Accuracy of venous thromboembolism assessment and compliance to prophylaxis in a tertiary care center

Introduction

Venous thromboembolism (VTE), an entity inclusive of deep vein thrombosis (DVT) and pulmonary embolism (PE), is a leading cause of morbidity and mortality in hospitalized patients. VTE has been found to account for 5–10% of hospital deaths, and to be a contributing factor in another 15% of hospital deaths, making it the most common preventable cause of in-hospital mortality.^1–5 It is also associated with long-term risks of post-thrombotic syndrome and chronic thromboembolic pulmonary hypertension, which not only increase morbidity, but hospital length of stay and cost of treatment.^6,7

Risk factors for VTE have been clearly defined, and multiple studies have shown that using prophylaxis in hospitalized patients at risk for VTE is safe, effective, and cost effective.^{1,2,4–6,8–10} The American College of Chest Physicians (ACCP) and National Institute for Health and Care Excellence (NICE) in the UK have set standards for the prevention and treatment of thromboembolic disorders. The ACCP guidelines are updated every three years and are the most frequently cited standard for the prevention and treatment of VTE.^{3,4,6–8,11,12} Despite this, the Epidemiologic International Day for the Evaluation of Patients at Risk for Venous Thromboembolism in the Acute Hospital Care Setting (ENDORSE) study found that worldwide, appropriate VTE prophylaxis was not being administered.^4,6,7,12 Subsequently, numerous strategies have been employed to improve VTE assessment and prophylaxis practices, including the use of electronic alerts, education, and computerized clinical decision support (CDS) tools which utilize an electronic medical record (EMR).^{1,2,5,9,10,13–18}

Most institutions have implemented interventions that integrate into the patient care process to improve VTE assessment and ordering of prophylaxis. However, few studies have looked at the effectiveness of these systems to correctly assess patient risk level as well as compliance to ordering recommended prophylaxis. More specifically, the appropriateness of the initial assessment and the proper ordering of prophylaxis remain factors which may continue to hinder prevention of VTE. As such, the purpose of our study was to determine the appropriateness of the admitting service’s (AS’s) assessment. The assessment was performed using an electronic provider advisor program which uses CDS tools. Compliance of the AS with ordering of the recommended prophylaxis was also evaluated.

Methods

An examination of randomly selected patients assessed within 18 h of admission for VTE risk was performed at a tertiary care center over a five-month period. This was a retrospective review of prospectively collected data of a quality improvement project to assess for appropriateness of VTE assessments. As such, institutional review board approval was waived and patient consent was not required. Assessment of VTE risk was determined by utilization of a Cerner VTE advisor (an interactive application through the EMR) of our institution which is based on the validated Caprini model. VTE risk category was developed using a Caprini score which utilized multiple factors including, but not limited to, age, body mass index, patient type, previous medical history, diagnosis, documented problems, allergies, home medications, level of care, and laboratory results and activity level.^8,11 The VTE advisor utilizes information from the EMR and that which is entered by the provider to calculate individualized risk and prophylaxis recommendations. Essentially, this is a risk stratification and prophylaxis recommendation tool for all admitted patients based on the 2008 ACCP guidelines. ¹¹ Patients were first categorized as either medical or surgical. The medical patients were stratified as either as low risk or high risk. The surgical patients were stratified as low, moderate/high, or very high risk. The ordering provider is required to perform the assessment by recording all the risk factors, and the advisor provides the patient’s risk level and the appropriate prophylaxis recommendations based on ACCP guidelines. The assessment must be completed within 18 h of admission or order entry will be blocked.

Adult patients ≥ 18 years of age admitted to both surgical and medical units in varying acute levels of care were arbitrarily selected and assessed within 18 h of admission over a consecutive five-month period within the middle of the academic year. It was not felt the time period had any bearing on the accuracy of assessments. Patients were selected and assessed on different weekdays and different times during day-time shifts to obtain unbiased data. Patient selection was made from a list automatically generated by our EMR that only included patients admitted within the past 18 h. All admissions were emergent and came through our institution’s emergency department. Pediatric patients < 18 years of age, and those admitted to the psychiatric, obstetrics, and gynecology services were excluded. Patients with VTE on admission as determined by a VTE admission diagnosis or seen on imaging (duplex ultrasound, computed tomography angiography, ventilation/perfusion scan, cardiac magnetic resonance imaging, or cardiac catheterization) at the time of risk assessment were also excluded.

