reducing postoperative venous thromboembolism ... · vte risk stratiﬁcation and commensurate...

Disclosure Information: Nothing to

Presented at the New England SuHartford, CT, September 2013.

Received October 14, 2013; RevDecember 17, 2013.From the Department of Surgery, Band Boston Medical Center, Boston,Corresponding address: David McACenter, Boston University SchoolOncology and Endocrine Surgery, 8Boston, MA 02118. email: David.M

ª 2014 by the American College of S

Published by Elsevier Inc.

NEW ENGLAND SURGICAL SOCIETY ARTICLES

Reducing Postoperative Venous ThromboembolismComplications with a Standardized Risk-StratifiedProphylaxis Protocol and Mobilization Program

Michael R Cassidy, MD, Pamela Rosenkranz, RN, BSN, MEd, David McAneny, MD, FACS

BACKGROUND: Data revealed that our urban, academic, safety net medical center was a high outlier for post-operative venous thromboembolism (VTE). Our goal was to implement and determine theefficacy of a standardized intervention for reducing postoperative VTE complications.

STUDY DESIGN: We developed a strategy to decrease VTE complications, based on standardized electronicphysician orders that specify early postoperative mobilization and mandatory VTE risk strat-ification for every patient, using the “Caprini” grading system. The derived scores dictate thenature and duration of VTE prophylaxis, including on an outpatient basis. Electronicreminders about appropriate VTE prophylaxis are automatically generated before and afteroperations, and on discharge. Both mechanical (pneumatic compression boots) and pharma-cologic prophylaxis (unfractionated or low molecular weight heparin) are used, as indicatedby risk level. We conducted a before-and-after trial, comparing National Surgical QualityImprovement Program (NSQIP) VTE outcomes (deep vein thromboses and pulmonaryemboli) before and after implementing the standardized risk-stratified protocol combinedwith a postoperative mobilization program. Measured outcomes included NSQIP-reportedraw and risk-adjusted VTE outcomes during 2 years before and after implementing theVTE prevention program.

RESULTS: The incidence of deep venous thromboses decreased by 84%, from 1.9% to 0.3% (p < 0.01),with implementation of VTE prevention efforts; the pulmonary emboli incidence fell by55%, from 1.1% to 0.5% (p < 0.01). Risk-adjusted VTE outcomes steadily declinedfrom an odds ratio of 3.41 to 0.94 (p < 0.05).

CONCLUSIONS: A patient care program, emphasizing early postoperative mobilization along with mandatoryVTE risk stratification and commensurate electronic prophylaxis recommendations, signifi-cantly reduced the likelihood of VTE complications among our patients. (J Am Coll Surg2014;218:1095e1104. � 2014 by the American College of Surgeons)

Postoperative venous thromboembolism (VTE) events,which include deep venous thromboses (DVT) and pul-monary emboli (PE), are a leading cause of morbidityand mortality. In the United States, the estimated annualincidence of VTE is 117 per 100,000.1 Among patientswho undergo abdominal operations, symptomatic VTE

disclose.

rgical Society 94th Annual Meeting,

ised December 9, 2013; Accepted

oston University School of MedicineMA.neny, MD, FACS, Boston Medicalof Medicine, Section of Surgical20 Harrison Ave, FGH Suite 5003,[email protected]

1095urgeons

occurs in 0.4% to 3.1%.2 Pulmonary emboli may causesudden death and may independently reduce survivalfor up to 3 months after diagnosis.3 Those who livemay develop pulmonary hypertension. Deep venousthromboses result in venous hypertension, which canlead to debilitating swelling and chronic pain.4 One studyestimated the attributable cost of a VTE complication tobe $18,310.5 Interestingly, in some series, the majority ofVTE events in surgery patients occur after discharge fromthe hospital.2

Surgery patients are at particular risk for VTE becausemajor operations induce an inflammatory response and ahypercoagulable state, and necessarily create an endothe-lial vascular insult. Furthermore, the tendency for patientsto limit mobilization due to postoperative pain may pre-dispose to venous stasis.6 Other commonly encounteredrisk factors include older age, cancer, trauma, obesity,

ISSN 1072-7515/14/$36.00

http://dx.doi.org/10.1016/j.jamcollsurg.2013.12.061

mailto:[email protected]

http://dx.doi.org/10.1016/j.jamcollsurg.2013.12.061

Abbreviations and Acronyms

DVT ¼ deep venous thrombosisO/E ¼ observed to expected ratioOR ¼ odds ratioPE ¼ pulmonary embolismVTE ¼ venous thromboembolism

