Regionalizing Acute Stroke Care to Specialized Centers: Hype or Help? Lee H. Schwamm, MD Professor of Neurology, Harvard Medical School Director, TeleStroke.

Regionalizing Acute Stroke Regionalizing Acute Stroke Care to Specialized Centers: Care to Specialized Centers:

Hype or Help?Hype or Help?

Lee H. Schwamm, MDLee H. Schwamm, MD

Professor of Neurology, Professor of Neurology, Harvard Medical SchoolHarvard Medical SchoolDirector, TeleStroke & Acute Stroke Services, Director, TeleStroke & Acute Stroke Services,

Massachusetts General HospitalMassachusetts General HospitalChairChair, , National Steering Committee, National Steering Committee, AHA/ASA GWTG AHA/ASA GWTG

QCOR 2010

DisclosuresDisclosures

FederalFederalNIH (IMS3, MR RESCUE) NIH (IMS3, MR RESCUE)

Industry Clinical Trial Steering Committees Industry Clinical Trial Steering Committees Lundbeck (DIAS4), CoAxia (Flo24)Lundbeck (DIAS4), CoAxia (Flo24)

OtherOtherStroke Systems Consultant to MA DPH, CDC, Stroke Systems Consultant to MA DPH, CDC,

TJCTJCDirector TeleStroke Program at MGHDirector TeleStroke Program at MGH

Over my dead body!

Phil, you and I both know the American public needs

regionalization of acute stroke care

PrimaryStroke Center

CompStrokeCenter

tPA CapableHospitals

Emergency Stabilization Facilities

Field Response and Triage

Paradigm forGeographicRegionalization

Limited?Advanced

“Make the right diagnosis so you can get me to the right place in the right amount of time”

Comp StrokeCenter

Primary StrokeCenter

tPA CapableCenter

Easy to get inEasy to get in Hard to get outHard to get out

Ideal set of evidence-based Ideal set of evidence-based conditions for regionalizationconditions for regionalization

Patients can be identified in pre-hospital settingPatients can be identified in pre-hospital setting When rapidly achieved, patients benefit from an When rapidly achieved, patients benefit from an

intervention or an environment of careintervention or an environment of care More benefit at centers with specific resources More benefit at centers with specific resources

(eg, cath lab) or expertise (eg, trauma team)(eg, cath lab) or expertise (eg, trauma team) Variability in capability (infrastructure) or Variability in capability (infrastructure) or

performance (benefit or harm) across hospitalsperformance (benefit or harm) across hospitals Improved performance is generalizableImproved performance is generalizable Reallocation of patients is feasible, acceptable to Reallocation of patients is feasible, acceptable to

patients, and will lead to the improved outcomepatients, and will lead to the improved outcome

National Recommendations already National Recommendations already endorse acute stroke ROCendorse acute stroke ROC

Stroke 2005;36(3):690-703; Circulation 2005;111(8):1078-91

Acute Stroke ROC already in place in Acute Stroke ROC already in place in certain local and regional service areascertain local and regional service areas

Why Stroke Regionalization?Why Stroke Regionalization? Patients can be identified by stroke screensPatients can be identified by stroke screens Patient benefit occurs with IV tPA (1A), catheter-Patient benefit occurs with IV tPA (1A), catheter-

based reperfusion (1B) or stroke units (1A)based reperfusion (1B) or stroke units (1A) More benefit expected at centers with high More benefit expected at centers with high

volume (SAH) or expertise (stroke teams, INR)volume (SAH) or expertise (stroke teams, INR) Variability in capability (INR, tPA use) or Variability in capability (INR, tPA use) or

performance (sICH rates) across hospitalsperformance (sICH rates) across hospitals Improved performance is generalizable (GWTG)Improved performance is generalizable (GWTG) Reallocation of patients is feasible, acceptable to Reallocation of patients is feasible, acceptable to

patients, and will lead to the improved outcome patients, and will lead to the improved outcome (depends on how much triage you’re planning)(depends on how much triage you’re planning)

Albers GW et al. Chest. 1998;119:683S-698S.Albers GW. Personal communication. February 27, 2003.Rosamond WD et al. Stroke. 1999;30:736-743.

Cardioembolic (20%)Lacunar (25%)(small vessel disease)

Ischemic Stroke (80%) Hemorrhagic Stroke (20%)

Subarachnoid Hemorrhage (30%)

Cryptogenic (30%)

Atherothrombotic CerebrovascularDisease (20%)

IntracerebralHemorrhage (70%)

Cerebrovascular Disease: Cerebrovascular Disease: Stroke TypesStroke Types

?

