Venous Thromboembolism (VTE) Prevention in the Hospital

Greg Maynard MD, MScClinical Professor of Medicine and Chief,

Division of Hospital MedicineUniversity of California, San Diego

VTE: A Major Source of Mortality and Morbidity

• 350,000 to 650,000 with VTE per year• 100,000 to > 200,000 deaths per year • Most are hospital related. • VTE is primary cause of fatality in half-

– More than HIV, MVAs, Breast CA combined– Equals 1 jumbo jet crash / day

• 10% of hospital deaths– May be the #1 preventable cause

• Huge costs and morbidity (recurrence, post-thrombotic syndrome, chronic PAH)

Surgeon General’s Call to Action to Prevent DVT and PE 2008 DHHS

Risk Factors for VTE

StasisAge > 40ImmobilityCHFStrokeParalysisSpinal Cord injuryHyperviscosityPolycythemiaSevere COPDAnesthesiaObesityVaricose Veins

Hypercoagulability CancerHigh estrogen statesInflammatory BowelNephrotic SyndromeSepsisSmokingPregnancyThrombophilia

Endothelial Endothelial DamageDamageSurgerySurgeryPrior VTEPrior VTECentral linesCentral linesTraumaTrauma

Anderson FA Jr. & Wheeler HB. Anderson FA Jr. & Wheeler HB. Clin Chest MedClin Chest Med 1995;16:235. 1995;16:235.

Risk Factors for VTE

StasisAge > 40ImmobilityCHFStrokeParalysisSpinal Cord injuryHyperviscosityPolycythemiaSevere COPDAnesthesiaObesityVaricose Veins

Hypercoagulability CancerHigh estrogen statesInflammatory BowelNephrotic SyndromeSepsisSmokingPregnancyThrombophilia

Endothelial Endothelial DamageDamageSurgerySurgeryPrior VTEPrior VTECentral linesCentral linesTraumaTrauma

Anderson FA Jr. & Wheeler HB. Anderson FA Jr. & Wheeler HB. Clin Chest MedClin Chest Med 1995;16:235. 1995;16:235. Bick RL & Kaplan H. Bick RL & Kaplan H. Med Clin North AmMed Clin North Am 1998;82:409. 1998;82:409.

Most hospitalized patients have

at least one ris

k factor for V

TE

Failure to Do Simple Things Well

• Wash Hands – 60% Reliable

• Patients Understand Meds / Problems– 40% Reliable

• Central Lines Placed w/ Proper Technique– 60% Reliable

• Basal Insulin for Inpt Uncontrolled DM– 40% Reliable

• VTE Prophylaxis– 50% Reliable

Registry DataHighlight the Underuse of

ThromboprophylaxisDVT-FREE RIETE IMPROVE

BAD NEWS!Only a minority of hospitalized patients receive thromboprophylaxis

Goldhaber SZ, Tapson VF. Am J Cardiol 2004;93:259-62.Monreal M, et al. J Thromb Haemost 2004;2:1892-8.Tapson V, et al. Blood 2004;104:11. Abstract #1762.

Endorse Results

• Out of ~70,000 patients in 358 hospitals, appropriate prophylaxis was administered in:– 58.5% of surgical patients– 39.5% of medical patients

Cohen, Tapson, Bergmann, et al. Venous thromboembolism risk and prophylaxis in the acute hospital care setting (ENDORSE study): a

multinational cross-sectional study. Lancet 2008; 371: 387–94.

The “Stick” is coming….

NQF endorses measures already

Public reporting and TJC measures coming soon:- Prophylaxis in place within 24 hours of admit or risk

assessment / contraindication justifying it’s absence- Same for critical care unit admit / transfers- Track preventable VTE

CMS – DVT or PE with knee or hip replacement reimbursed as though complication had not occurred.

Why don’t we do better?