Patients were assessed for VTE risk independently by the AS and by the VTE consult service (CS) both within 18 h of admission. The AS providers were mostly junior resident physicians and fellows. The CS is a team of Advanced Practice Nurses with specialized education in VTE that has exclusively managed the hospital’s VTE team for the past five years. This team includes a physician leader and a phlebology fellow that assist with policy and procedure development regarding VTE prophylaxis, as well as monitors DVTs throughout the hospital. The CS assessment was done in a blinded fashion to the AS, based on information provided within the AS’s admission note. VTE prophylaxis orders were then reviewed based on AS assessment and compliance to the VTE advisor recommendations. Only the type of prophylaxis was considered (pharmacologic, mechanical, both, or none).

The comparison between the AS and CS in the different VTE risk groups was done with a two tailed Fisher’s exact test. The level of significance was set at a p value of < .05.

Results

A total of 104 patients were evaluated, four of whom were excluded because of VTE presence on admission. Mean age was 63 years. There were 50 women and 50 men; 91% were white, 4% African American, 2% Hispanic, and 3% unknown. There were 52 medical patients and 48 surgical patients.

Based on AS assessment in the medical group, 7.7% were low risk and 92.3% were high risk. In the surgical group, 6.3% were low risk, 52% were moderate/high risk, and 41.7% were very high risk.

Comparing AS to CS assessment there were patients who were both under-assessed and over-assessed. Incorrect assessment, as determined by when the AS did not correspond to the CS assessment, occurred in 13 patients (p < .001). There were six patients who were under-assessed (p = .029). Two of these patients were medical and four were surgical. There were seven patients who were over-assessed (p = .014). One patient was medical, and six were surgical.

Only looking at AS assessment, regardless of the accuracy of this assessment, there were eight patients who did not have the prophylaxis recommended by the VTE advisor ordered. Two patients were under-prophylaxed (both were surgical) and six patients received over-prophylaxis (two medical and four surgical). Unnecessary exposure to complications due to inappropriate prophylaxis occurred in five patients (three bleeding risk and two VTE risk) (Table 1).

OPEN IN VIEWER

Table 1.

Potential complications from inappropriately ordered VTE prophylaxis by the admitting service.

VTE risk assessment	Prophylaxis administered	Potential harm
Medical low	Pharmacologic	Bleeding
Medical low	Mechanical	None
Surgical low	Mechanical	None
Surgical low	Mechanical + pharmacologic	Bleeding
Surgical low	Mechanical + pharmacologic	Bleeding
Surgical moderate/high	Mechanical	VTE
Surgical moderate/high	Mechanical + pharmacologic	None
Surgical very high	Pharmacologic	VTE

Discussion

Numerous strategies have been used including education programs, consultation services, electronic alerts, and computerized CDS tools to improve VTE assessment and prophylaxis. Educational efforts have had the lowest impact on change in VTE prevention while computerized CDS systems have had a more consistent impact. Most improvement occurs with the use of multifaceted interventions.^{1,2,5,9,10,13–18}

In 2000, Durieux et al. ¹⁶ used instant messaging to clinicians through a computerized CDS system which improved VTE prophylaxis compliance from 82.8% to 94.9%. Kucher et al. ¹⁵ developed an electronic alert system to physicians that improved significantly the use of VTE prophylaxis and resulted in fewer episodes of symptomatic VTE (4.9% vs. 8.2%).

Mandatory CDS tools were the natural next step to further improve compliance in patient risk assessment and prophylaxis. Implementation of a trauma-specific VTE CDS tool demonstrated an improvement in compliance with guideline-appropriate prophylaxis (66.2–84.4%; p < .001) and a reduction in the rate of preventable harm from VTE (1.0–0.17%; p = .04). ¹⁷ Zeidan et al. ² compared data from pre- and post-implementation of a mandatory computerized CDS “smart order set” on the medical service. They found the prescription of risk-appropriate VTE prophylaxis increased from 65.6% to 90.1% (p < .0001) and radiographically documented symptomatic VTE within 90 days of hospital discharge declined from 2.5% to 0.7% (p = .002). In our institution, a mandatory computerized CDS tool was introduced five years ago which has resulted in a significant increase in compliance with VTE risk assessment (50% pre-implementation vs. 98% post-implementation), and an increase in ordering of appropriate prophylaxis (72% pre-implementation vs. 92% post-implementation).

