 |
 |
 |
 |
 |
 |
 |
|
|
|
|||||||||
|
|||||||||||
Clinical Pharmacy Specialist, Anticoagulation Management Service, Department of Pharmacy, Allegheny General Hospital, Pittsburgh, PA
Director, Anticoagulation Management Service; Assistant Professor of Medicine, Division of Hematology, Johns Hopkins Medical Institutions, Baltimore, MD
PharmD Student, Department of Pharmacy, Ohio Northern University, Ada, OH; The Johns Hopkins Hospital
Associate Director, Clinical and Decentralized Services, Department of Pharmacy, The Johns Hopkins Hospital
Director, Center for Pharmaceutical Outcomes and Policy, Department of Pharmacy, The Johns Hopkins Hospital
Reprints: Dr. Streiff, Johns Hopkins Medical Institutions, 1830 E. Monument St., Ste. 7300, Baltimore, MD 21205, fax 410/614-8601, mstreif{at}jhmi.edu
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionConclusionsReferences |
|---|
OBJECTIVE: To evaluate the clinical benefits of an inpatient AMS to cardiac surgery patients requiring warfarin anticoagulation therapy.
METHODS: After obtaining institutional review board approval, we
conducted a retrospective, single-center, cohort study of consecutive cardiac
surgery patients treated before (January 2003–May 2005) and after
(June–December 2005) establishment of an inpatient AMS. Demographic and
clinical characteristics as well as laboratory and clinical data were
retrieved from institutional electronic databases and compared between the 2
patient cohorts. Comparisons between study groups were conducted using a
2 or Fisher's Exact test for categorical variables and
a Student's t-test for continuous variables. Analysis of rare event
data was conducted using Poisson regression analysis.
RESULTS: Of 1919 patients admitted during the study interval, 826 received warfarin (674 pre-AMS, 152 post-AMS). The number of patients with postsurgical panic international normalized ratio (INR) values declined after initiation of the AMS (pre-AMS 90/674 [13.4%] vs post-AMS 11/152 [7.2%]; p = 0.036). There was a trend toward fewer clinically significant postoperative bleeding events (pre-AMS 21/674 [3.1%] vs post-AMS 2/152 [1.3%]; p = 0.22) and fewer repeat surgeries for late postoperative bleeding (pre-AMS 8/674 [1.2%] vs post-AMS 0/152 [0%]; p = 0.08). AMS intervention was associated with a 17% decrease in the average postsurgical length of stay (13.9 days vs 11.6 days; p = 0.015).
CONCLUSIONS: A multidisciplinary AMS can improve anticoagulation management, leading to fewer panic INR values and a reduced length of hospital stay.
Key Words: anticoagulation, cardiac surgery, warfarin
Published Online, May 6, 2008. www.theannals.com, DOI 10.1345/aph.1L027
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
To standardize anticoagulation management, the AMS partnered with the faculty and staff of the Cardiac Surgery service and Critical Care Medicine to develop an antithrombotic management protocol that would standardize anticoagulant start times for treatment and prophylaxis, dosing and monitoring strategies, therapeutic targets, and duration of therapy. Using this protocol as a guide, a pharmacist and a physician from the AMS (staffing level, 1.5 full-time equivalent units) began to provide daily consultation regarding anticoagulation management and venous thromboembolism (VTE) prophylaxis Monday through Friday for all inpatients on the Cardiac Surgery service. The AMS staff reviewed patient laboratory, clinical, and pharmacy data and maintained a flow sheet for each patient that contained patient demographics, baseline weight, pertinent medical history, active/past medications, transfused blood products, and laboratory/radiographic results (Appendix I, www.hwbooks.com/pdf/appendices/L027.pdf). This information allowed the AMS to identify potential drug–drug interactions/combinations that might influence anticoagulation or place patients at higher risk for bleeding. Laboratory results and clinical signs/symptoms were reviewed to identify patients who might be bleeding or developing medication-induced cytopenias, including heparin-induced thrombocytopenia. The flow sheet was also used to provide advice on warfarin dose selection and monitoring. When a patient's antithrombotic management appeared to deviate from the protocol, the AMS staff contacted individual patient care providers (either nurse practitioners or attending physicians) to recommend changes in management. In addition, the AMS facilitated the early discharge of patients on low-molecular-weight heparin and warfarin by identifying appropriate candidates for outpatient treatment and providing telephone and/or outpatient clinic follow-up to manage warfarin and low-molecular-weight heparin therapy.
