| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Clinical Pharmacy Anticoagulation Service (Drs. Witt, Sadler, and Shanahan), Kaiser Permanente of Colorado, Westminster, CO; the Anticoagulation Clinic (Dr. Mazzoli), Kaiser Permanente of Ohio, Cleveland, OH; and Clinical Research (Dr. Tillman), Exempla St. Josephs Hospital, Denver, CO.
Correspondence to: Daniel M. Witt, PharmD, FCCP, 11245 Huron St, Westminster, CO 80234; e-mail: dan.m.witt{at}kp.org
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Objective: To compare clinical outcomes associated with anticoagulation therapy provided by a clinical pharmacy anticoagulation service (CPAS) to usual care.
Design: Retrospective, observational cohort study, 6 months in duration.
Setting: Large nonprofit, group-model health maintenance organization.
Patients: A total of 6,645 patients receiving warfarin therapy were included in the final analyses (intervention group, 3,323 patients; control group, 3,322 patients).
Intervention: Anticoagulation therapy for patients in the intervention group was managed by a centralized, telephonic CPAS. Therapy for patients in the control group was managed in the usual manner by their personal physicians.
Main outcome measures: The primary outcome was the occurrence of anticoagulation therapy-related complications. A secondary outcome was the proportion of time spent in the target international normalized ratio (INR) range for each patient. Cox proportional hazards regression analyses were used to examine the risk of complications in relation to the study group.
Results: Patients in the CPAS were 39% less likely to experience an anticoagulation therapy-related complication than were patients in the control group (hazard ratio, 0.61; 95% confidence interval, 0.42 to 0.88). The number of patients needed to treat to prevent an anticoagulation therapy complication was 52. Additional analyses revealed that improved outcomes associated with CPAS were mediated largely through improved therapeutic INR control. Patients in the CPAS group spent 63.5% of study period days within their target INR range compared to 55.2% in the control group (p < 0.001).
Conclusions: A centralized, telephonic, pharmacist-managed anticoagulation monitoring service reduced the risk of anticoagulation therapy-related complications compared to that with usual care. The cumulative evidence supporting the superior care associated with implementing a pharmacist-managed anticoagulation monitoring service was sufficient to recommend widespread implementation.
Key Words: anticoagulant drugs • health-care quality assessment • pharmacists
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Much of the available information in the medical literature describes AMS models in which practitioners conduct in-depth face-to-face interviews at each patient visit. Descriptive reports78 of AMS models utilizing telephone or mail systems for most patient care activities exist, but formal assessments of clinical outcomes associated with these models are not available. The quantification of the clinical outcomes associated with telephonic services would be useful since this model allows large numbers of patients to be managed by relatively few anticoagulation therapy providers. This study addresses some of the limitations of prior studies and provides insights into the effectiveness of telephonic anticoagulation management by comparing clinical outcomes associated with anticoagulation therapy provided by a centralized, pharmacist-managed AMS that provides the majority of care via the telephone to usual care in a large, diverse sample of anticoagulated patients enrolled in a group model health maintenance organization (HMO).
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study Design
This was a retrospective, observational cohort study with a combined historical and parallel control group. The intervention group consisted of patients who received warfarin anticoagulation therapy that was managed by the staff of the KPCR Clinical Pharmacy Anticoagulation Service (CPAS) during the 6-month evaluation period (April 1, 1999, to September 30, 1999). The historical control group was composed of KPCR patients who received warfarin anticoagulation therapy that was managed by their KPCR physician during the 6-month evaluation period (April 1, 1996, to September 30, 1996, just prior to the implementation of CPAS in October 1996). The parallel control group was composed of KPOR patients who had received warfarin anticoagulation therapy that was managed by their physician during the 6-month evaluation period (April 1, 1999, to September 30, 1999, the same time period evaluated for the CPAS group). Some KPCR patients who were receiving long-term warfarin therapy were managed both by their physician and later by CPAS. These patients were randomly assigned to the control or intervention group (ie, patients only contributed data to either the CPAS or the control group).
Intervention
The CPAS is a centralized team of pharmacy technicians, clinical pharmacists (ie, those with a BS or PharmD degree without residency training), and clinical pharmacy specialists (ie, those with a PharmD degree with residency training), with specialized knowledge and skill in the coordination and management of anticoagulation therapy. This team is available by pager 24 h per day, 7 days per week. Most CPAS patient care activities are conducted via telephone and mail. CPAS staff members act as the agent of the referring physician and facilitate all aspects of anticoagulation therapy, including patient education, the ordering of relevant laboratory tests including international normalized ratios (INRs), the adjustment of anticoagulation medication doses, the planning for interruption of anticoagulation therapy during invasive procedures, and the management of adverse events. All related activities and outcomes were documented in a comprehensive computerized patient monitoring system (Dawn AC; 4S Systems, Ltd; Cumbria, UK).