• Lack of awareness or buy in of guidelines

• Underestimation of clot risk, overestimation of bleeding risk

• Lack of validated risk assessment model

• Translating complicated guidelines into everyday practice is difficult

E-Alerts Can Increase Prophylaxis

• 2506 hospitalized patients• VTE risk score ≥ 4• Randomized to intervention or control

InterventionTreatment Received

Mechanical, % Pharmacologic, %

E-Alert 10 23.6

Control 1.5 13

P-value 0.001 0.001

Kucher N, et al. N Engl J Med. 2005;352:969-977.

Intervention

Control

Time (days)

0 30 60 90

% F

reed

om

fro

m D

VT

/ PE

90

92

94

96

98

100

E-Alerts Decrease VTE

Kucher N, et al. N Engl J Med. 2005;352:969-977.

Intervention

Control

Number at risk

1255 977 900

1251 976 893 839

853

41%P = 0.001

Effectiveness can wane over time

Lecumberri R, et al. Thromb Haemost. 2008;100:699-704.

00.5

11.5

22.5

33.5

44.5

Overall MedicalPatients

SurgicalPatients

VT

E I

nci

den

ce/1

000

Pat

ien

ts

2005 (pre-intervention)

2006

2007

*P < 0.05

*

Human Alerts Increase Prophylaxis

• 2493 hospitalized patients• VTE risk score ≥ 4• Randomized to intervention or control

InterventionTreatment Received

Mechanical, % Pharmacologic, %

Hu-Alert 21 28

Control 8 14

95% CI 10.6-16.0 10.5-16.8

Piazza G, et al. Circulation. 2009;119:2196-2201.

% F

reed

om

fro

m D

VT

/ P

E

Human Alerts Decrease VTE

Time After Initial Enrollment (days)

P = 0.31

Piazza G, et al. Circulation. 2009;119:2196-2201.

Bottom Line - Alerts

• A Useful Strategy

• E – Alerts and Human Alerts can work

• Not a panacea

• Alert fatigue can be a problem

• Need a multifaceted approach

Medical Admission Order Sets Can Improve DVT Prophylaxis………

Baseline- Only 11% of inpatients on any VTE prophylaxis

Intervention – A simple prompt for UFH or Mechanical

Prophylaxis placed into voluntary admission order sets.

Post intervention:44% on any prophylaxis26% pharmacologic prophylaxis

O'Connor C, Adhikari N, DeCaire K, Friedrich Jan. Medical Admission Order Sets to Improve Deep Vein Thrombosis Prophylaxis Rates and Other Outcomes. J Hosp Med 2009

…but not enough by themselves, and design of the order set matters

• Best practice prophylaxis not defined Prompt ≠ Protocol

• No protocol = No guidance at the point of carein order set, heparin, mechanical devices, and no

prophylaxis presented as equal choices

• Implementation / ReliabilityAt 15 months, only about half of inpatient

admissions utilized standardized order set.

Other methods needed to enhance performance!

Education alone is not sufficient

….but it is essential to optimize other strategies that are effective

• Standardized order sets• Computerized decision support• E-alerts• Human alerts• Raising situational awareness• Audit and feedback

19

Percent of Randomly Sampled Inpatients with Adequate VTE Prophylaxis

20%

30%

40%

50%

60%

70%

80%

90%

100%

Baseline

Consensus building

Order Set Implementation & Adjustment

Real time ID & intervention

Percent of Randomly Sampled Inpatients with Adequate VTE Prophylaxis

20%

30%

40%

50%

60%

70%

80%

90%

100%

Baseline

Consensus building

Order Set Implementation & Adjustment

Real time ID & intervention

N = 2,944 mean 82 audits / month UCSD experience

UCSD VTE Protocol Validated

• Easy to use, on direct observation – a few seconds• Inter-observer agreement –

– 150 patients, 5 observers- Kappa 0.8 and 0.9

• Predictive of VTE • Implementation = high levels of VTE prophylaxis

– From 50% to sustained 98% adequate prophylaxis– Rates determined by over 2,900 random sample audits

