Venous thromboembolism (VTE), which includes both deep vein thrombosis (DVT) and pulmonary embolism (PE), affects 900,000 Americans annually and is a major cause of morbidity and mortality in hospitalized patients.1–3 In particular, a postoperative patient is at high risk for venous stasis, vascular endothelial injury, and a hypercoagulable state, each of which is predisposing to DVT. The most severe complication of DVT is PE, which occurs in 500,000 patients each year in the United States and is fatal in 300,000 patients annually.1–3 Underuse of pharmacologic VTE prophylaxis among surgical patients has been well-documented in the literature.4–6 Moreover, in the absence of formal VTE risk assessment, clinicians may underestimate overall risk by as much as 50%, leading to further underutilization of chemoprophylaxis.7 In 2005, after careful review of reports specific to our medical center from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), we noted that our institution had an unacceptably high rate of postoperative VTE.
Assuming that a significant number of these complications were potentially preventable, our institution focused substantial resources at reducing VTE rates among our surgical patients. The PA community at our medical center was soon identified as being well-positioned for such a patient safety initiative. PAs in the Department of Surgery designed and implemented a comprehensive clinical pathway to reduce VTE at the University of Michigan. Specifically, at the time the preoperative history was taken and the physical examination was performed, general surgical patients were assessed for the risk of VTE. Using evidence-based guidelines, the examining PA then prescribed the appropriate perioperative VTE prophylaxis.
In this article, we review our institutional efforts to reduce VTE. In particular, we highlight the clinical intervention and its effectiveness. Additionally, we detail the difficulties associated with implementation of a broad clinical protocol within a large institution and the importance of continuous feedback of high-quality clinical data such as that provided by the ACS-NSQIP. We discuss the critical role PAs play to assure patient safety at a large academic medical center. Finally, we describe how the PA community at our large academic medical center is increasingly taking the lead with hospital-wide patient safety initiatives.
We conducted a retrospective study of all adult general surgical patients on the endocrine surgery (GSE) and GI surgical (SGI) services at the University of Michigan who underwent a scheduled operation between July 2005 and June 2007. PAs in the Department of Surgery implemented a VTE assessment and prophylaxis intervention in June 2006. Patients receiving surgery prior to this intervention are referred to as the preintervention group. Patients receiving surgery following this intervention are described as the postintervention group. From an institutional perspective, there were no other clinical initiatives at the University of Michigan to address VTE prophylaxis during this time frame. This research was approved by the University of Michigan Institutional Review Board.
The VTE risk assessment tool Beginning in June 2006, we implemented risk assessment for VTE at the time the preoperative history was taken and the physical examination was performed for elective general surgery cases. The risk assessment tool (www.med.umich.edu/clinical/images/VTE-Risk-Assessment.pdf) was developed with the assistance of Joseph A. Caprini, MD, a national expert in VTE, as well as in consultation with guidelines from the American College of Chest Physicians (ACCP).8–12 The risk assessment was incorporated into the work flow of PAs who were already performing more than 90% of the elective preoperative clinical assessments for the GSE and SGI services. This was determined to be the optimal time to perform risk assessment since a comprehensive history and physical examination were already a part of the work flow. The additional risk assessment piece added approximately 30 to 60 seconds to each patient encounter and was typically performed while orders and other paperwork were being completed for the visit. Once the patient's VTE risk had been assessed, the PA would fill out the preoperative orders, inclusive of perioperative antibiotics, preoperative SC heparin (if indicated based on patient risk), sequential compression devices (SCDs), preoperative laboratory studies, and other necessary orders for the day of surgery. Our protocol required all physicians to actively opt out in advance if they had particular concerns about bleeding risk associated with the proposed procedure. Any patient whose surgeon did not opt out and who had a risk score of 3 or higher (indicating that he or she was at high or highest risk for VTE) would have an order for single-dose SC heparin 5,000 units, ideally to be administered 1 to 2 hours prior to surgery. Following completion of the surgery, the clinician (usually a surgical house officer) who was writing postoperative orders was expected to order appropriate postoperative VTE prophylaxis based on the VTE risk score generated during the preoperative history and physical examination. As an additional safety net, inpatient PAs and/or house staff were expected to ensure that appropriate prophylaxis orders were in place for each patient on morning rounds and to correct any inadequate or absent orders for prophylaxis when necessary.
