Effective Clinical Practice
Effective Clinical Practice, July/August 2000.
Leigh Hamby, MD, William B. Weeks, MD, MBA, and Carolyn Malikowski, PharmD
For author affiliations, current addresses, and contributions, see end of text.
Context. Anticoagulation with warfarin requires careful management to avoid hemorrhage or thrombosis. The anticoagulation clinic has been suggested as a mechanism to reduce complications related to anticoagulation.
Objective. To report our experience with anticoagulation complications, the cost of subsequent care, and the role of the anticoagulation clinic.
Design. Sequential patients who were receiving warfarin within a period of 4 months were followed to identify warfarin-related adverse events. An independent examiner reviewed medical records to determine whether events were preventable and to identify possible causes. Hospital-based accounting data were used to determine attributable costs.
Patients. 306 patients who received warfarin prescriptions at a rural Vermont university-affiliated VA hospital with an established anticoagulation clinic.
Results. 91% (278) of patients received follow-up at the anticoagulation clinic, and the remaining 9% (28) were followed by VA physicians without involving the anticoagulation clinic. A total of 12 patients had adverse events associated with either sub- or supratherapeutic international normalized ratios, with an attributable cost of approximately $90,000; 8 of these patients were not enrolled in the anticoagulation clinic. Thus, the estimated relative risk for adverse events for patients not at the clinic, compared with those who were, was almost 20 (95% CI, 6.4 to 61.8). Review of the remaining 4 patients revealed that their problems were attributable either to missed appointments or lack of coordination between other providers and the anticoagulation clinic.
Conclusions. Establishing an anticoagulation clinic is only the first step toward reducing complications related to anticoagulation. The larger challenge is ensuring that patients use the anticoagulation clinic and that providers communicate with it. Our results suggest that our institution could invest considerable resources to meet this challenge and still save money.
Warfarin is commonly used to anticoagulate patients for a variety of indications, both therapeutic (e.g., treating deep venous thrombosis, maintaining bypass graft patency) and prophylactic (e.g., preventing stroke in atrial fibrillation). However, warfarin has a narrow therapeutic window (i.e., it is easy to over or under dose) and is commonly associated with adverse events, mainly bleeding. (1) To reduce these events, use of special anticoagulation clinics for monitoring patients receiving warfarin has been advocated, (2-5) and consensus guidelines for anticoagulation clinic organization and management have been published. (6) Consequently, the number of health systems that use specialty anticoagulation clinics has increased.
Although there has never been a randomized trial of anticoagulation clinics, the Anticoagulation Guidelines Task Force recently summarized several observational studies and reported a combined major hemorrhage rate of 11% per patient per year among patients receiving routine medical care versus 3% among patients in anticoagulation clinics. (6) Many of these studies, however, were conducted when complication rates were higher than those of today (before standardized reporting of anticoagulation activity by using the international normalized ratio [INR]). In addition, the observational studies are all potentially confounded by selection bias. Specifically, independent factors of patients enrolled in anticoagulation clinics (e.g., above-average compliance and ability to engage the health care system) may contribute to observed improved outcome. These factors may also affect who participates in future randomized trials and may threaten their generalizability.
The challenge inherent in systems currently using an anticoagulation clinic model is in evaluating the experience of all patients receiving warfarin, including those who never enroll in clinic (or who miss their appointments) and who may or may not be managed by other providers. (7, 8) It is equally important to determine the financial cost of failed care processes so that organizations can determine what resources might be devoted to improving the processes. (9) In this case report, we sought to evaluate our system's experience with ambulatory patients receiving warfarin. Specifically, we ascertained the number, cause, and cost of preventable warfarin-related adverse events and identified opportunities for improvement.
The White River Junction VA facility is an acute care hospital in rural Vermont and part of the VA New England Network. Our facility currently has an established anticoagulation clinic staffed by a full-time Doctoral of Pharmacology. Patients are referred to the clinic by their clinicians. In conjunction with the patient's referring physician or primary care provider, the pharmacist uses protocols to determine warfarin dosing and monitoring interval.
Patient Selection and Follow-up
Figure 1 provides an overview of our study design. During a 4-month period, one of us (LH), who had no patient care responsibilities in the anticoagulation clinic, was electronically notified of each ambulatory patient receiving a warfarin prescription processed through the facility's pharmacy. Clinical information on all subsequent encounters for each patient in the cohort was collected by using the electronic medical record. Specifically, the investigator reviewed the location and frequency of follow-up monitoring and records of clinic visits or hospitalizations. Any encounter involving hemorrhage or thrombosis in patients who were prescribed warfarin were reviewed to determine whether the adverse event was preventable and to determine a possible cause. A preventable warfarin-related adverse event was defined as hemorrhage in a patient receiving warfarin who had an INR greater than 5 or thrombosis in a patient who had an INR less than 2. Clinical records of patient care encounters outside of the VA facility were not reviewed.
The cost of additional care given to patients experiencing warfarin-related adverse events was retrieved retrospectively from the VA Decision Support Software (DSS) cost accounting system. This system provides direct costs (variable and fixed) on the basis of a labor-time allocation strategy. Included in costs are labor; supplies; equipment depreciation; and a portion of professional services rendered, based on a calculated sum total of unit costs for all patient encounters related to adverse warfarin-related events. Cost estimates derived from DSS data are believed to be conservative. (10)
During a 4-month period, 395 patients received prescriptions for warfarin and were monitored in a variety of settings (Figure 2). To learn more about why 89 patients were not followed by the VA, we performed a simple telephone survey and found that most thought the VA was too far from home (Figure 3).
We then focused on the 306 patients who were followed at the VA facility. Fourteen patients (5%) had a warfarin-related adverse event, including one death. Two of these events were judged not to be preventable--while receiving warfarin, one patient developed gastrointestinal hemorrhage from an ulcer and another had hemoptysis from a new pulmonary nodule. Both patients had therapeutic INRs.
Table 1 provides clinical details of the 12 events that were deemed preventable. Most events involved bleeding related to a supratherapeutic INR. Table 1 also shows that most patients who had adverse events were not enrolled in the anticoagulation clinic. In fact, the relative risk for adverse events among patients not enrolled in the anticoagulation clinic was 20 times that of enrolled patients ([8 of 28]/[4 of 278] = 19.9; CI, 6.4 to 61.8). For the four patients enrolled in the clinic, the adverse events were attributable to either missed appointments or lack of coordination between other providers and the anticoagulation clinic.
The cost of preventable warfarin-related adverse events during the 4-month study was $90,000, suggesting an estimated annual cost of $270,000. The cost of preventable warfarin-related adverse events in patients not followed by the anticoagulation clinic was 75% of the total cost of preventable warfarin-related adverse events identified.
Our examination of the frequency and cause of warfarin-related adverse events in a single system showed that the events were common and generally involved patients who did not use the anticoagulation clinic. We also found the clinical sequelae and cost of these events to be significant. These results suggest that our institution, which uses an anticoagulation clinic, might invest considerable resources to improve the processes of care related to warfarin (as much as $270,000 annually) and still save money. These results are consistent with those of others who have found potential savings in the use of anticoagulation clinics. (11,12)
We believe that there are several ways our organization might respond to these findings. Most patients who did not use the anticoagulation clinic reported that it was too far from home; we envision two possible solutions to this problem. The simplest would be to eliminate the practice of prescribing warfarin to patients who cannot be monitored by the VA facility. An alternate strategy would be to develop a monitoring system that does not involve traveling to the clinic, such as the telephone or self-monitoring. Another problem we found was failure to inform the anticoagulation clinic of important care events (e.g., surgery, liver biopsy, cardioversion, and hospital discharges). To address this, the anticoagulation clinic might be supplemented with case-management staff. Finally, to prevent missed anticoagulation clinic appointments, investing in patient tracking programs with clinical-decision support capabilities should be considered. (13)
Our study has several limitations. We have no information about the experience of patients who were not monitored at the VA facility. In addition, observation of patients monitored at the VA facility was limited in both time and scope. The length of observation (not more than 4 months) undoubtedly underestimated the warfarin-related adverse events of all study patients. The scope was limited to events within the VA facility and did not include the experience of patients who had an adverse event treated outside of the facility. These limitations make our results an underestimate of the total impact of warfarin-related adverse events.
Perhaps the greatest limitation is that our data in no way prove that moving all patients into an anticoagulation clinic (and providing better coordination) would have prevented their adverse events. Lack of perfect evidence, however, should not be used to support inaction. Organizational efforts should be made to avoid the adverse events described here. Getting all patients into the anticoagulation clinic would seem to be a logical place to start.
Establishing an anticoagulation clinic is only the first step toward reducing adverse events related to anticoagulation. The larger challenge is ensuring that patients use the anticoagulation clinic and that providers can easily communicate with it. These results suggest that our institution might invest considerable resources to meet this challenge and still save money.
|Take Home Points
1. Landefeld CS, Beyth RJ. Anticoagulant-related bleeding: Clinical epidemiology, prediction and prevention. Am J Med. 1993;95:315-28.
2. Pubentz MJ, Calcagno DE Jr, Teeters JL. Improving warfarin anticoagulation therapy in a community health system. Pharm Pract Manage Q. 1998;18:1-16.
3. Meisel S, Sershon L, White D. Reducing adverse drug events and medication errors using rapid cycle improvement. Qual Manage Health Care. 1998;6:15-28.
4. Foss MT, Schoch PH, Sintek CD. Efficient operation of a high-volume anticoagulation clinic. Am J Health Sys Pharm. 1999;56:443-9.
5. Chiquette E, Amato MG, Bussey HI. Comparison of an anticoagulation clinic with usual medical care: Anticoagulation control, patient outcomes, and health care costs. Arch Intern Med. 1998;158:1641-7.
6. Ansell JE, Buttaro ML, Thomas OV, Knowlton CH. Consensus guidelines for coordinated outpatient oral anticoagulation therapy management. Anticoagulation Guidelines Task Force. Ann Pharmocother. 1997;31:604-15.
7. Samsa GP, Matchar DB, Goldstein LB, et al. Quality of anticoagulation management among patients with atrial fibrillation. Arch Intern Med. 2000;160:967-73.
8. Ansell JE. The quality of anticoagulation management. Arch Intern Med. 2000;160:895-6.
9. Johnson LC, Batalden PB, Corindia JT, Marrin CA, Nelson EC, Plume SK. Clinical process cost analysis: A promising tool for clinical improvement. Qual Manage Health Care. 1997;
10. Barnett PG, Rodgers JH. Use of decision support for VA cost-effectiveness research. Med Care. 1999;37(4 suppl Va):AS63-70.
11. Lafata JE, Martin SA, Kaatz S, Ward RE. The cost-effectiveness of different management strategies for patients on chronic warfarin therapy. J Gen Intern Med. 2000;15:31-7.
12. Health Insurance Association of America. In: Source Book of Health Insurance Data, 1996. Washington DC; 1997:854-60.
13. Walton R, Dovey S, Harvey E, Freemantle N. Computer support for determining drug dose: systematic review and meta-analysis. BMJ. 1999;318:984-90.
Leigh Hamby, MD, Dartmouth Medical School, Hanover, NH, National Quality Scholars, Fellowship Program, Department of Veterans Affairs, White River Junction, Vt; William B. Weeks, MD, MBA, Community and Family Medicine, Dartmouth Medical School, Hanover, NH, VA National Quality Scholars, Fellowship Program, Center for Learning and, Improvement of Patient Safety, VISN 1 Patient Safety Center, Department of Veterans Affairs, White River Junction, Vt; Carolyn Malikowski, , PharmD, Clinical Pharmacy Specialist, Department of Veterans Affairs, White River Junction, Vt
Leigh Hamby, MD, Fellow, VA National Quality Scholars Fellowship Program, VAMC (11Q), White River Junction, VT 05009; telephone: 802-291-6283; fax: 802-291-6286.