1096 Cassidy et al Reducing Venous Thromboembolism Complications J Am Coll Surg

sepsis, and inherited coagulopathies. The concept of a riskstratification score, calculated by assigning weighted nu-merical values to VTE risk factors, has been championedby Joseph Caprini and others.7,8 The Caprini system rec-ognizes that numerous factors confer differing degrees ofhazards, and it assigns relative values to those attributes toderive an estimate of VTE likelihood. The Caprini scorehas been shown to accurately predict the chances of aVTE in surgical patients and has guided prophylaxisdecisions.Comprehensive prophylaxis recommendations have

been proposed by the American College of Chest Physi-cians. These include strong endorsement of low doseunfractionated heparin, low molecular weight heparin,or fondaparinux as chemoprevention, in addition to inter-mittent compression boots and mobilization if possible,for all patients undergoing major operations.9 To achievemaximum benefit, certain high risk patients may requireextended VTE prophylaxis, including after discharge.10,11

Unfortunately, there is evidence that prophylaxis measuresare often underused, with at-risk patients receiving inap-propriate or no prophylaxis.12,13 A large multinationalstudy revealed that only 59% of surgical patients receivedevidence-based VTE prophylaxis.14

National Surgical Quality Improvement Program(NSQIP) data revealed that our medical center, the largestsafety-net medical center in New England and an aca-demic, urban hospital, was a high outlier for VTE. BostonMedical Center (BMC) is a merged entity of the formerBoston University Hospital and Boston City Hospital,with 509 licensed beds. More than half of its patientshave an annual income below $20,420. About one-quarter of the patients do not speak English, and racialand ethnic minorities constitute 70% of all patients.Recognizing an opportunity for improvement, we soughtto create a standardized VTE prevention program and todemonstrate its efficacy.

METHODS

Development of a standardized venousthromboembolism prevention program

A VTE prevention team met to consider strategies toreduce VTE complications. After reviewing the relevant

literature, the consensus was to use the Caprini risk strat-ification method for all general surgery and vascular sur-gery patients at our institution. Among the severalavailable VTE prophylaxis programs, we selected theCaprini system because it is adaptable to individual pa-tients’ risk factors, less likely to underestimate the hazardsof VTE, and is well validated. We developed a scoringsystem and integrated it into the electronic inpatientmedical record (Sunrise Acute Care, Allscripts). The sys-tem uses a check-box format so that each risk factor isexplicitly listed and may be selected with a simple click(Fig. 1). The risk score is automatically calculated basedon the selected factors, and the patient is placed into 1of 5 risk categories (lowest, low, moderate, high, or high-est risk). Our electronic order system is customized torequire that a Caprini score be calculated for every patientat the time of operation and/or admission within generalsurgery and vascular surgery standardized order sets. If thesurgery team does not calculate the Caprini score and acton the electronic recommendations, the orders cannot becompleted. Therefore, we made an effort to ensure thateach patient would be scored according to the Caprinimodel.In addition, we created standardized VTE prophylaxis

regimens and linked them to the Caprini risk categories.The prophylaxis regimens provide the recommendedmechanical and pharmacologic prophylaxis along with asuggested duration (Table 1), which is automatically dis-played. The duration of chemoprophylaxis may requireextended regimens continued on an outpatient basis.For example, patients whose Caprini scores place themin the high risk category (scores 5 to 8) are advised toreceive 7 to 10 days of chemoprophylaxis; those in thehighest risk category (scores � 9) are advised to receivea 30-day course of chemoprophylaxis, both of which typi-cally require outpatient treatment. The electronic ordersystem is designed to require that all patients receive stan-dardized prophylaxis regimens. When the discharge or-ders are recorded, the recommended VTE prophylaxisis automatically displayed for patients who require anextended course.Our electronic order system requires selection of a

prophylaxis regimen. Nevertheless, a surgeon may stilldecline VTE prophylaxis, when it is contrary to his orher judgment, by choosing the “opt out” selection inthe order sets. This prompts an automatic drop-downmenu that indicates reasons for not prescribing VTEchemoprophylaxis. The selections include active bleeding,risk of hemorrhage outweighing risk of VTE, surgeonpreference, heparin allergy or contraindication, or otherreason (supported by an explanation in the medical re-cord). As a result, the order set even documents the reason

Figure 1. Mandatory electronic venous thromboembolism risk assessment tool.