Failure to significantly reduce our hospital acquired VTE events and a significant persistent rate of preventable harm from VTE led to this study in order to identify areas for further improvement. We found that risk stratification is not always done correctly with a tendency to both under-assess and over-assess patients. This is the first study to our knowledge that compared AS assessment with an independent VTE CS. It showed that CDS tools have limitations in appropriate patient assessment which can result in inadequate prophylaxis. In addition, we found that mandatory completion of a risk assessment tool does not translate to compliance with recommended prophylaxis. As a result, strategies to increase the implementation of appropriate prophylaxis need to be re-examined.

Although a CDS tool is used, it is primarily limited by the provider who uses it. Education and training on the VTE advisor and the ACCP guidelines happened throughout the year. However, the VTE advisor form may still not be filled out correctly, resulting in an improper assessment. This can result in under- and over-assessment in 13% of patients, both of which can lead to patient harm. We believe this occurred because providers are not utilizing the advisor tool effectively, and entering incorrect data. Enhanced training, more frequent education sessions, and accountability may help this in the future. Another important issue with the computerized system is the prompt for prophylaxis therapy. Although the advisor itself gives the provider the correct prophylaxis recommendation based on the provider’s assessment, the ordering provider is free to order prophylaxis or omit it, as the program provides solely a recommendation. As such, even if the assessment is filled out correctly, the provider is not forced to follow the appropriate recommendations. Failure to order recommended prophylaxis occurred in 8% of the patients in our study resulting in unnecessary exposure to complications in 5% of the patients.

This weakness of CDS tools can be addressed with mandatory order entry of recommended prophylaxis. Mandatory ordering has been looked at by several institutions but often is confronted with significant resistance by medical staff.^2,15–18 This is because it may lead to patient harm, especially since assessment of bleeding risk may not be done appropriately. Many physicians may feel a patient is at higher risk of bleeding when they are not. Interestingly however, in our study, there were more patients who were over-prophylaxed. Another option would be automated surveillance programs evaluating patient’s risk level and ordered prophylaxis. Alerts are sent to physicians and nursing staff on patients who are not receiving appropriate prophylaxis.

This study was limited by the sample size. Although the 100 randomly selected patients gave us a good insight for the VTE prophylaxis patterns, it may not be adequate to explain the exact practice in the institution. Data collection occurred at a single time interval (within 18 h of admission or earlier) and may have missed patients who eventually received appropriate prophylaxis within 24 h of admission. Additionally, we only looked at patients who were admitted through the emergency department. The assessments of those patients admitted electively were not evaluated. The study site, which is a major limitation, was a university teaching hospital where most orders are placed by house staff physicians. The challenge of any institution that trains physicians is to provide appropriate patient care while integrating trainees into the program. As such, new residents may not be trained on how to properly complete the VTE assessment and may incorrectly order any form of prophylaxis. At our institution, the house staff are the providers who primarily complete the VTE assessment and orders. This may be further reflected in some of the AS assessments which occurred at night, when there is less staff to facilitate accurate completion of the assessment and ordering of prophylaxis. In contrast, all of the CS assessments were performed during day-time shifts with a full team in attendance. Therefore, it is uncertain whether our findings would apply to institutions where attending physicians place the orders, and would not be applicable to institutions without computerized physician order entry systems.

Conclusion

This study reviewed the appropriateness of an AS assessment of VTE risk using an EMR CDS tool as well as compliance of ordering recommended prophylaxis. Even with utilization of the computerized system, we found a significant number of patients are not being appropriately and accurately assessed. Furthermore, even when providers perform an assessment of VTE risk and are provided with a recommendation for prophylaxis, many patients are not receiving the correct prophylaxis they need.