Preventing Hospital-Associated Venous Thromboembolism

Appendix B: Risk Assessment Models, Protocols, and Order Sets (continued)

B.10 Caprini Individualized Point-Based (Quantitative) Model of VTE Risk Assessment

This Model of VTE Risk Assessment uses a point system to assess risk factors. Risk Factors representing 1 point each: Age 41-60 years; Acute myocardial infarction; Swollen legs (current); Congestive heart failure (less than one month ago); Varicose veins; Medical patient currently at bed rest; Obesity (BMI greater than 25); History of inflammatory bowel disease; Minor surgery planned; History of prior major surgery (less than one month ago); Sepsis (less than one month ago); Abnormal pulmonary function (COPD); Serious lung disease including pneumonia (less than one month ago); Oral contraceptives or hormone replacement therapy; Pregnancy or postpartum (less than one month ago); History of unexplained stillborn infant, recurrent spontaneous abortion (3 or more) , premature birth with toxemia, or growth-restricted infant; Other risk factors. Risk Factors representing 2 points each: Age 61-74 years; Central venous access; Arthroscopic surgery; Major surgery (more than 45 minutes); Malignancy (present or previous); Laparoscopic surgery (more than 45 minutes); Patient confined to bed (more than 72 hours); Immobilizing plaster cast (less than one month ago). Risk Factors representing 3 points each: Age 75 years or older; Family history of thrombosis (most frequently missed risk factor); History of DVT/PE; Positive Prothrombin 20210A; Positive Factor V Leiden; Positive Lupus anticoagulant; Elevated serum homocysteine; Heparin-induced thrombocyptopenia (HIT) (Do not use heparin or any low molecular weight heparin); Elevated anticardiolipin antibodies; Other congenital or acquired thrombophilia - If yes, Type. Risk Factors representing 5 points each: Stroke (less than one month ago); Multiple trauma (less than one month ago); Elective major lower extremity arthroplasty; Hip, pelvis or leg fracture (less than one month ago); Acute spinal cord injury (paralysis) (less than one month ago). Each relevant risk factor is to be checked and points subtotaled for each section, then added together for a Total Risk Factor Score.

Factors Associated with Increased Bleeding-Patient may not be a candidate for anticoagulant therapy

Consider SCDs if: Active bleeding. Already receiving prophylaxis, BMT patient without indication for anticoagulation per UM protocol, Glycoprotein IIB/IIIA inhibitors, hemophilia or significant bleeding disorder, recent CNS bleed, intracranial or spinal lesion at high risk of bleeding, recent major operation at high risk for bleeding, systemic anticoagulant (non warfarin or INR > 2.0), thrombocytopenia (platelets > 50,000)

Clinical Considerations for the Use of Sequential Compression Devices-Patient may not be a candidate for SCDs

Alternative prophylactic measures should be considered if: Severe peripheral vascular disease (ABPI ≤ 0.5), Severe CHF, compartment syndrome of affected extremity, fracture of affected extremity, local conditions such as: gangrene, recent skin graft, or open wound of the affected extremity, known or suspected acute/subacute DVT (apply SCDs to contralateral limb if indicated)

Total Risk Factor Score Risk Level Prophylaxis Regimen
0 Very low __ Early ambulation
1-2 Low __ Sequential Compression Device (SCD)
3-4 Moderate Choose ONE of the following medications +/- compression devices:
__ Sequential Compression Device (SCD) - Optional
__ Heparin 5000 units SQ TID
__ Enoxaparin/Lovenox:
     __ 40mg SQ daily (WT < 150kg, CrCl > 30mL/min)
     __ 30mg SQ daily (WT < 150kb, CrCl = 10-29mL/min)
     __ 30mg SQ BID (WT > 150kg, CrCl > 30mL/min)
(Please refer to Dosing Guidelines on the back of this form)
5 or more High Choose ONE of the following medications PLUS compression devices:
__ Sequential Compression Device (SCD)
__ Heparin 5000 units SQ TID (Preferred with Epidurals)
__ Enoxaparin/Lovenox (Preferred):
     __ 40mg SQ daily (WT < 150kg, Crcl > 30mL/min)
     __ 30mg SQ daily (WT < 150kb, CrCl = 10-29mL/min)
     __ 30mg SQ BID (WT > 150kg, CrCl > 30mL/min)
(Please refer to Dosing Guidelines on the back of this form)

