Venous Thromboembolism: Are Regulatory Requirements Reasonable?

Elliott R. Haut, MD*
Assistant Professor of Surgery,
Anesthesiology and Critical Care Medicine
Division of Trauma, Surgery and Critical Care
Department of Surgery
The Johns Hopkins Hospital
Baltimore, Maryland, USA

Is venous thromboembolism a significant problem in critically ill patients?
Venous thromboembolism (VTE), the combined term for deep vein thrombosis (DVT) and pulmonary embolism (PE), is one of the most common complications for patients in the intensive care unit (ICU). VTE causes approximately 200,000 deaths per year in the United States (more than AIDS and breast cancer combined).1 The most notable events result in cardiopulmonary arrest and death in patients where a DVT dislodges, becoming a saddle PE. However, DVT alone also can cause significant delayed death in the form of post-thrombotic syndrome. Patient demographics and other medical diagnoses common in ICU patients (e.g., age, cancer, previous thrombosis, heart failure, pregnancy, immobility and stroke) clearly are associated with development of VTE. In addition, the risk of VTE increases as more patients undergo major surgery and admittance into the ICU for advanced treatments, such as mechanical ventilation and central venous access.

The absolute prevalence of DVT in patients not receiving any prophylaxis is at least 10%, even in very low risk groups, but this rate can be as high as 80% in certain patient populations.2 In one study of ICU patients who were mechanically ventilated for seven days and screened with duplex ultrasound, nearly 25% of patients were diagnosed with DVT, even though all were receiving prophylaxis. Furthermore, 11.5% of patients with DVT also were diagnosed with PE.3 In the trauma population, the DVT rate is as high as 58%4, depending on how aggressively clinicians search for these often clinically silent DVTs.

Some readers may think these numbers seem high and that rates among their patients are much lower. This is a common misconception. In published, well-controlled studies, researchers aggressively search for DVTs in all patients and report these high rates. In routine clinical practice (where clinicians do not search for DVT in every patient), rates often seem much lower.

Do prevention strategies for VTE really work?
Prevention of VTE works remarkably well. The Agency for Healthcare Research and Quality (AHRQ) has identified 11 “Clear Opportunities for Safety Improvement,” based on the strength of the evidence supporting more widespread implementation. The top-ranked item on this list is “appropriate use of prophylaxis to prevent venous thromboembolism in patients at risk.”5 Many large, dependable clinical trials have proven the benefit of primary thromboprophylaxis for prevention of VTE. Appropriately dosed chemical prophylaxis (e.g., unfractionated or low molecular-weight subcutaneous heparin) reduces DVT, PE, and (most importantly) fatal PE. In addition to improving outcomes, prophylactic measures are relatively inexpensive and cost-effective. The American College of Chest Physicians (ACCP) has published the standard evidencebased guidelines for prevention and treatment of VTE. This comprehensive document summarizes the available evidence and is an invaluable resource.2

How good are we at providing the correct prevention measures?
In general, healthcare providers do poorly across the board when it comes to giving appropriate VTE prophylaxis. Samuel Z. Goldhaber of the DVT FREE Steering committee has suggested that, “the disconnect between evidence and execution as it relates to DVT prevention amounts to a public health crisis.”6 Huge numbers of patients at risk never receive the correct preventative measures. In a large prospective registry study, only 42% of patients with in-hospital diagnosis of DVT received prophylaxis before diagnosis. 7 A recent study of almost 70,000 patients in 358 hospitals in 32 countries, published in The Lancet, shows a similarly bleak picture. On the basis of the ACCP guidelines, 52% of these patients were determined to be at risk for VTE, but only 59% of surgical patients and 40% of medical patients received the recommended prophylaxis regimen.8

If this is so important, why don’t we do it?
One reason patients are not receiving proper prophylaxis may simply be providers’ lack of knowledge regarding its importance. Individuals often underestimate the risk of the disease, while overestimating the potential for bleeding associated with prophylaxis. Healthcare practitioners erroneously may think their patients are different from the thousands who have been studied closely to derive this excellent data. Even when the appropriate prophylactic regimens are prescribed, they often are not implemented correctly. Low-dose prophylactic heparin doses are held frequently (and incorrectly) for invasive procedures and may be missed while patients are away from the patient care floor. Sequential compression devices (SCDs) have a role in prophylaxis, but one study showed that they are used correctly only about 50% of the time even when ordered.9

Why is DVT suggested as a marker of quality of care?
Measuring quality in healthcare is at the foreground of medicine in the United States. Many national providers are implementing pay-for-performance programs. DVT has been considered a quality metric, mainly because many in-hospital DVT/PE events and subsequent sequelae are considered preventable with appropriate prophylaxis. DVT prophylaxis is aimed at preventing further complications, including PE, the most serious complication, as well as at permanent swelling and venous insufficiency. Unfortunately, the implementation of pay for performance and public reporting of supposed quality of care are moving more rapidly than the science of measuring quality. Identifying measures of quality based on evidence-based medicine remains difficult and has resulted in significant controversy. Pronovost et al have suggested that the true analytical science behind measures of quality may be lagging behind the reality of the public’s quest for knowledge and their desire to go to the “best” hospitals.10 VTE can be approached as a quality measure in all three of the classic Donabedian methods of quality assessment: structure, process and outcome. However, these methods are not as straightforward as they seem at first glance. What pieces of structure (i.e., physical equipment and facilities) impact VTE? How should we evaluate riskstratification schemes and prophylaxis ordering routines that have been major agenda items for the process of care? Even simply measuring an outcome, such as VTE rates, is more complicated than it appears, as the denominator (patients at risk) is often not abundantly clear. 11

