stroke Scale (NIHSS) score, modified Barthel Index (MBI) score (0 to 11 versus 12 to 20; 20=normal), hours after onset (<24 hours or not), stroke subtype (small-vessel occlusion or not),1 intensive care unit admission, and discharge destination (with home or others the reference group). Patients with initial NIHSS score of 0 to 6, 7 to 15, and 16 to 38 were categorized as having mild, moderate, and severe stroke, respectively.
In regression analysis, natural logarithm of costs was the dependent variable. To assess the influence of each independent variable on costs with adjustment for stroke severity, the categorical NIHSS variables were also included as the predictor variables. Multivariable analysis was performed with all prespecified independent variables being entered simultaneously. Analyses were conducted with the use of SPSS version 10.0 for Windows (SPSS Inc). Significant probability value was set at 0.05.
Characteristics of the patients are shown in Table 1. Patients who survived the acute phase were younger (mean, 64.5 versus 69.4 years), had less severe stroke at admission by NIHSS score (mean, 7.9 versus 27.6) or MBI score (mean, 11.4 versus 2.0), and had a shorter length of stay (LOS) than those who died (median, 7 versus 8 days).
Costs increased noticeably with stroke severity (Table 2). The daily component, varying slightly with stroke severity, constituted approximately 38% of the costs. Table 3 summarizes the results of the multivariable regression model.
This report is the first to present the direct costs of first-ever ischemic stroke in Taiwan by incorporating stroke severity, which may not have been well acknowledged in cost analyses of stroke.3 The implication is that stroke severity cannot be overlooked when costs of acute hospitalization are evaluated. Policies regarding introduction of the diagnosis related group–based payment system might carefully consider the role of disease severity.4
An observational hospital-based study of this kind may be subjected to observation bias and practice differences. We also recognize the limitations of using hospital charges to conduct this study.5,6 Costs found in this study are much lower than in other studies when it is considered that rehabilitation therapy was initiated immediately.2,5,7 However, direct comparisons with other studies are problematic because of different practice patterns.
Although LOS was associated with the variation in costs,8,9 costs could still vary widely among patients with similar LOS. Because LOS could, to some extent, substitute for costs as a measure of resource use, we focused our attention on the total costs of acute care rather than LOS.
Smoking was found to be inversely associated with costs, although the association was not significant when evaluated separately but adjusted for stroke severity (not shown). Ninety-five percent of the smokers were male, and smoking was not a significant determinant of costs if sex was removed from the model. A related study revealed that smoking significantly decreased LOS by approximately 1.2 days.1 We do not know the extent to which the change of significant association was attributed to sex or other suppresser variables or if it was primarily due to these study data.
In summary, stroke severity strongly affects costs, with patients classified as having severe stroke incurring twice the cost of those with moderate stroke. Clearly, disease severity should be included in any decisions regarding healthcare resource allocation and when the impact of certain therapeutic strategies is assessed.
This study was supported by a grant from the National Science Council (NSC91-2416-H-110-027), Taiwan.
- Received May 15, 2003.
- Revision received July 5, 2003.
- Accepted July 11, 2003.
Chang KC, Tseng MC, Weng HH, Lin YH, Liou CW, Tan TY. Prediction of length of stay of first-ever ischemic stroke. Stroke. 2002; 33: 2670–2674.
Mamoli A, Censori B, Casto L, Sileo C, Cesana B, Camerlingo M. An analysis of the costs of ischemic stroke in an Italian stroke unit. Neurology. 1999; 53: 112–116.
Diringer MN, Edwards DF, Mattson DT, Akins PT, Sheedy CW, Hsu CY, Dromerick AW. Predictors of acute hospital costs for treatment of ischemic stroke in an academic center. Stroke. 1999; 30: 724–728.
Reed SD, Blough DK, Meyer K, Jarvik JG. Inpatient costs, length of stay, and mortality for cerebrovascular events in community hospitals. Neurology. 2001; 57: 305–314.
Caro JJ, Huybrechts KF. Stroke treatment economic model (STEM): predicting long-term costs from functional status. Stroke. 1999; 30: 2574–2579.
Caro JJ, Huybrechts KF, Duchesne I. Management patterns and costs of acute ischemic stroke: an international study. Stroke. 2000; 31: 582–590.