__ Ambulatory Surgery – No orders for venous thromboembolic prophylaxis required
__ VTE Prophylaxis Contraindicated, Reason: _______________________________

UMHS Enoxparin Dosing Guidelines

*ABW = Actual Body Weight

Non-Pregnant Patients

CrCl ≥ 30 mL/min:

Enoxaparin 40 mg SQ once daily (for ABW ≤ 150 kg, non-trauma patient)
Enoxaparin 30 mg SQ every 12 hours (for ABW > 150 kg)
Enoxaparin 30 mg SQ every 12 hours (regardless of ABW for trauma patients)

CrCl < 30 mL/min: Enoxaparin 30 mg SQ once daily (regardless of ABW)

Pregnant Patients

Less than 20 weeks gestation: Enoxaparin 40 mg SQ once daily

20 weeks gestation until 1 week post-partum: Enoxaparin 40 mg SQ every 12 hours

Post-partum weeks 2-6: Enoxaparin 40 mg SQ once daily

*Actual body weight should be used for dose determination

Monitoring Recommendations

Antifactor Xa activity monitoring is not recommended for Enoxaparin prophylaxis dosing.

Source: University of Michigan Health System. Used with permission.                                     


  • It was found to be predictive of VTE risk in several different surgical populations.
  • A large number of VTE risk factors are included, which may increase sensitivity.
  • AT9 guidelines use Caprini scores to help define risk categories.
  • While it is most commonly used and studied in surgical patients, it is amenable to use in medical and surgical populations.


  • It is fairly labor intensive and time consuming and requires serial calculations.
  • It has poor interobserver agreement and variation in scoring noted in studies.
  • Many institutions have tried the model but have discarded it after they found physicians would either use it incorrectly or bypass it altogether.
  • It may require advanced clinical decision support and/or a safety net to achieve reliable adherence.
  • IPC is offered as an option for primary prophylaxis for scores of 3-4, but IPC adherence is typically low. Careful monitoring and interventions to keep adherence high are required for institutions offering this option.

Variations and other comments:

  • Variation: AT9 does not endorse extended duration prophylaxis for those with scores >8 unless they fall into groups such as abdominal or pelvic surgery for cancer or major orthopedic surgery. Scores of 5–8 and >8 would be prophylaxed the same way.
  • Variation: Many prior iterations with different menus of options for different scores exist in the literature, generally with a lower threshold for anticoagulant prophylaxis.
  • Variation: For scores of 3–4, offer only anticoagulant, or only IPC, depending on the population the order set serves (e.g., CABG patients get IPC but abdominal surgery patients would get LMWH or UFH, instead of both being offered).


  1. Caprini JA. Risk assessment as a guide for the prevention of the many faces of venous thromboembolism. Am J Surg 2010;199(1S):S3-S10. Accessed May 6, 2016.
  2. Pannucci CJ, Bailey SH, Dreszer G, Validation of the Caprini risk assessment model in plastic and reconstructive surgery patients. J Am Coll Surg 2011 Jan; 212(1):101-12. Accessed May 6, 2016.
  3. Beck MJ, Haidet P, Todoric K, et al. Reliability of a point-based VTE risk assessment tool in the hands of medical residents. J Hosp Med 2011;6:195-201.

Return to Contents

B.11. Brigham and Women's Hospital—Weighted Scoring System Used for Alerts

Risk factor Score
Active cancer 3
Previous VTE (excluding superficial thrombosis) 3
Already known thrombophilic condition/hypercoagulability 3
Major surgery 2
Age ≥70 years 1
Obesity (BMI ≥30) 1
Bed rest/immobility 1
Ongoing hormonal treatment 1

Note: Patients with a score ≥4 and on no prophylaxis were sent an electronic alert.


  • One of the more simple and straightforward quantitative point-based models.
  • Demonstrated to increase prophylaxis and decrease VTE by 41 percent at BWH when alert sent to providers of patients with scores ≥4 but patient not on prophylaxis.
  • Required physicians to acknowledge computer alert, and they could order or withhold prophylaxis at their discretion.
  • Accepted by medical staff.


  • Designed to capture patients at obvious risk for VTE. Many common risk factors (e.g., congestive heart failure, chronic obstructive pulmonary disease, active infection/inflammation) not included. Therefore, this model is specific to those at VTE risk but is not a sensitive instrument to capture all patients at risk.
  • Never tested or tried as VTE risk assessment model in admission order sets and not particularly well suited for that purpose.
  • Requires integrated databases and clinical decision support sophistication for an automated alert strategy. Most hospitals do not have these.
  • Subject to alert fatigue, which can lessen efficacy over time, and two-thirds of providers rejected guidance.