Vol. 218, No. 6, June 2014 Cassidy et al Reducing Venous Thromboembolism Complications 1097

for not using prophylactic anticoagulation. Although theultimate decision for each patient receiving VTE prophy-laxis is made by the surgeons, the electronic reminders aredesigned to encourage adherence to a standardized pre-vention strategy. The electronic system was implementedand became fully operational in February 2011.We combined the requirement for Caprini risk stratifi-

cation and commensurate prophylaxis with a standard-ized postoperative mobilization program. We createdspecific standardized mobilization instructions andincluded them in order sets used for all general surgeryand vascular surgery patients. The nursing orders requirethat each patient be out of bed at least 3 times daily,beginning on the day of the operation. When possible,

based on the physical capacity of the patient and specificsof the operation, early ambulation is encouraged. Nurseeducators and surgeons met with unit nurses, includingthose from the ICU, to review baseline outcomes dataand to establish expectations for mobilization. Attendingsurgeons and housestaff were also educated about VTEoutcomes data and prevention principles. The mobiliza-tion program was implemented in August 2010.

Audits of practice

During deliberations about the risk-stratified program, wereviewed preintervention VTE prophylaxis practices of oursurgeons. After introduction of the program, we trackedcompliance with the new VTE prevention guidelines by

Table 1. Recommended Prophylaxis Regimens and Duration Based on Caprini Score and Risk Stratification

Caprini score Risk category Recommended prophylaxisRecommended duration of

chemoprophylaxis

0 Lowest Early frequent ambulation only, ORAt discretion of surgical team:Compression boots ORLow dose heparin ORLow molecular weight heparin

During hospitalization

1e2 Low Compression boots ORLow dose heparin ORLow molecular weight heparin(Choose 1 item)


3e4 Moderate Compression boots ANDLow dose heparin ORLow molecular weight heparin(Choose 1 medication)


5e8 High Compression boots ANDLow dose heparin ORLow molecular weight heparin(Choose 1 medication)

7e10 days total

>9 Highest Compression boots ANDLow dose heparin ORLow molecular weight heparin(Choose 1 medication)

30 days total


recording the percentage of general and vascular surgerypatients for whom the prescribed prophylaxis regimenwas consistent with the risk-stratified recommended pro-phylaxis. To be considered in compliance, both the natureand duration of VTE prophylaxis must have matchedthe recommended regimen for the patient’s risk score.Compliance was measured based on prescribing informa-tion for both inpatient and outpatient prophylaxis. Actualcompliance at home could not be confirmed. A retrospec-tive review of patient records was performed for the periodfrom February 2013 to August 2013.In order to understand baseline care of postoperative

patients at our institution before development of theVTE prevention program, we had audited mobilizationpractices in the spring of 2010. All patients who had un-dergone elective open abdominal or pelvic operationswere visited at 8:00 AM, 1:00 PM, and 6:00 PM onthe day of surgery and during the 2 subsequent days.Nurses were unaware of these audits. Trained clinical staffrecorded whether each patient was in bed, sitting in achair, or walking at the time of the visit. We definedoptimal practice as patients being out of bed, eithersitting in a chair or walking. Auditors were independentof both the ordering surgeons and the responsible nursingstaff, and included a quality improvement nurse, resi-dents, and medical students. The audits were observa-tional only and were not intended to directly alterpatient management. The mobilization program was fully

implemented in August 2010, and further audits wereperformed between 8 and 14 weeks after implementation.

Outcomes measures

WeusedNSQIP data from our institution to determine theimpact of the standardized VTE risk stratified prophylaxisprotocol and mobilization program. (Trained NSQIP sur-gical clinical reviewers collect clinical data rather than exclu-sively administrative data. This includes 30-day surveillanceof patients to capture events that occur after hospitaldischarge.) The NSQIP defines DVT as a new diagnosisof venous thrombosis, confirmed by imaging study or au-topsy, which is treated with anticoagulation or placementof vena cava filter. Criteria are also satisfied if the patient re-fuses treatment. A PE is defined as a new diagnosis of a newblood clot in a pulmonary artery, which is confirmed by im-aging or autopsy. All patients who underwent an operationon the general and vascular surgery services at our institu-tion during the specified time periods, and who wereaccrued to the NSQIP database, including those admittedto an ICU or to a non-ICU, were captured in our analysis.The absolute incidences of DVT and PE during the first 30postoperative days, as defined by NSQIP, were comparedbetween 1-year time periods before and after implementa-tion of the programs. Comparisons were made by usingthe chi-square test, with statistical significance set at p <0.05. SigmaStat statistical software (Systat) was used forall analyses. Additionally, we queried the NSQIP database