Patients
Patients were included in the analysis if they were at least 18 years of age, had received at least one prescription for warfarin, and had at least two INR values measured during the 6-month evaluation period (Fig 1
). Patients who received anticoagulation therapy while residing in a nursing home facility were excluded from the study because access to INR data was not routinely available for these patients. Patients were also excluded if the sole indication for warfarin anticoagulation therapy was the prevention of thrombosis in an indwelling central venous catheter, since these patients typically receive unmonitored, fixed-dose warfarin therapy. Patients receiving warfarin for the prevention of venous thromboembolism following a high-risk surgical procedure (eg, hip or knee replacement surgery) were also excluded because the duration of warfarin therapy following high-risk surgery is often only days to weeks.
|
2 g/dL, or any intracranial, intraarticular, intraocular, or retroperitoneal bleeding. The occurrence of thromboembolic complications was identified using computerized claims and referral data, and was verified by medical record review. Because patients with recurrent deep vein thrombosis (DVT) are often managed as outpatients in our health-care system, low-molecular-weight heparin dispensing information and medical record review were used to identify and confirm recurrent DVT episodes that were treated outside of the hospital. Thromboembolic recurrence was defined as an embolic or thrombotic cerebrovascular accident, pulmonary embolism, DVT, or other systemic thromboembolic event. Fatal events that were directly attributable to bleeding or thromboembolic recurrence were verified through medical record and/or death certificate review. When available, the INR value occurring at the time of an anticoagulation therapy-related adverse event was recorded. If the INR was not measured on the day of the adverse event, the most recently measured INR prior to the event was used.
The secondary outcome was time spent in the target INR range. This was estimated using linear interpolation between actual INR values and was reported as a percentage of total patient-days of therapy for each patient. The elapsed time in days between successive measurements of INR values was used to calculate the mean interval between INR values for each patient. The elapsed time, in days, between INR values 4.0 or 
1.5 and the next measured INR was used to assess the timeliness of the response to INR values that would generally require clinical intervention.6
Demographic data (ie, gender and age), information regarding all indications for warfarin therapy, and the target INR range were collected for each patient. When multiple indications for warfarin therapy were recorded, the primary indication was designated as the one requiring the highest target INR range or the longest duration of therapy. When the target INR range was unknown (CPAS group, 1 patient; control group, 83 patients), a range of 2.0 to 3.5 (which encompasses the most commonly used target ranges) was arbitrarily assigned to allow the calculation of time spent in the target INR range during the evaluation period.
Statistical Analysis
To compare the baseline characteristics between patients in the intervention and control groups, the 
2 test, the Mann-Whitney rank-sum test, and the Student t test were used. The 2 test was also used to compare the intervention and control groups with regard to overall therapeutic INR control, as measured by the proportion of time that the patient spent in their target INR range. Kaplan-Meyer survival analysis was performed to compare the rates of complications between the groups. Cox proportional-hazards regression modeling with censoring was used to estimate the hazard ratios and 95% confidence intervals (CIs) for anticoagulation therapy-related complications among patients in the intervention group in relation to patients in the control group, with adjustment for the effects of age, gender, and indication for warfarin therapy. A global model was constructed with the time interval from enrollment in the study until the first occurrence of any anticoagulation therapy-related complication as the dependent variable. In addition, Cox modeling was performed to estimate the hazard ratios for each of the individual complications (ie, major bleeding, thromboembolism, and fatal events). To assess whether any apparent benefits of the anticoagulation clinic were mediated by differences in therapeutic INR control, an additional regression analysis was conducted that included the proportion of time spent within the target INR range as an additional covariate in the global model.
To assess the soundness of the method for assessing the relationship between the study group and the occurrence of anticoagulation therapy-related complications, we repeated the analyses using a general linear model. The outcome was the presence or absence of any anticoagulation therapy-related complication. Models were adjusted for age, gender, indication for warfarin therapy, and the duration of patient-days spent in the study. Since the results obtained in this manner were very similar to those obtained in the primary analysis, only the results from Cox proportional-hazards regression are presented.
Initial analyses were conducted to compare outcomes between the KPCR and KPOR control groups. The results of these analyses indicated that these groups were similar (Table 1 ). Thus, the analyses presented compare only the intervention group to the single, combined control group.
|
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
4.0 or 1.5 was significantly lower in the CPAS group compared to that in the control group (15.1% vs 20.4%, respectively; p < 0.001). The time between INR values of 4.0 or 1.5 and follow-up INR testing was also significantly lower in the CPAS group (p = 0.03).
|
|
|
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To assess whether any apparent benefits of the anticoagulation clinic were mediated by differences in therapeutic INR control, we conducted an additional regression analysis that included the percentage of time spent within the target INR range as a covariate in the model. The addition of therapeutic INR control to the model attenuated the effect of study group assignment on the occurrence of complications, suggesting that the clinical benefits associated with the intervention were largely mediated through the superior INR management achieved by CPAS pharmacists. Improved INR control in the CPAS group likely resulted from several factors, including the use of a computerized patient monitoring system that assisted in the timely identification of patients who failed to return for INR testing as instructed and the specialized anticoagulation management experience of CPAS pharmacists. Control group patients were managed by providers with less hands-on experience in the management of warfarin therapy who were using more basic monitoring systems (eg, paper flow-sheets stored in three-ring binders).
The results of this study support other research1235 that has shown that clinically trained pharmacists using a structured, specialized approach to managing oral anticoagulation therapy can achieve superior outcomes compared to those using an ad hoc approach. Other investigators6 have documented the feasibility of implementing various AMS models in managed care settings. These investigators noted variability in outreach efforts and physician acceptance among the study sites. Some physicians desired to maintain direct oversight of anticoagulation therapy, while others who cared for patients from several managed care organizations hesitated to use an AMS that was available only to certain patients. These potential barriers were overcome in our organization by the aligned relationship between physicians and CPAS pharmacists within the integrated structure of KPCR.
The large number of patients included in this analysis allowed for precise estimates of anticoagulation therapy-related complication rates, including fatal adverse events. The diverse study sample included a wide variety of indications for and durations of anticoagulation therapy, and patients with newly initiated and long-standing warfarin therapy were included. Thus, we think that the patients in this study provided a realistic representation of typical patients receiving anticoagulation therapy.
Of note, this study utilized a control group that consisted of patients managed by experienced physicians from two geographic regions who were practicing in an integrated managed care setting. Other studies239 evaluating the impact of an AMS on anticoagulation therapy outcomes have utilized control groups that included medical residents or other less experienced clinicians. The rate of anticoagulation therapy complications seen in our control group was comparable to that associated with a structured anticoagulation service in another study.3 Despite the high quality of anticoagulation therapy provided by the physicians in the control group, the CPAS was able to demonstrate a statistically significant reduction in anticoagulation therapy complications.
The CPAS had the greatest impact on reducing thromboembolic complications. This may be attributable to the greater percentage of time that CPAS patients spent in their target INR range compared to those in the control group. Subgroup analyses of major clinical trials10 evaluating stroke prevention in patients with atrial fibrillation have demonstrated that most patients who experienced strokes while receiving anticoagulation therapy had subtherapeutic INR values. Therefore, increasing the amount of time that patients spend in their target INR range is likely to reduce the number of patients experiencing a stroke. In this study, the rate of stroke in the control group was approximately three times that of the intervention group. The differences in occurrences of major bleeding and fatal complications between the CPAS and control groups were not statistically significant. However, an earlier evaluation11 of the outcomes of excessive anticoagulation (INR, > 6.0) in our CPAS demonstrated a dramatic reduction in the rate of major bleeding complications compared to usual care (1.3% vs 6.3%, respectively). Patients in the control group tended to present with higher INRs during major bleeding episodes compared to those in the intervention group (4.3 vs 3.1, respectively), although this difference was not statistically significant.
An acknowledged limitation of this study was the inability to randomly assign patients into the CPAS or control group. It is unlikely that unaccounted for differences between the groups explained the benefit of CPAS, since overall the groups appeared to be demographically and clinically similar. In addition, differences between the groups were controlled for in the analysis. Another limitation that was common to retrospectively collected data related to extracting complete historical information from patient medical records. The availability of comprehensive laboratory datasets facilitated the almost complete capture of INR data in both the KPCR and KPOR study samples. Similarly, the initial identification of potential anticoagulation therapy adverse events was facilitated by the availability of electronic claims and referral datasets. The use of predefined ICD-9 codes reduced potential bias in adverse event selection. This study was conducted in a group model HMO with an integrated structure. The results may therefore not be generalizable to other settings.
The control group was composed of patients from two different Kaiser Permanente regions (KPCR and KPOR) and two different time periods (1996 and 1999). This resulted in qualitative differences in the 1996 and 1999 medical records. Medical records from 1999 were available electronically, and were more complete and easily abstracted than were paper medical records from 1996. This may have led to an underestimation of control group complication rates. However, subgroup analyses of data from the two control samples did not reveal substantial differences in study outcomes between these groups.
Complications were identified through administrative claims and referral datasets using predefined ICD-9 codes, and were, therefore, mostly limited to events that resulted in hospitalization or emergency department visits. Because DVT is generally treated on an outpatient basis in our organization,12 we used information about the dispensing of low-molecular-weight heparin to identify this type of thromboembolic complication. All adverse events were verified through medical record review; however, it was not possible to blind review the study group assignment. We attempted to minimize the effects of bias in complication ascertainment by using a standardized abstraction form with objective criteria during medical record reviews. The cause of death was verified for all patients who died during the study period using medical records and death certificates. It is well-known that death certificates may not always provide accurate information regarding the cause of death. Eighty-three patients in the control group were arbitrarily assigned a target INR range of 2.0 to 3.5 per the study protocol compared to only one patient in the CPAS group. This likely biased the comparison of therapeutic INR control rates between the two groups toward the null hypothesis.
Most patients receiving warfarin therapy in the United States are not enrolled in a structured AMS.13 This study provides evidence that a centralized clinical pharmacist-managed AMS, in which the majority of patient care activities are conducted by phone or through the mail, improves anticoagulation therapy outcomes compared to those with usual care. A coordinated, systematic approach to anticoagulation therapy may be more important than the method of management (ie, by telephone or in-person). Health-care organizations should strive to develop AMS models that meet system-specific needs. An AMS utilizing point-of-care INR testing and serving a relatively small number of patients may efficiently utilize in-person management. A centralized telephonic model like CPAS may provide the leverage needed for a relatively small number of providers to efficiently manage a large patient population over a wide geographic area. AMS models that include a systematic process utilizing a knowledgeable provider, reliable laboratory monitoring, and an organized system for timely patient follow-up and education will result in improved outcomes regardless of model type.13 Health-care systems with less-than-adequate approaches for managing patients who are receiving warfarin are likely to receive even greater benefits from implementing a pharmacist-managed AMS, provided that there is acknowledgment of the need for the improvement in anticoagulation therapy practices. Not all health-care systems meet this criterion.6
Although we did not specifically evaluate the economic impact of the CPAS, other studies have demonstrated that reduced complication rates result in reduced health-care costs.2314 Significant cost savings associated with services provided by our CPAS have been demonstrated previously.12 Implementation of an AMS could also reduce the need to prescribe more expensive anticoagulants like ximelagatran that have not been shown to be superior to well-managed warfarin therapy. We think that the cumulative evidence supporting the superior care associated with implementing a pharmacist-managed AMS is sufficient to recommend their widespread implementation.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Received for publication October 25, 2004. Accepted for publication November 25, 2004.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. G. Johnson, K. Rogers, T. Delate, and D. M. Witt Outcomes Associated With Combined Antiplatelet and Anticoagulant Therapy Chest, April 1, 2008; 133(4): 948 - 954. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Lane and G. Y. H. Lip Maintaining Therapeutic Anticoagulation: The Importance of Keeping "Within Range" Chest, May 1, 2007; 131(5): 1277 - 1279. [Full Text] [PDF]
|
|
|
|
|
|
C. van Walraven, N. Oake, P. S. Wells, and A. J. Forster Burden of Potentially Avoidable Anticoagulant-Associated Hemorrhagic and Thromobembolic Events in the Elderly Chest, May 1, 2007; 131(5): 1508 - 1515. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. K. Wittkowsky, E. A. Nutescu, J. Blackburn, J. Mullins, J. Hardman, J. Mitchell, and V. Vats Outcomes of Oral Anticoagulant Therapy Managed by Telephone vs In-Office Visits in an Anticoagulation Clinic Setting. Chest, November 1, 2006; 130(5): 1385 - 1389. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Monte, A. Macchia, F. Pellegrini, M. Romero, V. Lepore, A. D'Ettorre, M. Saugo, L. Tavazzi, and G. Tognoni Antithrombotic treatment is strongly underused despite reducing overall mortality among high-risk elderly patients hospitalized with atrial fibrillation Eur. Heart J., September 2, 2006; 27(18): 2217 - 2223. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. M. Stack and K. M. Rudd Advanced pharmacy practice for new practitioners. Am. J. Health Syst. Pharm., June 15, 2006; 63(12): 1126 - 1127. [Full Text] [PDF]
|
|
|
|
 |
|
 A. G. Staresinic, C. A. Sorkness, B. M. Goodman, and D. W. Pigarelli Comparison of outcomes using 2 delivery models of anticoagulation care. Arch Intern Med, May 8, 2006; 166(9): 997 - 1002. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. van Walraven, A. Jennings, N. Oake, D. Fergusson, and A. J. Forster Effect of Study Setting on Anticoagulation Control: A Systematic Review and Metaregression Chest, May 1, 2006; 129(5): 1155 - 1166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M Zeolla, M. R Brodeur, A. Dominelli, S. T Haines, and N. D Allie Development and Validation of an Instrument to Determine Patient Knowledge: The Oral Anticoagulation Knowledge Test Ann. Pharmacother., April 1, 2006; 40(4): 633 - 638. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