Variations and other comments:

  • Padua risk assessment model derived from this model.
  • Screen shots of alert system available in reference 2 below.


  1. Kucher N, Koo S, Quiroz R, et al. Electronic alerts to prevent venous thromboembolism among hospitalized patients. N Engl J Med 2005 Mar 10;352(10):969-77. Accessed May 6, 2016.
  2. Goldhaber SZ. Rationale supporting an "opt out" policy for pharmacological venous thromboembolism prophylaxis in hospitalized medical patients. J Thromb Thrombolysis 2013;35:371-4.
  3. Piazza G, Rosenbaum EJ, Pendergast W, et al. Physician alerts to prevent symptomatic venous thromboembolism in hospitalized patients. Circulation 2009;119:2196-2201. Accessed May 6, 2016.

Return to Contents

B.12. Padua Risk Assessment Model for Nonsurgical Inpatients

Note: High risk was defined as a score ≥4. In this study, more than 1 percent of those with a score of 3 experienced a pulmonary embolism. Improvement teams should consider defining high risk as a score ≥3, versus 4.

Baseline Features Score
Active cancer * 3
Previous VTE (excluding superficial thrombosis) 3
Reduced mobility  3
Already known thrombophilic condition  3
Recent (≤1 month) trauma and/or surgery 2
Elderly age (≥70 years) 1
Heart and/or respiratory failure 1
Acute myocardial infarction or stroke 1
Acute infection and/or rheumatologic disorder 1
Obesity (BMI ≥30) 1
Ongoing hormonal treatment 1

* Patients with local or distant metastases and/or chemotherapy or radiation therapy in last 6 months.
† Bed rest with bathroom privileges (either due to patient limitations or physician order) for ≥3 days.
‡ Carriage of defects of antithrombin, protein C or S, factor V Leiden, G20210A prothrombin mutation, antiphospholipid syndrome.


  • One of the more simple and straightforward quantitative point-based models.
  • Score of ≥4 was highly predictive of VTE risk in an Italian cohort study, and those <4 had relatively low risk.


  • It was never tested or shown to be effective as a VTE risk assessment model in order sets.
  • While relatively straightforward, it calls for calculation by providers.
  • It uses more common risk factors compared with the Brigham and Women's Hospital (BWH) model, but many (e.g., varicose veins, dehydration, nephrotic syndrome) are not incorporated.
  • More than 1 percent of patients with a score of 3 experienced pulmonary embolism. The cutoff point was arbitrarily modeled after the BWH model without formal analysis.
  • This model is only for medical inpatients. Institutions using this model would need to have separate risk assessment strategies for surgical patients and other populations.
  • The operational definition of reduced mobility calls for prediction of mobility on admission, which may lead to inconsistent application of the standard and difficulty monitoring for protocol adherence. Also, a physician order for "up in chair" or "activity ad lib" correlates poorly with actual activity and is no guarantee of a protective degree of ambulation. Many institutions prefer definitions such as "able to ambulate outside room independently" or "ambulates at least 50 feet" and look at current ambulation, rather than predicted ambulation. (See case scenarios 1, 3, and 4. How would they fit the Padua definition of reduced mobility? Would different providers give the same answer?)


  • Variation: Use cutoff of ≥3 for definition of high risk, rather than ≥4.
  • Variation: See B6. This changes the Padua quantitative model into a "bucket" or qualitative model (with cutoff of ≥3), which some may find easier to implement.


  1. Barbar S, Noventa F, Rossetto V, et al. A risk assessment model for the identification of hospitalized medical patients at risk for venous thromboembolism: the Padua Prediction Score. J Thromb Haemost 2010;8(11):2450-7. Accessed May 9, 2016.
  2. Maynard G, Jenkins IH, Merli GJ. VTE prevention guidelines for medical inpatients: mind the (implementation) gap. J Hosp Med 2013;8(10):582-8. doi:10.1002/jhm.2071. Epub 2013 Aug 27. PMID:23983041.