How are we currently being measured? What is on the horizon for benchmarking DVT?
Many national organizations and accrediting bodies are working on the topic of VTE and quality of care. The AHRQ is playing a major role in determining quality measures (including those for VTE) and has listed postoperative DVT as one of its Patient Safety Indicators (PSI). The National Quality Forum (NQF) also has listed VTE as a potential indicator of quality. The NQF continues its long, ongoing process to determine how it will benchmark hospitals and individual practitioners regarding DVT and PE. It is expected to publish policy, preferred practices, and initial performance measures soon, which will focus on four main domains: risk assessment/stratification, prophylaxis, diagnosis, treatment and monitoring.12 The final recommendations for suggested and endorsed performance measures are expected to be announced in April 2008.13 The Joint Commission recently mandated risk stratification for VTE for all patients upon admission to the hospital and upon admission to an ICU setting.

What can my hospital do to ensure appropriate VTE risk assessment on all patients?
Many hospitals already have VTE risk-stratification tools in place, but their use is sporadic at best. Unless this process is quick and efficient, providers will not be able to comply. A large systematic review of 30 studies looking at strategies that improved VTE prophylaxis came to some not-so-surprising conclusions: passive dissemination of guidelines simply does not work; a single, isolated strategy is less effective than multiple strategies used in combination; and the most effective systems use an active system to remind clinicians to assess for VTE risk through either paper or computerized decision-support mechanisms.14 At Johns Hopkins Hospital, the VTE collaborative recently implemented a novel solution to this complex problem. Old risk-assessment paper forms were converted into electronic formats and incorporated into the new computerized provider order entry (CPOE) system. Providers simply click on the appropriate risk factors and/or contraindications; the computer pulls in other information (e.g., age, renal function), and the appropriate level of risk and recommended prophylaxis appear. To ensure compliance, the VTE collaborative constructed order sets in such a way that risk assessment must be completed or the CPOE system will not accept any other orders. The patients get the Right Care, Right NowTM, and the providers are educated about the risk-stratification process. Johns Hopkins Hospital believes that implementing this mandatory step will ensure that patients are stratified by risk and that staff order risk appropriate VTE prophylaxis.  This group is also examining this project from a performance improvement standpoint and is planning a future report, which will compare the number of patients with documented VTE risk stratification and risk-appropriate VTE prophylaxis ordering before and after the CPOE implementation.

What about surveillance bias in DVT reporting? The trauma example.
Due to conflicting study results, ongoing debate continues in the literature regarding the utility of screening high-risk asymptomatic trauma patients for occult DVT. Some studies suggest that routine screening of asymptomatic patients in high-risk populations may benefit patients if DVTs are identified early and are treated appropriately to avoid fatal PE and the long-term sequelae.15-19 However, some authors claim that such screening is not cost effective20 and that strict adherence to DVT prophylaxis guidelines makes more of a difference than screening duplex ultrasound imaging.21

As a result, many practice variations exist in the use of screening duplex ultrasound for DVT in high-risk asymp tomatic trauma patients. Surveillance bias is a type of selection or information bias that occurs when an exposure (e.g., duplex ultrasound) may result in a higher probability of detection (e.g., DVT) in exposed patients. Surveillance bias causes any data acquired and reported to be called into question. Johns Hopkins Hospital researchers recently published two studies showing the possible effects of surveillance bias in the reporting of DVT after trauma. They compared patient outcomes before and after implementing practice management guidelines promoting duplex ultrasound surveillance of high-risk asymptomatic trauma patients admitted to their level 1 trauma center. Trauma patients were four times more likely to have a duplex ultrasound performed and 10 times more likely to be diagnosed with DVT after guidelines were implemented22 (See Figure 1). The researchers also have shown the possible effect of surveillance bias on a national level using the National Trauma Data Bank, the largest combined trauma registry in the world. When hospitals are separated into quartiles by duplex rate, the DVT rate in the highest quartile was seven-fold higher than the combined DVT rate in the first three quartiles.23 In the current system, hospitals may screen patients aggressively, find more DVTs, and be unfairly labeled as providing worse care. DVT rates may be related to how often caregivers look for these events rather than the quality of care provided. The use of DVT rate alone as an independent quality measure for hospital benchmarking and for pay-for-performance programs may be questionable.

Is “zero tolerance” a realistic goal for VTE?
The Centers for Medicare and Medicaid Services (CMS) recently announced that hospitals will no longer be reimbursed for certain “reasonably preventable” events, beginning in October 2008. Under this new “zero tolerance” policy, certain conditions, such as retained foreign bodies, air embolism, fall from bed, and blood incompatibility, will not be reimbursed if they occur after hospital admission.24 CMS is considering adding DVT and PE to this list. While I completely agree that some things (e.g., wrong-site surgery) are “never events,” DVT should not be on that list, in my opinion. Even with best practice measures and optimal prophylaxis, thromboembolic complications may still occur. We should continue to try our best, but we may never be able to reach the goal of zero DVTs. Clinicians need to help define quality measures using data rather than waiting for administrators and legislators to define them.

References