Various public education and first-Various public education and first-responder screening toolsresponder screening tools

EMS can reliably identify stroke EMS can reliably identify stroke syndromes in the fieldsyndromes in the field

Paramedic performance when completing Paramedic performance when completing the LAPSS for identifying AIS and HS the LAPSS for identifying AIS and HS sensitivity of 91% (95% CI, 76% to 98%), sensitivity of 91% (95% CI, 76% to 98%), specificity of 97% (95% CI, 93% to 99%), specificity of 97% (95% CI, 93% to 99%), negative predictive value of 98% (95% CI, negative predictive value of 98% (95% CI,

95% to 99%).95% to 99%).positive predictive value of 86% (95% CI, 70% positive predictive value of 86% (95% CI, 70%

to 95%), 97% (95% CI, 84% to 99%) after to 95%), 97% (95% CI, 84% to 99%) after correcting for documentation errorscorrecting for documentation errors

Kidwell Stroke 2000;31;71-76

Why Stroke Regionalization?Why Stroke Regionalization? Patients can be identified by stroke screensPatients can be identified by stroke screens Patient benefit occurs with IV tPA (1A), catheter-Patient benefit occurs with IV tPA (1A), catheter-

based reperfusion (1B) or stroke units (1A)based reperfusion (1B) or stroke units (1A) More benefit expected at centers with high More benefit expected at centers with high

volume (SAH) or expertise (stroke teams, INR)volume (SAH) or expertise (stroke teams, INR) Variability in capability (INR, tPA use) or Variability in capability (INR, tPA use) or

performance (sICH rates) across hospitalsperformance (sICH rates) across hospitals Improved performance is generalizable (GWTG)Improved performance is generalizable (GWTG) Reallocation of patients is feasible, acceptable to Reallocation of patients is feasible, acceptable to

patients, and will lead to the improved outcome patients, and will lead to the improved outcome (depends on how much triage you’re planning)(depends on how much triage you’re planning)

NINDS t-PA Stroke TrialNINDS t-PA Stroke TrialFavorable outcome at 3 monthsFavorable outcome at 3 months

52%

38%43%

26%

45%

31%34%

21%17%

21%

0%

10%

20%

30%

40%

50%

60%

BarthelIndex

RankinScale

GlasgowOutcome

NIHSSscore

90 d Death

TPA Placebo

Global outcome statistic OR=1.7; Absolute ~ +13%, Relative 30%

ECASS-3 t-PA for AIS at 3-4.5 hr ECASS-3 t-PA for AIS at 3-4.5 hr Favorable outcome at 3 monthsFavorable outcome at 3 months

52%

45%

8% 8%

0%

10%

20%

30%

40%

50%

60%

Rankin Scale 90 d Death

TPAPlacebo

OR=1.34; Absolute ~ 7%, Relative ~15%

(p= 0.04)

IV tPA is Beneficial 0-4.5 hrIV tPA is Beneficial 0-4.5 hr

PROACT II Stroke TrialPROACT II Stroke Trialof MCA occlusion < 6 hrof MCA occlusion < 6 hr

Favorable outcome at 3 months (1B)Favorable outcome at 3 months (1B)66%

18%

40%

25%

11%

3%

25% 27%

0%

10%

20%

30%

40%

50%

60%

70%

Open M1 mRS 0-2 Sx ICH Death

ProUKPlacebo

15% absolute benefit in functional independence

P=.001

P=.04

P=.06

P=NS

Comparisons of Efficacy Comparisons of Efficacy TrialsTrials Pts RxPts Rx To Prevent a MajorTo Prevent a Major

CEACEA for Sx ICAfor Sx ICA 66 1 Stroke 1 Stroke

IA Lysis (0-6h)IA Lysis (0-6h) 77 1 Stroke1 Stroke

IV tPA (0-3h)IV tPA (0-3h) 88 1 Stroke 1 Stroke

IV tPA (3-4.5 h) IV tPA (3-4.5 h) 1414 1 Stroke1 Stroke

Stroke UnitsStroke Units 1818 1 Stroke or Death1 Stroke or Death

CEA for Asx ICA CEA for Asx ICA 15-2015-20 1 Stroke1 Stroke

Oral Anticoag AFOral Anticoag AF 2020 1 Stroke /yr1 Stroke /yr

IV tPA in AMIIV tPA in AMI 2626 1 Death from MI1 Death from MI

NASCET (n=659); PROACT; NINDS (n=624); ECASS3; BMJ 1997; ACAS; BAATAF; Lancet 1994

Why Stroke Regionalization?Why Stroke Regionalization? Patients can be identified by stroke screensPatients can be identified by stroke screens Patient benefit occurs with IV tPA (1A), catheter-based Patient benefit occurs with IV tPA (1A), catheter-based

