Journal of Stroke 2020;22(3):412-415

https://doi.org/10.5853/jos.2020.01928

Letter to the Editor

National Trends in Clinical Outcomes of Endovascular

Therapy for Ischemic Stroke in South Korea between
2008 and 2016
Kwon-Duk Seo,a Min Jin Kang,b Gyu Sik Kim,a Jun Hong Lee,a Sang Hyun Suh,c Kyung-Yul Leed
a
Department of Neurology, National Health Insurance Service Ilsan Hospital, Goyang, Korea
b
Institute of Health Insurance and Clinical Research, National Health Insurance Service Ilsan Hospital, Goyang, Korea
c
Department of Radiology, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea
d
Department of Neurology, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea

Dear Sir: Insurance Service Ilsan Hospital (NHIMC 2019-01-006).

Between 2008 and 2016, a total of 12,501 patients with
The advent of mechanical thrombectomy (MT) in 2015 drasti- acute ischemic stroke were treated with ET. In 2015, ET was
cally changed the treatment of acute stroke.1 The correspond- administered at a higher rate than that in the previous year.
ing guidelines were revised in 2019 to broaden the scope of MT During the study period, the annual rate of discharge to home
within 24 hours of symptom onset.2 In previous studies, MT significantly increased, whereas the rates of cerebral hemor-
have improved the prognoses of patients in real-world clinical rhage, disability, and death significantly decreased (P<0.0001)
settings.3,4 Using nationwide representative data, this study (Figure 1). The rate of home discharge was 35.8% in 2008 and
aimed to evaluate the national trends in clinical outcomes of 43.8% in 2016. The 3-month mortality rate was 27.2% in 2008
patients treated with endovascular therapy (ET) in South Korea. and 19.7% in 2016. The ratio of patients receiving surgery for
From the Korean National Health Insurance Service data-
base,5 we extracted the claims data of patients aged >45 years
who were hospitalized due to cerebral infarction (International (%)
50 3,000
Classification of Diseases, 10th revision, code: I63) and under-
45
went ET (based on the registered procedure codes) between 40
2,500

2007 and 2017. We categorized the analysis period as follows:

Number of patients
35 2,000
the “non-advanced MT period (January 2008 to December
Percent

30
2010),” when stent retrievers were rarely used; “transitional 25 1,500

period (January 2011 to July 2014),” when stent retrievers were 20

1,000
used off-label and were not reimbursed; and the “MT period 15
10
(August 2014 to December 2016),” when the insurance claims 500
5
data verified the frequency of stent retriever use. We catego-
0 2008 0
rized the status of discharge to home as a good outcome and 2009 2010 2011 2012 2013 2014 2015 2016

statuses of cerebral hemorrhage, disability, and death as poor Total ET 3-month mortality 1-year mortality

outcomes. SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) Cerebral hemorrhage Home discharge Disability

was used to perform statistical analysis. Methods are provided

Figure 1. Annual trends among patients treated for ischemic stroke with
in detail in the Supplementary methods. This study was ap-
endovascular therapy (ET) from 2008 to 2016 and their outcomes. All P-
proved by the Institutional Review Board of the National Health values were <0.0001 on the Cochran-Armitage trend test.

Copyright © 2020 Korean Stroke Society

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which
permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

412 http://j-stroke.org pISSN: 2287-6391 • eISSN: 2287-6405

Vol. 22 / No. 3 / September 2020

cerebral hemorrhage within 30 days gradually decreased from ties, and use of tissue plasminogen activator, the risk of death
6.8% in 2008 to 1.8% in 2016. Furthermore, the rate of dis- and disability were significantly lower in the MT and transi-
ability was 36.8% in 2008 and 21.7% in 2016. tional periods than in the non-advanced MT period. Logistic
A period-based analysis revealed that the patients’ average regression analysis adjusted for the same covariates revealed
age and the prevalence of risk factors increased gradually dur- the MT period was an independent predictor of home discharge
ing the study period. The highest proportion of comorbidities (Supplementary Table 1).
and Charlson comorbidity index was noted in the MT period. In agreement with other studies,3,4 our analysis of the na-
The rate of home discharge increased and the mortality rates tionwide data showed that the patients treated with ET had
decreased continuously from the non-advanced MT period to better outcomes in the MT period than in the non-advanced
the MT period. Additionally, the ratio of the patients receiving MT period. Over time, the baseline prognostic characteristics,
surgery for cerebral hemorrhage after ET decreased during the reflected by age and comorbidities, got worse; however, the
study period. The rate of disability after ischemic stroke also outcomes of the patients generally improved; this was sup-
showed a significant decrease from the non-advanced MT pe- ported by the annual trend analyses (Supplementary Table 2).
riod to the MT period (Table 1). The 1-year survival rate tended In this study, the rates of mortality, disability, and cerebral
to increase from the non-advanced MT period to the MT period hemorrhage decreased with the increased application of ET.
(P<0.0001) (Figure 2). After adjusting for age, sex, comorbidi- The 3-month mortality rate in the MT period observed in the

Table 1. Baseline characteristics and outcomes of patients according to the period of endovascular therapy adoption
Standardized difference (%)
Non-advanced Transitional Non-advanced MT Non-advanced MT
Characteristic MT period P Transitional period
MT period period period and period
and MT period
transitional period and MT period
Number 3,028 4,113 5,360
Male sex 1,691 (55.8) 2,246 (54.6) 2,968 (55.4) 0.5619 2.490 0.951 –1.539
Age (yr) 69.00±10.70 70.97±10.81 71.08±11.02 <0.0001 –18.294 –19.158 –1.032
<60 630 (20.8) 706 (17.2) 931 (17.4) <0.0001 9.293 8.753 –0.541
60–69 785 (25.9) 938 (22.8) 1,251 (23.3) 7.270 6.003 –1.267
70–79 1,124 (37.1) 1,519 (36.9) 1,850 (34.5) 0.390 5.436 5.045
≥80 489 (16.1) 950 (23.1) 1,328 (24.8) –17.563 –21.507 –3.935
Hypertension 2,119 (70.0) 3,192 (77.6) 4,125 (77.0) <0.0001 –17.410 –15.857 1.548
Diabetes mellitus 975 (32.2) 1,807 (43.9) 2,574 (48.0) <0.0001 –24.346 –32.713 –8.211
Dyslipidemia 1,147 (37.9) 2,400 (58.4) 3,648 (68.1) <0.0001 –41.861 –63.435 –20.234
Atrial fibrillation 1,411 (46.6) 2,137 (52.0) 2,861 (53.4) <0.0001 –10.734 –13.588 –2.844
Ischemic heart disease 798 (26.4) 1,465 (35.6) 2,007 (37.4) <0.0001 –20.136 –23.964 –3.791
Chronic kidney disease 202 (6.7) 373 (9.1) 427 (8.0) 0.0011 –8.913 –4.975 3.950
CCI >5 1,255 (41.4) 2,410 (58.6) 3,427 (63.9) <0.0001 –34.812 –46.233 –10.982
Intravenous tPA 1,025 (33.9) 1,283 (31.2) 2,032 (37.9) <0.0001 5.674 –8.472 –14.160
3-month mortality 777 (25.7) 903 (22.0) 1,045 (19.5) <0.0001 8.709 14.784 6.068
1-year mortality 1,046 (34.5) 1,239 (30.1) 1,414 (26.4) <0.0001 9.461 17.808 8.322
Cerebral hemorrhage (I61) 195 (6.4) 152 (3.7) 133 (2.5) <0.0001 12.536 19.264 7.023
Cerebral hemorrhage 149 (4.9) 112 (2.7) 102 (1.9) <0.0001 11.481 16.681 5.458
(operation)
Disability 1,016 (33.6) 1,092 (26.5) 1,254 (23.4) <0.0001 15.320 22.653 7.292
Home discharge without 1,086 (35.9) 1,632 (39.7) 2,233 (41.7) <0.0001 –7.873 –11.916 –4.034
events
Home discharge and re-ad- 48 (1.6) 71 (1.7) 67 (1.3) 0.1938 –1.105 2.836 3.934
mission within 30 days
Home discharge and death 113 (3.7) 99 (2.4) 109 (2.0) <0.0001 7.686 10.163 2.535
within 30 days
Values are presented as number (%) or mean±standard deviation.
MT, mechanical thrombectomy; CCI, Charlson comorbidity index; tPA, tissue plasminogen activator.

https://doi.org/10.5853/jos.2020.01928 http://j-stroke.org 413

 Seo et al. National Trends in the Era of Endovascular Therapy

common ET instrument, were only used off-label or when the

1
solitaire stents were not covered by insurance. Improved out-
0.95
0.9
comes of patients treated with ET in the transitional period
compared to those of patients treated with ET in the non-ad-
Survival probability

0.85
0.8 vanced MT period is likely attributed to the increasing use of
0.75 MT. However, we had no data on the proportion of patients
0.7
treated with MT during the transition period and our interpre-
0.65
0.6
tation is not supported by formal analysis.
0.55 In conclusion, our analysis of nationwide insurance data ver-
0.5 ified that the rates of mortality and disability associated with
0 1 2 3 4 5 6 7 8 9 10 11 12
stroke have decreased with the increasingly widespread adop-
Follow-up duration (mo)
tion of stent retrievers.
Number at risk
Non-advanced MT period 3,028 2,484 2,383 2,306 2,251 2,193 2,152 2,121 2,089 2,055 2,030 2,003 1,982
Transitional period
MT period
4,113 3,531 3,386 3,280 3,210 3,141 3,100 3,048 3,008 2,956 2,923 2,899 2,874
5,360 4,670 4,503 4,394 4,315 4,247 4,196 4,144 4,097 4,051 4,014 3,986 3,946 Supplementary materials
Non-advanced MT period Transitional period MT period Supplementary materials related to this article can be found
online at https://doi.org/10.5853/jos.2020.01928.
Figure 2. Kaplan-Meier curve of 1-year mortality in the patients treated
for ischemic stroke by the period of endovascular therapy adoption
(P<0.0001). MT, mechanical thrombectomy.
References
1. Powers WJ, Derdeyn CP, Biller J, Coffey CS, Hoh BL, Jauch EC,
current study was higher than that reported in a meta-analysis et al. 2015 American Heart Association/American Stroke As-
of pivotal clinical trials,6 but it was lower than that reported in sociation focused update of the 2013 guidelines for the early
a real-world data study.7 The 3-month mortality rate during management of patients with acute ischemic stroke regard-
the transitional period was similar to that reported by a study ing endovascular treatment: a guideline for healthcare pro-
performed around the time when stent retrievers were intro- fessionals from the American Heart Association/American
duced.8 We could not verify the clinical data from this claim Stroke Association. Stroke 2015;46:3020-3035.
data; therefore, we used disability registration data following 2. Ko SB, Park HK, Kim BM, Heo JH, Rha JH, Kwon SU, et al.
cerebral injury rather than modified Rankin Scale (mRS) to 2019 Update of the Korean clinical practice guidelines of
evaluate patient prognosis. The ratio of patients with an mRS of stroke for endovascular recanalization therapy in patients
3–5 (vs. 0–2) at 3-months in the meta-analysis of clinical stud- with acute ischemic stroke. J Stroke 2019;21:231-240.
ies and real-world data was higher than the rate of patients 3. Rai AT, Crivera C, Kalsekar I, Kumari R, Patino N, Chekani F, et
with cerebral injury-related disability among the patients in the al. Endovascular stroke therapy trends from 2011 to 2017
national database used in our study.6,7 This difference could show significant improvement in clinical and economic out-
possibly because some patients with an mRS of 3–5 at 3 comes. Stroke 2019;50:1902-1906.
months would have returned to their daily lives after rehabili- 4. Stein L, Tuhrim S, Fifi J, Mocco J, Dhamoon M. National
tation. When patients survive with a disability after a stroke trends in endovascular therapy for acute ischemic stroke:
event, their direct medical costs and indirect social costs in- utilization and outcomes. J Neurointerv Surg 2020;12:356-
crease.9 Therefore, verifying decreases in the number of such 362.
disabled patients is meaningful. 5. National Health Information Database. National Health In-
This study has several limitations. We used insurance claims surance Service. https://nhiss.nhis.or.kr/bd/ay/bdaya001iv.do.
data rather than well-designed clinical trial or prospective reg- 2019. Accessed September 4, 2020.
istry data. Therefore, we were unable to adjust for detailed and 6. Goyal M, Menon BK, van Zwam WH, Dippel DW, Mitchell PJ,
important prognostic factors such initial National Institutes of Demchuk AM, et al. Endovascular thrombectomy after large-
Health Stroke Scale scores, occlusion sites, recanalization sta- vessel ischaemic stroke: a meta-analysis of individual patient
tus, and interval from symptom onset to recanalization. In ad- data from five randomised trials. Lancet 2016;387:1723-
dition, we were unable to verify the precise number of patients 1731.
who received ET in the period when solitaire stents, the most 7. Wollenweber FA, Tiedt S, Alegiani A, Alber B, Bangard C, Ber-

414 http://j-stroke.org https://doi.org/10.5853/jos.2020.01928

Vol. 22 / No. 3 / September 2020

rouschot J, et al. Functional outcome following stroke throm-

bectomy in clinical practice. Stroke 2019;50:2500-2506. Correspondence: Kyung-Yul Lee
Department of Neurology, Gangnam Severance Hospital, Yonsei University
8. Kim BJ, Han MK, Park TH, Park SS, Lee KB, Lee BC, et al. Trends College of Medicine, 211 Eonju-ro, Gangnam-gu, Seoul 06273, Korea
in the effectiveness of endovascular recanalization for acute Tel: +82-2-2019-3325
Fax: +82-2-3462-5904
stroke: is a change taking place? J Stroke Cerebrovasc Dis E-mail: kylee@yuhs.ac
2015;24:866-873. http://orcid.org/0000-0001-5585-7739
9. Joo H, George MG, Fang J, Wang G. A literature review of in- Co-correspondence: Sang Hyun Suh
direct costs associated with stroke. J Stroke Cerebrovasc Dis Department of Radiology, Gangnam Severance Hospital, Yonsei University
College of Medicine, 211 Eonju-ro, Gangnam-gu, Seoul 06273, Korea
2014;23:1753-1763.
Tel: +82-2-2019-4551
Fax: +82-2-3462-5472
E-mail: suhsh11@yuhs.ac
http://orcid.org/0000-0002-7098-4901

Received: May 22, 2020

Revised: August 10, 2020
Accepted: August 20, 2020

This work was supported by the National Health Insurance Ilsan Hospital grant
(2019-1-167). This study used NHIS-NHID data (2019-1-167), provided by Na-
tional Health Insurance Service (NHIS). The authors alone are responsible for the
content and writing of the paper.

The authors have no financial conflicts of interest.

https://doi.org/10.5853/jos.2020.01928 http://j-stroke.org 415

 Supplementary Table 1. HRs and ORs for death, disability, and home discharge in patients between the MT, transitional period and the non-advanced MT period
Death Disability Home discharge
Variable Unadjusted HR Adjusted HR Unadjusted HR Adjusted HR Unadjusted OR Adjusted OR
P P P P P P
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
Non-advanced MT period 1 1 1 1 1 1
Transitional period 0.908 0.0029 0.816 <0.0001 0.677 <0.0001 0.689 <0.0001 1.178 <0.0001 1.214 <0.0001
(0.852–0.968) (0.765–0.871) (0.619–0.741) (0.629–0.755) (1.069–1.298) (1.100-1.340)
MT period 0.762 <0.0001 0.693 <0.0001 0.744 <0.0001 0.750 <0.0001 1.273 <0.0001 1.302 <0.0001
(0.711–0.816) (0.645–0.744) (0.681–0.812) (0.684–0.822) (1.161–1.395) (1.183–1.434)
Vol. 22 / No. 3 / September 2020

https://doi.org/10.5853/jos.2020.01928
Adjusted HR and OR for age, sex, comorbidities, and use of tissue plasminogen activator.
HR, hazard ratio; OR, odds ratio; MT, mechanical thrombectomy; CI, confidence interval.

1http://j-stroke.org
2
Supplementary Table 2. Baseline characteristics and outcomes of patients according to the calendar year
Cochran-Armitage
Characteristic 2008 2009 2010 2011 2012 2013 2014 2015 2016
trend test, P
Number 869 1,000 1,159 1,356 992 1,093 1,359 2,059 2,614
Male sex 502 (57.8) 556 (55.6) 633 (54.6) 760 (56.0) 545 (54.9) 590 (54.0) 724 (53.3) 1,115 (54.2) 1,480 (56.6) 0.5631

http://j-stroke.org
Age (yr) 68.29±10.88 69.37±10.54 69.21±10.68 69.84±10.61 70.69±10.78 71.77±11.02 71.72±10.78 70.82±10.99 71.29±11.08
<60 202 (23.2) 193 (19.3) 235 (20.3) 252 (18.6) 181 (18.2) 179 (16.4) 210 (15.5) 360 (17.5) 455 (17.4) <0.0001
60–69 227 (26.1) 280 (28.0) 278 (24.0) 344 (25.4) 217 (21.9) 231 (21.1) 295 (21.7) 491 (23.8) 611 (23.4) 0.0043
70–79 305 (35.1) 352 (35.2) 467 (40.3) 505 (37.2) 376 (37.9) 387 (35.4) 513 (37.7) 724 (35.2) 864 (33.1) 0.0084
≥80 135 (15.5) 175 (17.5) 179 (15.4) 255 (18.8) 218 (22.0) 296 (27.1) 341 (25.1) 484 (23.5) 684 (26.2) <0.0001
Hypertension 554 (63.8) 705 (70.5) 860 (74.2) 1,026 (75.7) 776 (78.2) 855 (78.2) 1,084 (80.0) 1,562 (75.9) 2,014 (77.0) <0.0001
Diabetes mellitus 251 (28.9) 311 (31.1) 413 (35.6) 559 (41.2) 433 (43.6) 508 (46.5) 633 (46.6) 956 (46.4) 1,292 (49.4) <0.0001
Dyslipidemia 272 (31.3) 373 (37.3) 502 (43.3) 701 (51.8) 565 (57.0) 677 (61.9) 914 (67.3) 1,364 (66.2) 1,827 (69.9) <0.0001
Atrial fibrillation 390 (44.9) 466 (46.6) 555 (47.9) 689 (50.8) 511 (51.5) 558 (51.1) 736 (54.2) 1,119 (54.3) 1,385 (53.0) <0.0001
Ischemic heart disease 165 (19.0) 280 (28.0) 353 (30.5) 436 (32.2) 364 (36.7) 400 (36.6) 516 (38.0) 766 (37.2) 990 (37.9) <0.0001
Chronic kidney disease 61 (7.0) 69 (6.9) 72 (6.2) 108 (8.0) 80 (8.1) 107 (9.8) 133 (9.8) 160 (7.8) 212 (8.1) 0.0312
CCI >5 269 (31.0) 424 (42.4) 562 (48.5) 728 (53.7) 564 (56.9) 681 (62.3) 864 (63.6) 1,267 (61.5) 1,733 (66.3) <0.0001
Intravenous tPA 290 (33.3) 337 (33.7) 398 (34.3) 426 (31.4) 313 (31.6) 318 (29.1) 497 (36.6) 751 (36.5) 1,010 (38.6) <0.0001
3-month mortality 236 (27.2) 273 (27.3) 268 (23.1) 306 (22.6) 219 (22.1) 232 (21.2) 287 (21.1) 389 (18.9) 515 (19.7) <0.0001
1-year mortality 316 (36.4) 365 (36.5) 365 (31.5) 414 (30.5) 295 (29.7) 322 (29.5) 393 (28.9) 534 (25.9) 695 (26.6) <0.0001
Cerebral hemorrhage (I61) 77 (8.9) 65 (6.5) 55 (4.7) 64 (4.7) 35 (3.5) 36 (3.3) 38 (2.8) 44 (2.1) 66 (2.5) <0.0001
Cerebral hemorrhage 59 (6.8) 36 (3.6) 54 (4.7) 48 (3.5) 25 (2.5) 23 (2.3) 30 (2.2) 42 (2.0) 46 (1.8) <0.0001
Seo et al.

(operation)
Disability 320 (36.8) 323 (32.3) 373 (32.2) 415 (30.6) 253 (25.5) 253 (23.1) 359 (26.4) 500 (24.3) 566 (21.7) <0.0001
Home discharge without 311 (35.8) 357 (35.7) 418 (36.1) 537 (39.6) 394 (39.7) 424 (38.8) 541 (39.8) 825 (40.1) 1,144 (43.8) <0.0001
event
Home discharge and re-ad- 14 (1.6) 18 (1.8) 16 (1.4) 23 (1.7) 15 (1.5) 22 (2.0) 16 (1.2) 29 (1.4) 33 (1.3) 0.2068
mission within 30 days
Home discharge and death 50 (5.8) 42 (4.2) 21 (1.8) 38 (2.8) 21 (2.1) 24 (2.2) 36 (2.6) 35 (1.7) 54 (2.1) <0.0001
within 30 days
Values are presented as number (%) or mean±standard deviation.
CCI, Charlson comorbidity index; tPA, tissue plasminogen activator.