for the same outcomes during the same time periodsat comparable institutions, which we considered to beacademic medal centers with more than 500 beds. Wealso compared our risk-adjusted VTE outcomes, whichNSQIP reports as observed/expected ratios (O/E) for pe-riods before calendar year (CY) 2010, and as odds ratios(OR) for CY 2010 and later. The O/E and OR values areconsidered to be statistically comparable for large samplesizes. A risk-adjusted ratio of 1.0 indicates that the numberof observed events is equal to the number of events thatwould be expected based on the risk profile of the patients.Ratios exceeding 1.0 indicate a higher number of adverseevents than would be expected; ratios less than 1.0 suggestfewer adverse events than expected. Risk-adjusted data inthe NSQIP database account for patient comorbiditiesand severities of illness, which limit the effects that variationin patients’ disease characteristics between the 2 study pe-riods may have had on the measured outcomes. This studywas approved by the Institutional Review Board of the Bos-ton University School of Medicine.

RESULTS

Preintervention practice

Before development of our standardized program, no VTEprevention guidelines were formally used. Our surgeonsgenerally acknowledged the American College of ChestPhysicians guidelines, but no structured system existedand no individualized risk stratification was performed.Pneumatic compression boots and subcutaneous heparinwere frequently used, even in combination, but outpatientVTE chemoprophylaxis was rarely, if ever, used. Therewere no electronic reminders about VTE prophylaxis, andno surgeons used the Caprini system to guide decisions.Preintervention analysis of practice revealed that pa-

tients generally remained in bed more than desired.Mobilization orders were often absent or vague; forexample, orders might have simply stated “ambulate”without specifying a frequency. Nurses were not requiredto document details about ambulation.

Impact on practice

There was an excellent level of adherence to recommen-ded prophylaxis regimens after implementation of theelectronic risk-stratification and prophylaxis program.During the audit period, we observed 100% adherenceto recommended prophylaxis regimens among patientsstratified to low risk and moderate risk groups (n ¼749). Among patients stratified to the high-risk category(n ¼ 99), 89% received appropriate pharmacologic pro-phylaxis and duration. Ten percent of patients did notreceive the recommended prophylaxis, based on surgeon

discretion. One percent had a contraindication to phar-macologic prophylaxis. Adherence to the recommendedprophylaxis and duration was 77% for patients in thehighest risk category (n ¼ 13), with 23% of patients hav-ing contraindications to pharmacologic prophylaxis. Nopatients received inappropriate or inadequate prophylaxiswithout a documented reason, by virtue of the electronicsystem. Documented reasons for choosing an alternativeregimen included ambulatory operations, history ofheparin-induced thrombocytopenia, early terminationdue to low platelet count, outpatient warfarin therapy,and administration of clopidogrel. Only the patient dis-charged on clopidogrel instead of the recommended pro-phylaxis developed a VTE complication.Before introduction of a standardized early postop-

erative mobilization program, nursing practice auditsrevealed that only 19.6% of patients were out of bed atthe time of the visit (n ¼ 250). In contrast, postinterven-tion audits 8 to 14 weeks after implementation of theprogram revealed a significant difference in practice,with 69.1% of patients out of bed (n ¼ 250, p < 0.001).

Impact of risk-stratified prevention and mobilizationon incidence of venous thromboembolismcomplications

During calendar year 2009, before implementation of therisk-stratified prophylaxis and mobilization program, theincidence of DVT at our institution was 1.9% of 1,569patients (Fig. 2). The incidence of DVT fell by 84% to0.3% of 1,323 patients, by the reporting period of July2011 to June 2012, after implementation of VTE preven-tion efforts in 2010 and 2011 (p < 0.01). The averageincidence of DVT at comparable NSQIP hospitals (aca-demic medical centers with more than 500 beds) duringthe same time periods are presented in Figure 2 for pur-poses of comparison, and remained steady at 0.8%.The incidence of PE at our institution decreased by

55%, from 1.1% of 1,569 patients before the preventionprogram to 0.5% of 1,323 patients (Fig. 3, p < 0.01).The same outcomes at comparable NSQIP hospitals aredemonstrated in Figure 3, and remained 0.4% over time.In terms of risk-adjusted NSQIP data, our O/E for

VTE was 3.41 before implementation of the program(n ¼ 1,569, 95% confidence interval [CI] 2.40 to4.70), and steadily decreased to an OR of 0.94 (n ¼1,323, 95% CI 0.56 to 1.58, p < 0.05) after risk-stratified prophylaxis was introduced (Fig. 4).