• Safe – no discernible increase in HIT or bleeding• Effective – 40% reduction in HA VTE

– 86% reduction in risk of preventable VTE

UCSD - Decrease in Patients with Preventable HA VTE

0

2

4

6

8

10

12

14

Q 1 '0

5

Q2 '05

Q3 '05

Q4 '05

Q1'06

Q2 '06

Q3 '06

Q4 '06

Q1 '07

Quarter

# o

f P

ati

en

ts

Medicine

Surgery

Ortho

Other

Total

21

Level 5 Oversights identified and addressed in real timeOversights identified and addressed in real time 95+%

Dr.

May

nard

, th

e C

Is a

re d

iffer

ent

here

and

in

the

pro

of.

Whi

ch a

re c

orre

ct?

Maynard GA, et al. J Hosp Med. 2009;

Hospital Acquired VTE by Year2005 2006 2007

Patients at Risk 9,720 9,923 11,207

Cases w/ any VTE 131 138 92Risk for HA VTE 1 in 76 1 in 73 1 in 122Unadjusted RR 1.0 1.03 0.61#

(95% CI) (0.81-1.31) (0.47- 0.79)

Cases with PE 21 22 15Risk for PE 1 in 463 1 in 451 1 in 747

Unadjusted RR 1.0 1.02 0.62 (95% CI) (0.54-1.86) (0.32-1.20)

Cases with DVT (and no PE) 110 116 77Risk for DVT 1 in 88 1 in 85 1 in 146

Unadjusted RR 1.0 1.03 0.61* (95% CI) (0.80-1.33) (0.45-0.81)

Cases w/ Preventable VTE 44 21 7Risk for Preventable VTE 1 in 221 1 in 473 1 in 1,601

Unadjusted RR 1.0 0.47# 0.14*(95% CI) (0.28-0.79) (0.06-0.31)

# p < 0.01 *p < 0.001

2008

80

12

68

6

VTE Prevention Guides Modeling a Multifaceted Approach

http://ahrq.hhs.gov/qual/vtguide/

http://www.hospitalmedicine.org/ResourceRoomRedesign/RR_VTE/VTE_Home.cfm

VTE QI Resource Room www.hospitalmedicine.org

Collaborative Efforts

• SHM VTE Prevention Collaborative I - 25 sites• SHM / VA Pilot Group - 6 sites• SHM / Cerner Pilot Group – 6 sites• AHRQ / QIO (NY, IL, IA) - 60 sites• IHI Expedition for VTE Prevention – 60 sites

• Effective across wide variety of settings– Paper and Computerized / Electronic – Small and large institutions– Academic and community

Basic Ingredients for Success

• Institutional support, will to standardize the process

• Designated multidisciplinary team with physician leadership

• Specific goals and metrics• VTE Protocol guidance built into order sets• Education / consensus• Alerts / feedback to clinicians in real time

Enlist Key Groups / Leaders

• Section Heads

• Hospitalists – (most groups receive some direct support

from the hospital)

• Other high volume providers

• Find some more physician champions

Educational Detailing - PR

Quote ACCP 8 Guidelines

Don’t use aspirin alone for DVT prophylaxis

Mechanical prophylaxis is not first line prophylaxis in the absence of contraindications to pharmacologic prophylaxis

Geerts WH et al. Chest. 2008;133(6 Suppl):381S-453S

Use the powerful anecdote and data

• Look for VTE case that could have been prevented

• Personalize the story• Enlist a patient / family to help you tell the

story• Get data on VTE in your medical center

– (it occurs more often than the doctors think it does)

Q and A

Q. What is the best VTE risk assessment model?

A. Simple, text based model with only 2-3 layers of VTE Risk

Q. Who should do the VTE risk assessment?

A. Doctors (via admit transfer order sets), with back up risk assessment by front line nurses or pharmacists, focusing on those without prophylaxis.