Methodology Patients who underwent surgery in the preintervention period were compared with patients who underwent surgery in the postintervention period. Patients who underwent surgery prior to the intervention period did not have a preoperative VTE risk score documented in the clinical chart. For this group, a retrospective VTE risk-scoring method was developed to calculate the VTE risk score using data from electronic sources, including the OR information system, the clinical data repository, and the International Statistical Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes from hospital billing data. This method was tested by applying the calculated VTE risk score method to patients in the postintervention period and comparing the results to the risk score documented by PAs using the risk assessment tool. The sensitivity and specificity of the calculated risk score were reported. Exclusion criteria for patients with a calculated risk score included active bleeding, coagulopathies, or use of anticoagulant medications as well as other exclusion criteria (Table 1). We then compared the frequency with which patients in the preintervention and postintervention periods received VTE prophylaxis that satisfied ACCP guidelines.
Postoperative DVT or PE was determined using ICD9-CM diagnosis codes for DVT or PE and identified as “acquired in hospital” or readmitted with a diagnosis code for a preexisting DVT or PE within 30 days following surgery. The diagnoses used to identify patients with VTE were the same as those used by the Agency for Healthcare Research and Quality for its DVT/PE Patient Safety Indicator. (Table 2 lists the diagnoses and their ICD-9-CM codes.) Because the identification of DVT/PE cases relies on an indicator of whether the diagnosis was preexisting or acquired (information that was not captured at our institution prior to November 1, 2005), the preintervention period covered 7 months, from November 2005 through May 2006. The odds ratio of developing VTE within the preintervention group was compared with the rate of VTE within the postintervention group in order to determine the effectiveness of the intervention on incidence of VTE. In addition, the casemix index was compared between the two groups using the all-patient refined diagnosis-related group (APR-DRG) severity index for each patient. The APR-DRG determined the severity of illness of our patient populations, ensuring that any reduction in VTE in the postintervention group was not associated simply with a lower severity of illness for this patient population. Determination of illness severity is based on the age of the patient, the principal and secondary diagnoses, and the use of various surgical and nonsurgical procedures.
In an effort to better understand the patients who had a VTE at our center, we performed an individual chart review for each patient known to have experienced VTE as a complication in either the preintervention or postintervention period. The chart review looked at the clinical characteristics of these patients, including preoperative VTE risk score, body mass index (BMI), and operative duration.
Overall, 2,046 patients underwent surgery during the study period—1,079 patients in the preintervention period and 967 patients in the postintervention period. Among the postintervention group, 772 patients had a PA-documented Caprini risk score of 3 or higher (this number includes those at high and highest risk for VTE), 761 of whom had no documented contraindication to pharmacologic VTE prophylaxis. When applying our retrospective scoring method to calculate risk scores for the postintervention group, the sensitivity and specificity in identifying high and highest risk patients was 93% and 90%, respectively, as compared with the VTE risk score generated by PAs, illustrating the reliability of this retrospective method. Using the same method to calculate retrospective VTE risk scores for the preintervention group, 512 patients had a VTE risk score of 3 or higher, 441 of whom had no documented contraindication to pharmacologic VTE prophylaxis. For patients without documented contraindications to pharmacologic VTE prophylaxis, compliance with VTE prophylaxis recommendations was assessed between preintervention and postintervention groups. Additionally, all patients with a VTE score of 3 or higher in the preintervention period and postintervention period were evaluated for the rates of postoperative DVT/PE.
Compliance with prophylaxis recommendations We compared compliance rates for the recommended VTE prophylaxis between the preintervention and postintervention groups, according to VTE risk score (Figure 1 and Figure 2). Within each graph, we show the percentage of patients in six mutually exclusive subgroups according to the specific VTE prophylaxis that was ordered: (1) no prophylaxis, (2) SCDs only, (3) drug prophylaxis that did not satisfy ACCP guidelines (ie, SC heparin twice a day), (4) a combination of SCDs and drug prophylaxis that did not satisfy ACCP guidelines, (5) drug prophylaxis that satisfied ACCP guidelines (enoxaparin or SC heparin three times a day), and (6) a combination of SCDs and drug prophylaxis that satisfied ACCP guidelines.