DISCUSSIONMore than 20 years since Joseph Caprini introduced aVTE risk score, and despite guidelines from multiple

Figure 2. Incidence (%) of deep venous thromboses at Boston Medical Center (blue line) and atcomparable NSQIP institutions (red line) during consecutive reporting periods before and afterthe implementation of a standardized mobilization program and a risk-stratified prophylaxisprogram. *indicates p < 0.01 compared with calendar year (CY) 2009.


sources for VTE prophylaxis regimens, PE and DVTremain significant problems among hospitalized patientsin the United States. Data show that high-risk patientstend to receive insufficient prophylaxis; low-risk patientsmay be overtreated.15 It has been suggested that the solu-tion to this problem is standardized risk assessment andcommensurate prophylaxis.16

We were distressed to discover that our hospital was ahigh outlier for NSQIP-defined VTE events, but weregarded this as an opportunity to improve care. Byadopting mandatory VTE risk calculation and associatedrisk-stratified prophylaxis for every patient, along withstandardized early postoperative mobilization, we demon-strated excellent adherence to prophylaxis guidelinesand a dramatic reduction in postoperative VTE eventsamong our patients. Although the Caprini scoring sys-tem has been well validated in terms of its predictivevalue for VTE,17-19 to our knowledge this is the first studyto demonstrate a reduction of VTE events based on itsstandardized and required use in conjunction with aformal mobilization program.We have demonstrated that the incidence of VTE events

was reduced after the introduction of a postoperative

mobilization program and a mandatory risk stratificationsystem with corresponding risk-based prophylaxis guide-lines, including recommendations for extended chemopro-phylaxis. Our data (Figs. 2 to 4) indicate that the incidenceof VTE events began to decline before the official introduc-tions of the programs, although they still remained greaterthan expected. Our faculty had deliberated and plannedthese new programs for quite a while, likely influencingchanges in practice before formal implementation of theprograms. Once the programs were operational, weobserved a continued reduction in adverse VTE events.The risk of VTE conferred by some operations persists

after discharge, and up to one-third of VTE events aftercancer operations may be diagnosed after the index admis-sion.11 Several randomized controlled trials have demon-strated excellent risk reduction by extending the courseof VTE prophylaxis on an outpatient basis.10,20,21 Despitethis evidence, adherence to extended prophylaxis regimensin actual practice has been poor.22,23 In the experience atour medical center, extended regimens of prophylaxiswere rarely used before introduction of the mandatoryrisk stratification and electronic reminder system. The re-minders have particularly increased awareness among our

Figure 3. Incidence (%) of pulmonary embolism at Boston Medical Center (blue line) and atcomparable NSQIP institutions (red line) during consecutive reporting periods before and afterthe implementation of a standardized mobilization program and a risk-stratified prophylaxisprogram. *indicates p < 0.01 compared with calendar year (CY) 2009.


surgeons about the recommendation for extended prophy-laxis in high-risk patients. Indeed, audits show that major-ity of high-risk patients at our institution now routinelyreceive extended duration prophylaxis. This may accountfor much of the dramatic reduction that we observed inthe overall incidence of DVT and PE at 30 days. Our de-cision to construct the risk-stratified regimens withextended courses of prophylaxis was based on the collectedliterature on Caprini scores and the commensurate risk forVTE at each level. In particular, a report by Bahl and col-leagues17 demonstrated the greatest likelihood of VTE inpatients with scores greater than 5. Furthermore, therewas a significant increase in the likelihood of VTE betweenscores of 7 and 8 (2.58% incidence) and scores of 9 orhigher (6.51% incidence). Therefore, we chose to recom-mend extended duration chemoprophylaxis of 7 to 10days for patients with scores 5 to 8, and a more extended30-day duration for those with scores of 9 or greater.The risk of VTE in patients with scores less than 5 is quitelow (less than 1%) so that the cost and risk of extendedprophylaxis does not seem to be justified in those groups.The concept of electronic reminders has been shown to

increase use of VTE prophylaxis and to decrease the inci-dence of VTE events. In a study by Kucher and associates,24

patients randomized to an intervention group in whichphysicians were shown an automatically generated elec-tronic reminder of VTE prophylaxis recommendationswere nearly twice as likely to receive pharmacologic pro-phylaxis, when compared with the group for which no elec-tronic reminders were generated. Importantly, theincidence of DVT and PEwas reduced by 41% in the inter-vention group. We used this strategy to increase adherenceto VTE prophylaxis among our surgeons and demonstratedsimilar success. This lends support to the strategy of elec-tronic reminders that are tailored to individual risk factors.Early frequent mobilization has been suggested as a strat-

egy to prevent VTE, based largely on the fact that immobi-lized patients are at particular risk for VTE.9,25 However,there are no randomized controlled trials toprove the benefitof postoperative mobilization. Despite the lack of evidence,VTE prevention guidelines emphasize early mobilization asa key component of prophylaxis and as the only necessaryprophylactic measure in lowest risk patients. We embracedthe idea of standardized, early, frequent postoperativemobi-lization as a mechanism of VTE prevention, and as anelement of excellent patient care, to decrease the likelihoodof postoperative respiratory complications.26 We are notable to determine the extent to which mobilization may