Return to Contents

B.13. Medical VTE Order Set, Using Padua Risk Model Depicted as a "Bucket" or Grouping Model, Instead of a Point Model, if High Risk Defined as Score ≥3 (versus 4)

Venous Thromboembolism Risk Assessment and Prophylaxis Physician Order Sheet

Complete Assessment at Admission, Post-Op, And Transfer

DVT/ PE Risk Level & Prophylaxis Orders

___ Moderate to High Risk

  • Active cancer (with mets, or on active therapy in last 6 months).
  • Previous VTE or known thrombophilia.
  • Reduced mobility and acute illness.


  • Recent trauma (≤1 month) and/or surgery and any other listed risk factor.


  • Any 3 of the following:
    • Elderly age (>70 years).
    • Heart or respiratory failure.
    • Acute MI or stroke.
    • Acute infection and/or rheumatologic disorder.
    • Obesity (BMI ≥30).
    • Ongoing hormonal treatment.
  • ___ Enoxaparin 40 mg SC q 24 hrs.
  • ___ Enoxaparin 30 mg SC q 24 hrs (renal insufficiency dosing).
  • ___ Heparin 5,000 units SC q 8 hrs.
  • ___ Heparin 5,000 units SC every 12hrs (if weight <50kg or age >75).

___ Low Risk.

  • Observation patients; expected LOS <48 hrs.
  • Fully ambulatory without multiple VTE risk factors.
  • Patients no longer/never ill, awaiting disposition.
  • Patients not meeting criteria for Moderate to High Risk.
  • ___ Early ambulation, education.
  • ___ Education.


Contraindication to pharmacologic prophylaxis, including already on therapeutic anticoagulation
(see reverse): ________________________________

  • Mechanical prophylaxis with sequential compression device OR
  • Contraindicated (peripheral vascular disease or wounds).


Signature / Provider ID Date / Time


  • Relatively intuitive and easy to use, no math or calculation required.
  • Fairly simple to monitor and audit.
  • Captures VTE risk, and risk level is directly linked to appropriate choices for prophylaxis.
  • Aligned with Padua risk assessment model featured in AT9 guidelines.


  • The caveats are the same as for the Padua model (i.e., many VTE risk factors not included, Padua model never prospectively verified as effective in order sets, and model based on one Italian study).
  • Operational definition of mobility needs definition (in the original study it was based on limited mobility for 3 or more days, which would call for prediction at the time of admission or on physician orders, which often do not correlate with what the patient is actually achieving).
  • Separate risk model for surgical populations is required, which may cause confusion.
  • Padua study authors endorsed high risk as a score of ≥4, but this version is consistent with high risk equating to a score of ≥3. This cutoff should definitely be considered, as more than 1 percent of patients with a score of 3 had pulmonary embolism, but there is no consensus on this issue.


  • Variation: Many hospitals include critically ill patients in ICU settings to highest risk groups that receive combination mechanical and anticoagulant prophylaxis. (Note that the benefit of this practice has little direct evidence and is extrapolated from high-risk surgical patients).
  • Variation: Blend B.6 (classic 3 bucket model) and this model to serve medical and surgical patients.


  1. Barbar S, Noventa F, Rossetto V, et al A risk assessment model for the identification of hospitalized medical patients at risk for venous thromboembolism: the Padua Prediction Score. J Thromb Haemost 2010; 8(11):2450-7. Accessed May 9, 2016.
  2. Maynard G, Jenkins IH, Merli GJ. VTE prevention guidelines for medical inpatients: mind the (implementation) gap. J Hosp Med 2013;8(10):582-8. doi:10.1002/jhm.2071. . Epub 2013 Aug 27. PMID:23983041.
  3. Kahn SR, Lim W, Dunn AS, et al. Prevention of VTE in nonsurgical patients. Chest 2012;141:2 suppl e195S-e226S; doi:10.1378/chest.11-2296. Accessed May 9, 2016.

Return to Contents

B14. IMPROVE VTE Risk Model

IMPROVE VTE Risk Model—4 Factor Version (Risk Factors Available on Admission)

VTE Risk Factor Points
Previous VTE 3
Known thrombophilia 3
Cancer 1
Age >60 years 1

IMPROVE VTE Risk Model—7 Factor Version (Includes Risk Factors That May Develop During the Hospital Stay)

VTE Risk Factor Points
Previous VTE 3
Known thrombophilia 2
Cancer 2
Current lower limb paralysis 2
Immobility ≥7 days 1
ICU/CCU stay 1
Age >60 years 1

Return to Contents

B15. Case Scenarios to Test Models and Example Order Sets

These cases are provided to encourage a test drive of the models and example order sets that follow. Instead of evaluating each risk assessment model (RAM) in a vacuum, assemble several physicians or other ordering providers from representative services and have them apply the RAM/order set to these cases or other case scenarios that you may construct. Afterward, debrief them and see if they filled out the order set in a consistent manner and have feedback about ease of use and clarity.