Figure 1 depicts all patients with a VTE risk score of 3 or higher (ie, high and highest risk combined), whereas Figure 2 depicts only patients at highest risk (VTE risk score of 5 or higher) based on our risk assessment tool. As evidenced by the trends over time depicted in these graphs, prophylaxis practices improved after the intervention. Use of SCDs alone or in combination with twice-a-day heparin decreased, while use of the recommended dose of medication, alone or in combination with SCDs, increased. According to our protocol, patients with a Caprini score of 3 or higher are candidates for high-risk pharmacologic prophylaxis with either enoxaparin or SC heparin three times a day. For all patients with a risk score of 3 or higher (high and highest risk combined), orders for appropriate pharmacologic prophylaxis improved from an average of 23.1% (range 0%-43% per month) in the preintervention group to an average of 63.7% (range 29%-85% per month) in the postintervention group. For patients at highest risk (Figure 2), attention should be focused on the dark blue portion of the graph, which corresponds to patients receiving combined high-risk pharmacologic prophylaxis plus mechanical prophylaxis (SCDs plus medication recommended). For all patients with a risk score of 5 or higher (highest risk), orders for appropriate prophylaxis improved from an average of 29.4% (range 0%-59% per month) in the preintervention group to an average of 69.5% (range 43%-88% per month) in the postintervention group. Overall, more patients at highest risk (risk score of 5 or higher [Figure 2]) received prophylaxis satisfying the ACCP guidelines as compared with all patients who had a risk score of 3 or higher (Figure 1). While the data demonstrate that suboptimal VTE prophylaxis orders were still being written (ie, SC heparin twice a day for a high-risk patient), this practice was greatly reduced by our intervention, as evidenced by the overall reductions in “medication not recommended” orders depicted in Figure 1 and Figure 2.
Characteristics of patients with a known DVT/PE Over the course of the study, 23 patients were diagnosed with a DVT or PE during their inpatient stay or when readmitted within 30 days of surgery. As measured by the APR-DRG inpatient classification system, the preintervention group had a lower case-mix index (1.82) as compared with the postintervention group (2.10), indicating that patients in the preintervention group had fewer comorbid conditions. Despite the higher severity of illness in the postintervention group, this patient population had a lower incidence of VTE. The overall rate of VTE in the preintervention group was 2.2% compared with 1.6% in the postintervention group (P = not significant [NS]). Similarly, the odds ratio of VTE in the preintervention group compared with the postintervention group was 1.39 (95% confidence interval 0.61–3.18, P = NS [Figure 3]).
The clinical characteristics of patients known to have acquired a DVT or PE after surgery, before and after intervention were detailed (Table 3). A medical record review for each of the 23 patients was then conducted to determine whether each had received appropriate VTE prophylaxis commensurate with individual patient risk during the preoperative, intraoperative, and postoperative periods. We determined that all patients had a risk assessment score of 3 or higher (range, 3–9; mean, 6) (Figure 4). (Three patients were underscored, as determined by our manual chart review.) All 23 patients should have received SC heparin three times a day (or enoxaparin one to two times a day) per protocol. In the preintervention period, eight of 11 (73%) patients were prescribed inadequate prophylaxis. In contrast, only three of 12 (25%) patients with a known VTE had inadequate prophylaxis prescribed in the postintervention period, further underscoring the effectiveness of the intervention.
In addition, 15 of 23 (65%) patients were male with an age range of 36 to 84 years (mean 60.0 years), with 21 of 23 (91%) patients at VTE risk because of their age. Perhaps most noteworthy was that 16 of 23 (70%) patients had either active cancer or a history of cancer. Inflammatory bowel disease was identified in five of 23 (22%) patients, only two of whom had concomitant cancer. Prior history of VTE was documented in six of 23 (26%) patients. Obesity was a factor, with BMI that ranged from 18 to 51 kg/m2 (mean 31.8 kg/m2) (Figure 5); 12 of 23 (52%) patients had a BMI higher than 30 kg/m2, traditionally recognized as obese, whereas 19 of 23 (83%) patients had a BMI higher than 25 kg/m2, placing them at increased risk for VTE according to our Caprini risk assessment tool. Operative duration ranged from 1.5 to 10 hours (mean 6.1 hours) (Figure 6). VTE diagnosis was confirmed between 0 and 27 days postoperatively (mean 8.2 days).
Center-specific reports from the ACS-NSQIP clearly indicated that patients at the University of Michigan had a high rate of VTE. In response, we instituted a broadranging quality improvement initiative that focused on reducing VTE. The unique aspect of this program was that it was developed and initiated by PAs within the Department of Surgery. This PA-driven initiative achieved greater compliance with published guidelines for VTE prophylaxis, which has resulted in significant improvements in how patients are cared for at the University of Michigan. The success of this initiative has resulted in expansion of these efforts to other medical and surgical services at the university, with PAs providing leadership and oversight. We have developed a PA-run center that now provides preoperative risk assessment for more than 90% of our elective cases across all surgical services. Additionally, the University of Michigan Health System now requires that this risk assessment be performed on all medical and surgical patients at the time of admission—a significant system-wide change.