Hierarchy of Reliability

No protocol* (“State of Nature”)

Decision support exists but not linked to order writing, or prompts within orders but no decision support

Protocol well-integrated

(into orders at point-of-care)

Protocol enhancedProtocol enhanced

(by other QI / high reliability strategies)(by other QI / high reliability strategies)

Oversights identified and addressed in Oversights identified and addressed in real timereal time

Level

4

1

2

3

5

Predicted

Prophylaxis rate

40%

50%

65-85%

90%

95+%

* Protocol = standardized decision support, nested within an order set, i.e. what/when

The Essential First Intervention

1) a standardized VTE risk assessment, linked to…2) a menu of appropriate prophylaxis options, plus…3) a list of contraindications to pharmacologic VTE

prophylaxis

Challenges: Make it easy to use (“automatic”)

Make sure it captures almost all patientsTrade-off between guidance and ease of use /

efficiency 32

VTE Protocol

Map to Reach Level 3Implementing an Effective VTE Prevention

Protocol• Examine existing admit, transfer, periop order

sets with reference to VTE prophylaxis.• Design a protocol-driven DVT prophylaxis order

set (w/ integrated risk assessment model [RAM])• Vette / Pilot – PDSA• Educate / consensus building• Place new standardized DVT order set ‘module’

into all pertinent admit, transfer, periop order sets.

• Monitor, tweak - PDSA

34

Is your order set in a competition?

Too Little GuidancePrompt ≠ Protocol

DVT PROPHYLAXIS ORDERS

Anti thromboembolism Stockings Sequential Compression Devices UFH 5000 units SubQ q 12 hours UFH 5000 units SubQ q 8 hours LMWH (Enoxaparin) 40 mg SubQ q day LMWH (Enoxaparin) 30 mg SubQ q 12 hours No Prophylaxis, Ambulate

No Math!Critiques of VTE Risk Assessment

Model using point scoring techniques

• Point based systems -– low inter-observer agreement in real use– users stop adding up points– too large to be modular (collects dust)– point scoring is arbitrary– never validated

Low Medium HighAmbulatory with no other risk factors. Same day or minor surgery

CHF

COPD / Pneumonia

Most Medical Patients

Most Gen Surg Patients

Everybody Else

Elective LE arthroplasty

Hip/pelvic fx

Acute SCI w/ paresis

Multiple major trauma

Abd / pelvic CA surgery

Early ambulation

UFH 5000 units q 8 h (5000 units q 12 h if > 75 or weight <50 kg)

LMWH Enox 40 mg q day

Other LMWH

CONSIDER add IPC

Enox 30 mg q 12 h or

Enox 40 q day or

Other LMWH or

Fondaparinux 2.5 mg q day or

Warfarin INR 2-3

AND MUST HAVE

IPC 37

IPC needed if contraindication to AC exists

Example from UCSD Keep it Simple – A “3 bucket” model

Paper Version – “3 Bucket” RAM DVT Prophylaxis Order Set

Module

See separate paper version demonstrating 3 bucket model

Integrate order set as a module

• Make order set even more portable

• Incorporate module into current heavily used order sets

Or

Strip out VTE orders from popular order sets and refer to the standardized orders

Clip orders to all admit / transfer orders

Most Common Mistakes in VTE Prevention Orders

• Point based risk assessment model

• Improper Balance of guidance / ease of use– Too little guidance - prompt ≠ protocol

– Too much guidance- collects dust, too long

• Failure to revise old order sets

• Too many categories of risk

• Allowing non-pharm prophy too much

• Failure to pilot, revise, monitor

• Linkage between risk level and prophy choices are separated in time or space

Hierarchy of Reliability

No protocol* (“State of Nature”)

Decision support exists but not linked to order writing, or prompts within orders but no decision support

Protocol well-integrated

(into orders at point-of-care)

Protocol enhancedProtocol enhanced

(by other QI / high reliability strategies)(by other QI / high reliability strategies)

Oversights identified and addressed in Oversights identified and addressed in real timereal time

Level

4

1

2

3

5

Predicted

Prophylaxis rate

40%

50%

65-85%

90%

95+%

* Protocol = standardized decision support, nested within an order set, i.e. what/when

Daily measurement drives concurrent intervention

(i.e. same as Level 5 in Hierarchy of Reliability)

Identify patients not receiving VTE prophylaxis in real time

1) Suitable for ongoing assessment, reporting to governing body

Archive-able data (!)