This exercise is also useful in raising explicit questions about operational definitions for important parameters such as ambulation, bleeding risk, and time interval after surgery when it is considered safe to start anticoagulant prophylaxis.

Case 1
A 69-year-old male is admitted from the emergency department to a noncritical care unit with shortness of breath x 3-4 days. He has subjective fever and cough, a past medical history of congestive heart failure, chronic obstructive pulmonary disease, hypertension, and hyperlipidemia. He is a one-pack per day smoker.

Physical exam confirms a temperature of 101° F, respiratory rate 22/minute, pulse 108/minute, BP 120/70.

Height – 67 inches; Weight – 200 pounds (91 kg).

Obese moderately dyspneic male in bed, PICC line placed in ED in left arm.

Dull breath sounds, right lower base. Actively wheezing with a prolonged expiratory phase.

Cor-RRR without obvious S3 or murmur.

2+ pedal edema with acute and chronic stasis, bilateral varicose veins.

Case 2
A 40-year-old man is admitted to the hospital for acute myocardial infarction and congestive heart failure thought to be associated with polysubstance abuse, including tobacco. The patient has no prior history of cardiovascular disease. He was in the hospital 3 weeks ago for sepsis associated with skin-popping-induced skin abscesses and cellulitis, from which he seemed to fully recover. He is at bed rest on admission. He has no other known VTE risk factors.

Physical exam shows low-grade fever. RR 14, pulse 94, BP 97/60.

Height – 70 inches; Weight – 161 pounds (73 kg).

Well-developed male in no acute distress, oxygen at 2 liters per nasal cannula in place.

Some jugular vein distention present. Cardiac exam – Normal except S3 gallop. Chest – basilar rales.

Extremities – signs of skin popping. Peripheral IV in place, left arm. No edema.

Case 3
A 61-year-old woman is admitted after a syncopal episode. In the emergency department she was found to have orthostasis and an elevated BUN/creatinine, attributed to over-diuresis with furosemide. She has a history of nephrotic syndrome, massive proteinuria, and anasarca. The nephrotic syndrome is thought to be secondary to primary membranous nephropathy. She had a mastectomy for breast cancer and radiation therapy about 10 months ago, but the details are not available. She is not on hormonal therapy. She generally is able to ambulate within her home with a walker, but is admitted with fall precautions and instructions to not get out of bed without assistance, pending rehydration. She has no history of liver disease.

Physical shows she is afebrile. RR 14, pulse 100, BP 120/70 (supine).

Height – 62 inches; Weight – 150 pounds (68 kg).

Moderately obese female in no acute distress, lying down, alert and oriented, edematous.

Pitting edema to buttocks, moderate ascites on exam, dull to auscultation at lung bases. Bilateral varicose veins.

Case 4
A 49-year-old obese woman is admitted overnight for observation after arthroscopic knee surgery. She is placed in a cast and crutches postoperatively. She is on oral contraceptives and has a past medical history of pulmonary emboli after colectomy for inflammatory bowel disease 4 years ago.

Physical exam post op in recovery room: Afebrile. RR 12, pulse 98, BP 130/80.

Obese female still groggy post-anesthesia. No acute distress.

Chest – clear to auscultation. No rub, murmur, or gallop on cardiac exam, no JVD. Graduated compression stockings are in place, 1+ peripheral edema.

Case 5
A 64-year-old man is admitted for colorectal surgery to remove colonic neoplasm and associated mesenteric lymphadenopathy. He is a smoker with COPD and on inhalers. He is generally ambulatory and has a high level of function. He has a PICC line in for venous access, as attempts to obtain peripheral venous access failed.

Physical exam reveals a well-appearing man with normal vital signs.

Next Page
Return to Contents

Page last reviewed May 2016
Page originally created May 2016
Internet Citation: Appendix B: Risk Assessment Models, Protocols, and Order Sets (continued). Content last reviewed May 2016. Agency for Healthcare Research and Quality, Rockville, MD.