SEE THE THE ONLINE VERSION OF THIS ARTICLE TO LINK TO
Caprini Risk Assessment Tool
Not only was the initiative successful, but it also demonstrated the unique role PAs can fulfill in assuring patient safety. Large academic medical centers present unique challenges with respect to implementing institution-wide patient safety initiatives. Potentially competing priorities for physicians at academic medical centers include research, administrative duties, education, and clinical care. Within this context, care protocols are frequently enacted by physician trainees. Unfortunately, physician trainees frequently rotate through various departments, affecting the consistency with which care can be delivered to patients. These factors create an environment in which relying on physician trainees to implement broad patient safety protocols is difficult. In our large academic center, PAs are increasingly being employed to help bridge this gap by providing greater continuity and coordination of care. Indeed, our experience with the traditional resident model prior to this study resulted in unacceptably low adherence to published guidelines for VTE prophylaxis as compared with that achieved by the PAs. Consequently, as stable front-line providers of care, PAs are ideal for implementing and providing oversight for patient safety initiatives, such as VTE risk assessment and prophylaxis.
The identification of DVT/PE cases relies on an indicator of whether the diagnosis was preexisting or acquired, information that was not captured at our institution prior to November 1, 2005; consequently we were unable to identify patients with an in-hospital-acquired DVT or PE prior to that date, resulting in analysis for only a portion of the desired 1-year preintervention time frame. This likely resulted in an underestimation of the total number of VTE events as well as an underestimation of the number of patients with a VTE who were prescribed inadequate VTE prophylaxis in the preintervention time period. In addition, some patients with acquired VTE might not be identified because of incomplete coding of these conditions. Similarly, we are unable to identify patients with acquired VTE who received treatment at an outside hospital, potentially resulting in a further underestimation of the overall incidence of this complication. Thus the number of acquired VTE cases both before and after intervention is likely understated.
Our protocol increased the number of patients within our health system who were prescribed appropriate orders for VTE prophylaxis according to published guidelines and according to individual patient risk. While we did not reduce our overall VTE complication rates by statistically significant numbers within the time frame analyzed, we are encouraged by the dramatic reductions in the practice of prescribing twice-a-day heparin for at-risk patients. Consequently, we are confident that we have improved the care of patients who are at risk for VTE complications in our hospital. Implementation of large-scale patient safety initiatives is fraught with many challenges, particularly in large academic medical centers. As hospitals across the nation continue to struggle with how to optimize VTE prophylaxis for their patients, they should take note that PAs are well-positioned to develop, implement, and provide oversight for important patient safety initiatives such as this. Similarly, PAs should recognize the opportunities that exist to become active leaders in their institutions’ patient safety endeavors.
1. Heit JA. The epidemiology of venous thromboembolism in the community. Arterioscler Thromb Vasc Biol.
2. Raskob GE, Silverstein R, Bratzler DW, et al. Surveillance for deep vein thrombosis and pulmonary embolism. Am J Prev Med.
4. Conti S, Daschbach M. Venous thromboembolism prophylaxis: a survey of its use in the United States. Arch Surg.
5. Anderson FA Jr, Wheeler HB, Goldberg RJ, et al. Physician practices in the prevention of venous thromboembolism. Ann Intern Med.
6. Bratzler DW, Raskob GE, Murray CK, et al. Underuse of venous thromboembolism prophylaxis for general surgery patients: physician practices in the community hospital setting. Arch Intern Med.
7. Holley AB, Moores LK, Jackson JL. Provider preferences for DVT prophylaxis. Thromb Res.
8. Caprini JA. Thrombosis risk assessment as a guide to quality patient care. Dis Mon.
9. Caprini JA, Tapson VF, Hyers TM, et al. Treatment of venous thromboembolism: adherence to guidelines and impact of physician knowledge, attitudes, and beliefs. J
10. Monagle P, Chan A, Massicotte P, et al. Antithrombotic therapy in children: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest.
11. Levine MN, Raskob G, Beyth RJ, et al. Hemorrhagic complications of anticoagulant treatment: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest.
2004; 126(3 suppl):287S-310S.
© 2010 American Academy of Physician Assistants.
12. Geerts WH, Pineo GF, Heit JA, et al. Prevention of venous thromboembolism: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest.