CLEVELAND ACUTE STROKE EXPERIENCE Cleveland Health Quality Choice –stroke data collected by trained abstractors since 1991 –IV tPA datapoints added 1996.

CLEVELAND ACUTE STROKE EXPERIENCE

• Cleveland Health Quality Choice– stroke data collected by trained abstractors since 1991– IV tPA datapoints added 1996– disbanded 1998

• Cuyahoga County Operation Stroke 1999– data collection instrument designed by the Medical Committee and

The Stroke Group (ETHOSR)

• grants from Genentech/Astra Zeneca

• Cleveland Clinic Health System Stroke QI Program– presented to JCAHO 4/01

Cleveland Health Quality Choice IV tPA Utilization: Ischemic Strokes

July 1997 - June 1998

n=4275

n=70

tPA 1.8%

No tPA

98.2%

Katzan IL etal. JAMA 2000;283:1151

Cleveland Health Quality Choice IV tPA Utilization by Hospital

July 1997 - June 199811

5

32 2

12

1

0

2

4

6

8

10

12

0 1 2 4 5 6 9 17

# tPA cases

# H

ospi

tals

Katzan IL etal. JAMA 2000;283:1151

N = 27

(N = 70)

Cleveland Health Quality Choice IV tPA Symptomatic Intracranial Hemorrhage (ICH)

July 1997 - June 1998

n = 59n = 11

Symptomatic ICH

15.7%*

* 95% CI 8.1% - 26.4%

Katzan IL etal. JAMA 2000;283:1151

Cleveland Health Quality Choice IV tPA: Identified Protocol Deviations

July 1997 - June 1998

Antithrombotics < 24hr

65%

Beyond Time Window

22.5%

High Blood Pressure 12.5%

Katzan IL etal. JAMA 2000;283:1151

Deviations in 50%

Intracranial Hemorrhage after IV tPA

Series # pts Total ICH Sxatic ICH Fatal ICH

NINDS Trial 312 11% 6% 3%

STARS 296 10% 3% n/a

Cologne 100 11% 5% 1%

Multicenter tPA Survey 189 9% 6% 2%

Houston 30 10% 7% 3%

Minnesota 97 13% 9% 6%

Michigan 54 15% 9% n/a

Indianapolis 41 22% 12% 10%

Cleveland 70 22% 16% 9%

Connecticut* 63 17% 6% na

Connecticut IV tPA experience(Bravata DM etal. Arch Intern Med 2002;162:1994)

• Retrospective cohort of 16 community based hospitals 5/96-12/98

• 67% (42/63) major protocol deviations– dosing errors– >3 hours– known increased bleeding risk (eg low platelets)

• Serious extracranial hemorrhage 17% (NINDS 2%)

• In-hospital mortality 31% (NINDS 13%)

Cuyahoga County Operation Stroke Door to Doctor

Median Times

1012

2

10 10 10

20

7

0

5

10

15

20

25

NINDS ALL IV tPA NoExclusions

<2 hr 2-3 hr 3-6 hr >6

Patient population

min

1/00 - 3/01

(N=65) (N=224) (N=253) (N=65)(N=692) (N=101) (N=59)

Katzan IL etal. Stroke 2003 in press

Cuyahoga County Operation Stroke Time to Initiation of CT

Median Times

25

65

20.5

62 68 67.5

108

50

0

20

40

60

80

100

120

NINDS ALL IV tPA NoExclusions

<2 hr 2-3 hr 3-6 hr >6

Patient population

min

1/00 - 3/01

(N=56) (N=241) (N=262) (N=53)(N=671) (N=78) (N=58)

Katzan IL etal. Stroke 2003 in press

Center line = median, box=25-75% quartiles, whiskers=1.5x interquartile range

Katzan IL etal. Stroke 2003 in press

Cuyahoga County Operation Stroke

Cleveland Clinic Health SystemStroke Quality Improvement Program Symptomatic Intracranial Hemorrhage

Symptomatic ICH13.8%*

n=4

No symptomatic ICH86.2%

n=25

*95% CI = 5.5% - 30.6%

7/97-6/98 CHQC

CLEVELAND ACUTE STROKE EXPERIENCE

• Stroke QI requires data– quality of data varies (many hospitals = no data)– multiple barriers must be overcome (behavioral, political, resources)– team building through trust building

• Performance varies widely across hospitals– physicians and hospitals may not like their data– outliers may not mean bad care– community effectiveness may differ from NINDS efficacy

• Protocol deviations are very common – linked with bad outcomes

• Data can change behavior and improve stroke care – community performance improves over time with systematic QI

• Community hospitals can & should give IV tPA IF they are able to demonstrate they know how