Figure 4. Risk-adjusted ratio for total venous thromboembolism at Boston Medical Center,during consecutive NSQIP reporting periods before and after the implementation of a stan-dardized mobilization program and a risk-stratified prophylaxis program. For reporting periodsbefore calendar year (CY) 2010, NSQIP reports risk-adjusted data as observed/expected (O/E).For CY 2010 and later, NSQIP reports risk-adjusted data as odds ratios (OR). Vertical linesindicate 95% confidence intervals.


have reduced our incidence ofVTE relative to the benefits ofmandatory risk calculation and risk-based prophylaxis.Nevertheless, our incidence of VTE had decreased withinstitution of the mobilization program alone while therisk-stratified VTE program was being designed. Ulti-mately, our aimwas not to demonstrate the superior efficacyof a solitary intervention, but rather to support the conceptthat standardization of practices can improve patient out-comes. In reality, the 2 practicesdmobilization and risk-based prophylaxisdmay be synergistic.We perceived little resistance to the program among

our surgeons, which is reflected in the high compliancerates with prophylaxis regimens. We found that surgeonsappreciated the simplicity of the program, including theelectronic automation, as an easy and thoughtful way toindividualize VTE prophylaxis based on risk. Althoughthe risk-stratification element was required, the specificprophylaxis regimens were not mandated, and an “optout” choice was provided, along with documented rea-sons for declining the therapy. Therefore, surgeons main-tain autonomy and gain awareness of VTE hazards byvirtue of the risk-assessment and electronic reminders.

There are several methodologic limitations of this study.We present a before-after analysis, comparing our NSQIP-reported VTE outcomes before implementation of ourstandardized prevention protocol to subsequent outcomes.Therefore, we have not conducted a randomized trial ofthe described interventions. Because the basis of ourwork was a quality improvement effort founded on stan-dardization, we believed it necessary to implement the pro-gram for all of our general and vascular surgery patients atthe same time. Furthermore, the usage of NSQIP dataallowed us to present risk-adjusted data for our entire studypopulation during the periods before and after interven-tion. Admittedly, NSQIP has some inherent limitations,such as capturing only a representative sampling of pa-tients rather than including all patients. Nevertheless, itsstatistical methods are well validated and accepted by sur-gery audiences. The risk-adjusted outcomes account forpatient comorbidities, limiting the impact of possible vari-ability in patient disease between the study periods, andshow a definite improvement in VTE events.Because of changes in the way NSQIP reports risk-

adjusted outcomes, data are presented as O/E ratios for


periods before calendar year 2010 and as odds ratios for cal-endar year 2010 and later. As a result, we compare O/E ra-tios for the periods before our VTE prevention program toORs for the later intervals. However, O/E and OR areconsidered to be statistically comparable for large samplesizes. We have received formal NSQIP assurance that theinterchangeability of these data for large groups offers avalid comparison. The dramatic reduction of absolute andrisk-adjusted VTE events provides convincing evidencethat an actual benefit was achieved among our patients.Health systems that have higher use of imaging studies

have a higher likelihood of diagnosing VTE complica-tions.27 Such a surveillance bias could theoretically influencethe results of our study if use of VTE diagnostic imagingchanged during the study period. Although we did not tracktesting patterns, we have no reason to believe that our insti-tutional imaging practices changed substantively.

CONCLUSIONSIn conclusion, our patient care program emphasizing earlypostoperative mobilization, mandatory VTE risk stratifica-tion, and commensurate electronic prophylaxis recom-mendations, significantly reduced the likelihood of VTEcomplications among our patients. Risk-based prophylaxisfor VTE provides a distinct benefit to patients, and stan-dardization of care ensures that best-practice guidelines arefollowed to the greatest extent possible. We are encour-aged by the success in reducing these devastating eventsamong our patients by implementing such a strategy,and we are optimistic that postoperative complicationsmay be diminished by adherence to risk-stratified andstandardized patient care standards.

Author Contributions

Study conception and design: Cassidy, Rosenkranz,McAneny

Acquisition of data: Cassidy, Rosenkranz, McAnenyAnalysis and interpretation of data: Cassidy, Rosenkranz,McAneny

Drafting of manuscript: Cassidy, McAnenyCritical revision: Cassidy, McAneny

Acknowledgment: The authors acknowledge the gracioussupport and guidance of Joseph Caprini, MD, in developingour VTE prevention program.

REFERENCES

1. Silverstein MD, Heit JA, Mohr DN, et al. Trends in the inci-dence of deep vein thrombosis and pulmonary embolism. ArchIntern Med 1998;158:585.

2. Spyropoulos AC, Hussein M, Lin J, Battleman D. Rates ofsymptomatic venous thromboembolism in US surgical patients:

a retrospective administrative database study. J ThrombThrombolysis 2009;28:458e464.

3. Heit JA. Venous thromboembolism: disease burden, outcomesand risk factors. J Thromb Haemost 2005;3:1611e1617.

4. Wakefield TW, Caprini J, Comerota AJ. Thromboembolicdiseases. Curr Probl Surg 2008;45:844e899.

5. Dimick JB, Chen SL, Taheri PA, et al. Hospital costs associ-ated with surgical complications: a report from the private-sector National Surgical Quality Improvement Program.J Am Coll Surg 2004;199:531e537.

6. Motykie GD, Zebala LP, Caprini JA, et al. A guide to venousthromboembolism risk factor assessment. J Thromb Throm-bolysis 2000;9:253e262.

7. Arcelus JI, Candocia S, Traverso CI, et al. Venous thrombo-embolism prophylaxis and risk assessment in medical patients.Semin Thromb Hemost 1991;17[Suppl 3]:313e318.

8. Caprini JA, Arcelus JI, Hasty JH, et al. Clinical assessment ofvenous thromboembolic risk in surgical patients. SeminThromb Hemost 1991;17[Suppl 3]:304e312.

9. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention ofvenous thromboembolism: American College of Chest Physi-cians Evidence-Based Clinical Practice Guidelines (8th Edi-tion). Chest 2008;133[6 Suppl]:381Se453S.

10. Bottaro FJ, Elizondo MC, Doti C, et al. Efficacy of extendedthrombo-prophylaxis in major abdominal surgery: what doesthe evidence show? A meta-analysis. Thromb Haemost 2008;99:1104e1111.

11. Merkow RP, Bilimoria KY,McCarter MD, et al. Post-dischargevenous thromboembolism after cancer surgery: extending thecase for extended prophylaxis. Ann Surg 2011;254:131e137.

12. Amin A, Stemkowski S, Lin J, Yang G. Thromboprophylaxisrates in US medical centers: success or failure? J Thromb Hae-most 2007;5:1610e1616.

13. Amin AN, Stemkowski S, Lin J, YangG. Inpatient thrombopro-phylaxis use in U.S. hospitals: adherence to the seventh Amer-ican College of Chest Physicians’ recommendations for at-riskmedical and surgical patients. J Hosp Med 2009;4:E15eE21.

14. Cohen AT, Tapson VF, Bergmann JF, et al. Venous thrombo-embolism risk and prophylaxis in the acute hospital caresetting (ENDORSE study): a multinational cross-sectionalstudy. Lancet 2008;371:387e394.

15. Deheinzelin D, Braga AL, Martins LC, et al. Incorrect use ofthromboprophylaxis for venous thromboembolism in medicaland surgical patients: results of a multicentric, observationaland cross-sectional study in Brazil. J Thromb Haemost2006;4:1266e1270.

16. Caprini JA. Risk assessment as a guide for the prevention ofthe many faces of venous thromboembolism. Am J Surg2010;199[1 Suppl]:S3e10.

17. Bahl V, Hu HM, Henke PK, et al. A validation study of aretrospective venous thromboembolism risk scoring method.Ann Surg 2010;251:344e350.

18. Pannucci CJ, Bailey SH, Dreszer G, et al. Validation of theCaprini risk assessment model in plastic and reconstructivesurgery patients. J Am Coll Surg 2011;212:105e112.

19. Zhou HX, Peng LQ, Yan Y, et al. Validation of the Cap-rini risk assessment model in Chinese hospitalized patientswith venous thromboembolism. Thromb Res 2012;130:735e740.

20. Bergqvist D, Agnelli G, Cohen AT, et al. Duration of prophy-laxis against venous thromboembolism with enoxaparin aftersurgery for cancer. N Engl J Med 2002;346:975e980.

http://refhub.elsevier.com/S1072-7515(14)00186-0/sref1






















































































1104 Dankers and Ashley Invited Commentary J Am Coll Surg

21. Rasmussen MS, Jorgensen LN, Wille-Jorgensen P, et al. Pro-longed prophylaxis with dalteparin to prevent late thromboem-bolic complications in patients undergoing major abdominalsurgery: a multicenter randomized open-label study. J ThrombHaemost 2006;4:2384e2390.

22. Yu HT, Dylan ML, Lin J, Dubois RW. Hospitals’ compliancewith prophylaxis guidelines for venous thromboembolism. AmJ Heal Syst Pharm 2007;64:69e76.

23. Muntz J. Duration of deep vein thrombosis prophylaxis in thesurgical patient and its relation to quality issues. Am J Surg2010;200:413e421.

24. Kucher N, Koo S, Quiroz R, et al. Electronic alerts to preventvenous thromboembolism among hospitalized patients.N Engl J Med 2005;352:969e977.

25. Agnelli G. Prevention of venous thromboembolism in surgicalpatients. Circulation 2004;110[24 Suppl 1]:IV4e12.

26. Cassidy MR, Rosenkranz P, McCabe K, et al. I COUGH:Reducing postoperative pulmonary complications with a multi-disciplinary patient care program. JAMA Surg 2013:1e6.

27. Bilimoria KY, Chung J, Ju MH, et al. Evaluation of surveil-lance bias and the validity of the venous thromboembolismquality measure. JAMA 2013;310:1482e1489.

Invited Commentary

Christian A Dankers, MD, MBA,Stanley W Ashley, MD, FACS

Boston, MA

Cassidy and colleagues should be commended for theirwork to develop a standardized protocol to reduce post-operative venous thromboembolism (VTE). Althoughsurveillance bias complicates the interpretation and com-parison of VTE rates between hospitals (increased hospitalVTE event rates are associated with increasing hospital im-aging use rates),1 few organizations would likely argue thatthey can’t improve in the implementation of postoperativeVTE reduction strategies. It is heartening to see that an or-ganization can experience such an impressive reduction inVTE rates in a relatively short period of time.Cassidy and colleagues present us with a program that

includes mandatory risk stratification (in the form of anelectronic system that provides guidance on prophylaxis se-lection and provides reminders at key times duringhospitalization) and early mobilization. In addition, theelectronic guidance included recommendations for postdi-scharge prophylaxis for certain groups of high risk patients.As we seek to reduce VTE events in our own organiza-

tions, our challenge is to determine which particularinterventions have the most impact and therefore, are

worthy of time and resource investment. Fully imple-menting the VTE reduction program proposed by Cas-sidy and associates would require a fair institutionalcommitment, both in information technology and pro-cess improvement resources. It would be helpful toknow which particular aspects of the program weremost impactful. Could the reduction in VTE eventshave been driven exclusively by the electronic risk strati-fication module? Was much benefit attributable to mobi-lization? Or, was providing postdischarge prophylaxis forhigh risk patients really the key intervention? The authorsstated that it was not their intent to demonstrate the su-periority of a particular intervention, but rather to sup-port the concept that standardization of practices canimprove patient outcomes. This is understandable, butit makes it more difficult for us to make judgments aboutfocused interventions. In addition, it is not clear howchemoprophylaxis practices actually changed as a resultof the intervention. It would be helpful to have dataabout rates and other aspects of VTE chemoprophylaxispractices pre-, peri- and postintervention, especiallybecause most of the reduction in the VTE risk-adjustedratio actually came before any of the interventions wereformally launched. The authors suspected that this isbecause surgeons were already starting to implementcertain aspects of the program before the formal launch,but we don’t have the data on practices to validate thishypothesis.Certainly, few would argue that we shouldn’t work to

improve decision-making about prophylaxis, improvethe consistency of prophylaxis administration when or-dered, and improve postoperative mobilization in asafe manner. With time and further study, it shouldbecome clearer whether these are synergistic practices,as the authors suggest, or whether certain high-impactinterventions drive the bulk of improvement. Cassidyand colleagues have done great work to develop andimplement a VTE reduction strategy, and they haveseen a remarkable improvement in outcomes. Weshould applaud their success as we continue the workto refine our understanding of the impact of VTE reduc-tion strategies.

REFERENCE

1. Bilimoria KY, Chung J, Ju MH, et al. Evaluation of surveillancebias and the validity of the venous thromboembolism qualitymeasure. JAMA 2013;310:1482e1489.


































reducing postoperative venous thromboembolism ... · vte risk stratiﬁcation and commensurate...

Documents