reperfusion (1B) or stroke units (1A)reperfusion (1B) or stroke units (1A) More benefit expected w/ operators (CEA) or centers More benefit expected w/ operators (CEA) or centers

with high volume (SAH) or expertise (stroke teams, INR)with high volume (SAH) or expertise (stroke teams, INR) Variability in capability (INR, tPA use) or performance Variability in capability (INR, tPA use) or performance

(sICH rates) across hospitals(sICH rates) across hospitals Improved performance is generalizable (GWTG)Improved performance is generalizable (GWTG) Reallocation of patients is feasible, acceptable to Reallocation of patients is feasible, acceptable to

patients, and will lead to the improved outcome patients, and will lead to the improved outcome (depends on how much triage you’re planning)(depends on how much triage you’re planning)

Better Outcomes at High Volume Better Outcomes at High Volume Centers for Care of Aneurysmal SAHCenters for Care of Aneurysmal SAH

Bardach Stroke 2002;33;1851-1856

Surgeon CEA volume (low, med, high) Surgeon CEA volume (low, med, high) vs. in-hospital mortality, stroke or LOSvs. in-hospital mortality, stroke or LOS

J Am Coll Surg. 2002 Dec;195(6):814-21.

National Inpatient Sample 1996-1997 in 35,821 patients who underwent CEA

Why Stroke Regionalization?Why Stroke Regionalization? Patients can be identified by stroke screensPatients can be identified by stroke screens Patient benefit occurs with IV tPA (1A), catheter-Patient benefit occurs with IV tPA (1A), catheter-

based reperfusion (1B) or stroke units (1A)based reperfusion (1B) or stroke units (1A) More benefit expected at centers with high More benefit expected at centers with high

volume (SAH) or expertise (stroke teams, INR)volume (SAH) or expertise (stroke teams, INR) Variability in performance (sICH rates) or Variability in performance (sICH rates) or

capability (INR, tPA use) across hospitalscapability (INR, tPA use) across hospitals Improved performance is generalizable (GWTG)Improved performance is generalizable (GWTG) Reallocation of patients is feasible, acceptable to Reallocation of patients is feasible, acceptable to

patients, and will lead to the improved outcome patients, and will lead to the improved outcome (depends on how much triage you’re planning)(depends on how much triage you’re planning)

Marked variations in IV tPA by hospital type, in Ontario, from 2002/03 vs. 2004/05

Courtesy of Moira Kapral

All Ischemic Strokes

Ischemic Strokes <2.5 hr

What’s different about stroke?What’s different about stroke?“Stuff vs. Eyeballs”“Stuff vs. Eyeballs”

TraumaTrauma Blood bank, OR, trauma service, imagingBlood bank, OR, trauma service, imaging

Decompression sickness or Air embolismDecompression sickness or Air embolism Hyperbaric Chamber Hyperbaric Chamber

STEMISTEMI IV thrombolysis: EKG/Trop (+test), IV, EMS/ED MD PCI: Cath lab, cardiologist, CCU, anesthesia (?)PCI: Cath lab, cardiologist, CCU, anesthesia (?)

Acute Stroke Acute Stroke IV tPA: CT (-test), ED MD, IV +/- stroke expertise for

a clinical basis for diagnosis and prediction of outcome

IAT: Biplane, INR/ES, NICU, anesthesia (?)IAT: Biplane, INR/ES, NICU, anesthesia (?)

b-ii.

b-i.

a.

Not easy to read a subtle CTNot easy to read a subtle CT

Subtle SDH

Early CT change

The impact of external forces on The impact of external forces on provider tolerance for medical provider tolerance for medical

uncertainty at the bedsideuncertainty at the bedside

Low volume centers struggleLow volume centers struggle 1997-1998, 3948 patients at 29 hospitals in 1997-1998, 3948 patients at 29 hospitals in

ClevelandCleveland 1.8% received IV tPA. 1.8% received IV tPA. 15.7% (8.1%-26.4%) had SICH 15.7% (8.1%-26.4%) had SICH In-hospital mortality was 15.7% In-hospital mortality was 15.7% 50% (37.8%-62.2%) had protocol deviations50% (37.8%-62.2%) had protocol deviations

After training and protocol implementation After training and protocol implementation supervised by the Cleveland Clinic in 1999supervised by the Cleveland Clinic in 1999 18.8% of patients received IV tPA 18.8% of patients received IV tPA SICH rate was 6.4%.SICH rate was 6.4%. protocol deviations occurred in 19.1% of patientsprotocol deviations occurred in 19.1% of patients

Katzan JAMA. 2000;283:1151-1158 and JAMA. 2000;283:1151-1158 and Stroke 2003;34;799-800

Fancy Stuff: Fancy Stuff: Dissolve or Remove ClotsDissolve or Remove Clots

MCA occluded Device placed Clot removed

post 15 mins

Courtesy Dr. Huang-Hellinger

Even Fancier: Make Clots (1A)Even Fancier: Make Clots (1A)

Why Stroke Regionalization?Why Stroke Regionalization? Patients can be identified by stroke screensPatients can be identified by stroke screens Patient benefit occurs with IV tPA (1A), catheter-Patient benefit occurs with IV tPA (1A), catheter-

based reperfusion (1B) or stroke units (1A)based reperfusion (1B) or stroke units (1A) More benefit expected at centers with high More benefit expected at centers with high

volume (SAH) or expertise (stroke teams, INR)volume (SAH) or expertise (stroke teams, INR) Variability in capability (INR, tPA use) or Variability in capability (INR, tPA use) or

performance (sICH rates) across hospitalsperformance (sICH rates) across hospitals Improved performance is generalizable (GWTG)Improved performance is generalizable (GWTG) Reallocation of patients is feasible, acceptable to Reallocation of patients is feasible, acceptable to

patients, and will lead to the improved outcome patients, and will lead to the improved outcome (depends on how much triage you’re planning)(depends on how much triage you’re planning)

Quality Improvement from 2003-2009 in AHAQuality Improvement from 2003-2009 in AHA“Get With the Guidelines-Stroke”“Get With the Guidelines-Stroke”

0%

20%

40%

60%

80%

100%

Baseline YR1 YR2 YR3 YR4 YR5

3943=76/wk

5511=106/wk

7440=143/wk

9840=189/wk

CT (18)

RI (5)

GWTG Stroke Hospitals in Every State1,272 Stroke Contracts Last updated: 6/11/2009

WA(28)

OR(18)

CA(88)

ID(1)

NV(9)

MT(4)

WY(2)

CO(37)

NM(3)

ND(2)

SD(3)

NE (10)

KS(6)

OK(9)

TX(78)

MN(13)

IA(16)

MO(22)

AR(8)

LA(15)

WI(61) MI

(42)

UT(10)

AZ(17)

IL(45)

IN(23)

KY(16)

TN (26)

MS(12)

AL(16)

GA(58)

FL(115)

SC(11)

NC(19)

VA(28)

OH(44)

WV(5)

PA(83)

NY(110)

MD (35)

ME(6)

VT (1)

NH (4)

NJ (39)

MA (60)

DE (4)

DC (4)

1-9

10-19

20-49

50-99

>100

# of Hospitals

Median Beds 300; Major or minor Teaching, 51%

Stroke Discharges <100, 30%; 101-300, 54%; >300, 16%

All triage in America is LocalAll triage in America is Local

The Inequitable Distribution of The Inequitable Distribution of Regionalization to Advanced CentersRegionalization to Advanced Centers

[Ann Emerg Med. 2009;54:261-269.]

Time lost in transit is brain lost in transitTime lost in transit is brain lost in transit

Virtual Regionalization Brings Virtual Regionalization Brings High-Volume Providers to PatientsHigh-Volume Providers to Patients

Spoke Hospital

Hub Hospital

Real Time Audio and Video with full far-end camera controls

Digital Imaging

Remote MD

MD / Fellow / Resident “In-House”

Nurse

Transfer

Comparison of TeleStroke to Conventional Care Comparison of TeleStroke to Conventional Care at a Regional (CA) and Urban (TX) Centerat a Regional (CA) and Urban (TX) Center

Stroke Center

Patient

PopulationSample

Symptom-to-Door

(min)

Door-to-Consult

(min)

Consult-to-Needle

(min)

Symptom-to-Needle

(min)

MGH

(TeleS)Rural

Received tPA(n=6) 36 70 36 = 142

REACH

(TeleS)Rural

Received tPA (n=12) 71 45 18 = 134

TEMPiS

(TeleS)Rural

Received tPA

(n=106) 65 15 61 = 141

Ontario*

(Tx-tPA)Rural

Received tPA (n=23) 34 89 49 = 173

Houston

(Conv)Urban

Received tPA

(n=269) 67 70 = 137

Chapman Stroke. 2000;31:2920-2924

shipship & drip & drip vs. drip & ship

Distribution of TeleStroke Hospitals in Massachusetts

MGH& BWH

21 hospitals in Network6 hospitals Considering------------------------------66 tPA-Capable Centers

Regionalization of Stroke Regionalization of Stroke Thrombolysis in MA (2005 -)Thrombolysis in MA (2005 -)

On site inspection by DPH with mandatory quality On site inspection by DPH with mandatory quality and safety data submision and safety data submision

As of July 2005 when re-routing began, 56/72 As of July 2005 when re-routing began, 56/72 (78%) hospitals had already gained licensure(78%) hospitals had already gained licensure

Licensed hospitals were more likely than non-Licensed hospitals were more likely than non-licensed hospitals to be licensed hospitals to be larger (259 ± 174 v 103 ± 81 beds, p=0.004) larger (259 ± 174 v 103 ± 81 beds, p=0.004) urban (93% v 44%, p<0.001)urban (93% v 44%, p<0.001) teaching (60% v 18%, p=0.02)teaching (60% v 18%, p=0.02)

Effect of DPH PSS on tPA use Effect of DPH PSS on tPA use among eligible patientsamong eligible patients

Variable Pre-PSS (%) Post-PSS (%) P Value

IV tPA when Onset to ED Arrival < 2 hr

40.0 63.7 .0004

IV tPA when Onset to ED Arrival < 3 hr

31.2 53.3 .0001

reason for no IV tPA

83.4 92.0 .0003

Data before re-routing (PrePSSr: 2004-Q4 to 2005-Q2) were compared to after re-routing (PostPSSr: 2005-Q3 to 2006-Q2) by means of Chi-square or Fisher’s exact test.

Raw Rates of IV tPA use in MA by year Raw Rates of IV tPA use in MA by year in all AIS patients arriving <2 hr*in all AIS patients arriving <2 hr*

0%

5%

10%

15%

20%

25%

30%

2005 20006 2007 2008

% IV tPA

*No eligibility criteria applied other than time of arrival minus time of LSW

The MGH TeleStroke Experience:The MGH TeleStroke Experience:Rates of IV-tPA Treatment by Consult TypeRates of IV-tPA Treatment by Consult Type

TeleStroke (TS) vs. In House/Phone Only (ASL)

Yr TS tPA %tPA ASL tPA %tPA2004 10 3 30.0% 257 38 14.8%2005 33 16 48.5% 482 65 13.5%2006 83 28 33.7% 481 59 12.3%2007 98 40 40.8% 512 63 12.3%2008 146 47 32.2% 572 75 13.1%2009 247 96 38.9% 935 114 12.2%Total 617 230 37.3% 3239 414 12.8%

Overall tPA Rate in 2009 = 17%

Why Stroke Regionalization?Why Stroke Regionalization? Patients can be identified by stroke screensPatients can be identified by stroke screens Patient benefit occurs with IV tPA (1A), catheter-based Patient benefit occurs with IV tPA (1A), catheter-based

reperfusion (1B) or stroke units (1A)reperfusion (1B) or stroke units (1A) More benefit expected at centers with high volume More benefit expected at centers with high volume

(SAH) or expertise (stroke teams, INR)(SAH) or expertise (stroke teams, INR) Variability in capability (INR, tPA use) or performance Variability in capability (INR, tPA use) or performance

(sICH rates) across hospitals(sICH rates) across hospitals Improved performance is generalizable (GWTG)Improved performance is generalizable (GWTG) Smart triage and early treatment initiation via Smart triage and early treatment initiation via

telemedicine with escalation of care environment for telemedicine with escalation of care environment for patients requiring tertiary care is feasible, acceptable to patients requiring tertiary care is feasible, acceptable to patients, and will lead to the improved outcomespatients, and will lead to the improved outcomes

World Leaders Embrace Acute Stroke Care Regionalization.

Middle East Declares Lasting Peace

Tea Party Suffers Stunning Nov. Losses