2) Can be used for real time interventionActionable data (!)

42

Measure-vention

Map to Reach Level 595+ % prophylaxis

• Use MAR or Automated Reports to Classify all patients on the Unit as being in one of three zones:

GREEN ZONE - on anticoagulationYELLOW ZONE - on mechanical

prophylaxis only RED ZONE – on no prophylaxis

Act to move patients out of the RED!

Situational Awareness and Measure-vention: Getting to

Level 5• Identify patients on no anticoagulation• Empower nurses to place SCDs in

patients on no prophylaxis as standing order (if no contraindications)

• Contact MD if no anticoagulant in place and no obvious contraindication– Templated note, text page, etc

• Need Administration to back up these interventions and make it clear that docs can not “shoot the messenger”

45

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86

Pre

vale

nce o

f V

TE

Pro

ph

yla

xis

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

Pre

vale

nce o

f V

TE

Pro

ph

yla

xis

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

Pre

vale

nce o

f V

TE

Pro

ph

yla

xis

Effect of Situational Awareness on Prevalence of VTE Prophylaxis by

Nursing Unit

Hospital A, 1st Nursing Unit Baseline Post-Intervention

UCL: 93% 104%Mean: 73% 99% (p < 0.01)LCL: 53% 93%

Hospital A, 2nd Nursing Unit Baseline Post-Intervention

UCL: 90% 102%Mean: 68% 87% (p <

0.01)LCL: 46% 72%

Hospital B, 1st Nursing Unit Baseline Post-Intervention

UCL: 89% 108%Mean: 71% 98% (p <

0.01)LCL: 53% 88%______________________________________________UCL = Upper Control Limit UCL = Upper Control Limit LCL = Lower Control LimitLCL = Lower Control Limit

Hospital Days

Intervention

Intervention

Intervention

Most Common Mistakes in Measurement of DVT

Prophylaxis

• Not doing it at all

• Not doing it concurrently

• Failure to make measured poor performance actionable

Key Points - Recommendations

• QI building blocks should be used• Multifaceted approach is needed• VTE protocols embedded in order sets• Simple risk stratification schema, based on VTE-

risk groups (3 levels of risk should do it)• Institution-wide if possible (a few carve outs ok)• Local modification is OK

– Details in gray areas not that important

• Use measure-vention to accelerate improvement

47

Maynard G, Morris T, Jenkins I, Stone S, Lee J, Renvall M, Fink E, Schoenhaus R (2009) Optimizing prevention of hospital acquired venous thromboembolism: prospective validation of a VTE risk assessment model. J Hosp Med 4(7). doi:10.1002/jhm.562

Maynard G, Stein J. Preventing Hospital-Acquired Venous Thromboembolism: A Guide for Effective Quality Improvement. Prepared by the Society of Hospital Medicine. AHRQ Publication No. 08-0075. Rockville, MD: Agency for Healthcare Research and Quality. August 2008, last accessed September 15, 2008 at http://www.ahrq.gov/qual/vtguide/.

Maynard G, Stein J. Preventing Hospital-Acquired Venous Thromboembolism: A Guide for Effective Quality Improvement, version 3.3. Society of Hospital Medicine supplement The Hospitalist August 2008, Vol 12 (8) 1-40.

Maynard G, Stein J. Designing and Implementing Effective VTE Prevention Protocols: Lessons from Collaboratives. J Thromb Thrombolysis DOI 10.1007/s11239-009-0405-4 published online Nov 10, 2009