https://doi.org/10.5853/jos.2020.01928
National Trends in the Era of Endovascular Therapy
Vol. 22 / No. 3 / September 2020

Supplementary methods within 30 days of discharge because their prognosis was ex-
pected to be poor even if they experienced recurrence or were
Study population rehospitalized for rehabilitation.
We used data from the Korean National Health Insurance da- Poor outcomes were defined as cerebral hemorrhage after
tabase, which has data on 97% of the South Korean popula- the procedure, significant disability, and death. To identify the
tion. We extracted all insurance claims data of patients aged incidence rate of cerebral hemorrhage, we screened patients
>45 years who were hospitalized due to cerebral infarction with insurance claims of ICD-10 code I61 as the main diagno-
(International Classification of Diseases, 10th revision, code sis, as the first subdiagnosis, or with a procedure code related
[ICD-10 code]: I63) as the main or first subdiagnosis between to cerebral hemorrhage within 30 days of ET. The procedure
2007 and 2017. After considering the washout and follow-up codes used for screening were N0322 (burr Hole), N0333 (cra-
periods, we excluded data from the years 2007 and 2017. Fur- niectomy), and S4756/S4622 (hematoma removal). We tracked
ther, we sorted the claims data by procedure codes related to deaths within 3 months and 1 year of ET using mortality data.
thrombectomy and determined whether a tissue plasminogen In South Korea, the registration of disability is possible when at
activator (tPA) was used during thrombectomy in these pa- least 6 months have passed from the time of the stroke event,
tients by referring to the Anatomical Therapeutic Chemical permitting patients to register for the cerebral injury disability.
Classification System code B01AD02 or product code We further excluded patients who received ET in 2017 as we
653500661. could not track their deaths within 1 year or their status of dis-
We classified the patients based on tPA use and the period ability registration for cerebral injury from the selected data.
when endovascular therapy (ET) was performed. It was estimat- Consequently, we analyzed patients and outcome variables by
ed that off-label stent retrievers were used in South Korea be- year and then conducted a comparison analysis based on the
tween October and December of 2010. Even though its usage non-advanced MT, transitional, and MT periods.
was officially approved, stent retrievers were not reimbursed by We identified the comorbidities (hypertension, diabetes mel-
the South Korean National Health Insurance between May litus, dyslipidemia, and atrial fibrillation) that could increase
2013 and July 2014. Therefore, we categorized the period when the risk of stroke and, thus, affect the prognosis and the risk
stent retrievers were rarely used as the “non-advanced MT peri- factors of patients from the insurance claims data prior to ET
od (January 2008 to December 2012)”; when we could not pre- treatment and during hospitalization for ET. We analyzed the
cisely identify stent retriever usage as the “transitional period Charlson comorbidity index of patients to verify the influence
(January 2011 to July 2014)”; and when the insurance claim of comorbidities on death.
data verified the frequency of stent retriever use as the “MT pe-
riod (August 2014 to December 2016).” Statistical analysis
This study was approved by the Institutional Review Board of SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) was used for
the National Health Insurance Service Ilsan Hospital (NHIMC data analysis. To compare descriptive statistics and the fre-
2019-01-006). quency of risk factors, one-way analysis of variance and the
chi-square test were used. To compare the distribution of base-
Outcome and covariates line covariates between the analysis periods, we used the stan-
Good outcomes were defined as discharge of patients to their dardized differences test. To confirm the difference in the incli-
homes after 30 days without significant issues such as death nations of the outcome variables by year, we performed the
or rehospitalization. To screen patients with good outcomes, Cochran-Armitage trend test. We performed a time-dependent
we tracked the insurance claims of patients who were dis- Cox proportional hazard regression analysis to identify the risk
charged home after hospitalization for stroke treatment. To in- factors for death and cerebral injury disability. To identify the
crease the reliability of good outcomes, we excluded patients factors related to home discharge, multiple logistic regression
who had died within 30 days of discharge and those who had analysis was used. P-values of <0.05 for the two-sided tests
claimed rehospitalization under either ICD-10 code I63 or I61 were considered to indicate statistical significance.

https://doi.org/10.5853/jos.2020.01928 http://j-stroke.org 3