ANTITHROMBOTIC PROTOCOL
Warfarin 2.5 mg daily was started between postoperative days 1 and 3, once
the patient was taking oral medications. In patients who were taking warfarin
prior to surgery, we recommended that the initial postoperative dose of
warfarin be 50% lower to compensate for initially reduced postoperative oral
intake, use of potentially interacting medications, and reduced hepatic
function. Subsequent doses were recommended based on daily international
normalized ratio (INR) values. Dose reductions were recommended for any
patient whose INR value increased more than 0.4 INR units in a 24-hour period.
Empiric warfarin dose reductions were recommended on a case-by-case basis when
drugs known to interact with warfarin were initiated. All patients received
risk-adjusted VTE prophylaxis. Mechanical prophylaxis consisting of T.E.D.
compression stockings (Tyco Healthcare, Princeton, NJ), and sequential
compression devices (Tyco Healthcare) were placed upon patient arrival in the
cardiac surgery intensive care unit. In patients without active bleeding or
platelet counts less than 100 x 103/µL,
unfractionated heparin prophylaxis was initiated 24 hours postoperatively at a
dose of 5000 units administered subcutaneously every 12 hours (body weight
100 kg) or 5000 units administered subcutaneously every 8 hours (body
weight >100 kg). Pharmacologic VTE prophylaxis was discontinued upon
initiation of intravenous heparin infusions or if the INR value exceeded
1.4.
PROGRAM ANALYSIS
To evaluate the effectiveness of the AMS, we obtained approval from the
Johns Hopkins Institutional Review Board to conduct a retrospective cohort
study of consecutive adults (aged 
18 y) who underwent coronary artery
bypass grafting (CABG) and/or cardiac valve surgery (replacements, repairs, or
both) and were discharged on warfarin between January 1, 2003, and December
31, 2005. Two authors (PJB-H, TRU) identified patients who received warfarin
while hospitalized by reviewing the electronic patient record (EPR). Patients
undergoing aneurysm repairs or heart or lung transplantation were excluded
because their postoperative course and length of stay are driven by
complications unrelated to anticoagulation. All patients discharged between
January 1, 2003, and May 31, 2005 (AMS launched on June 1, 2005) were included
in the pre-AMS patient cohort, and all patients discharged afterward (June 1,
2005–December 31, 2005) were considered part of the post-AMS patient
cohort. Demographic, clinical, and laboratory data were retrieved (by PJB-H
and TRU) from the EPR, a clinical and administrative database of all cardiac
surgery patients, maintained by the Cardiac Surgery department and an
institutional administrative database. To assess the impact of the AMS, we
examined the following set of outcome measures:
Using the EPR, the number of patients with at least one panic INR value, patients receiving vitamin K, and episodes of bleeding and thromboembolism were collected from the operative date until hospital discharge. Bleeding episodes were determined to be clinically significant if they were mentioned in the EPR and required intervention (surgery, transfusion, change in medical therapy). Bleeding events were classified as either a return to the operating room within 24 hours of the initial procedure, a return to the operating room after 24 hours of the initial procedure, a surgical site hematoma not requiring additional major surgery, or a nonsurgical site bleed (eg, gastrointestinal bleed). The number of patients who returned to the operating room within 24 hours of the index surgery was recorded, but those patients were excluded from analysis because the bleeding events were related to the surgical procedure itself rather than to anticoagulation. Thromboembolic events required objective radiologic documentation to fulfill study criteria and were classified as a lower-extremity deep vein thrombosis, a pulmonary embolism, or a stroke that occurred 48 hours or more after surgery.
STATISTICAL ANALYSIS
Descriptive statistics were used to summarize all outcome measures. A
sensitivity analysis was performed on the postoperative length of stay data.
In this analysis, patient information beyond day 25 of admission was censored
(ie, a length of stay that exceeded 25 days was recorded as 25 days). This
analysis was conducted to eliminate the impact of a prolonged postoperative
length of stay that was unlikely to be related to perioperative
anticoagulation.8
Comparisons of outcome measures between the study groups were conducted
using a 2 or Fisher's Exact test for categorical
variables and a Student's t-test for continuous variables. When
non-normal distributions were encountered, a log transformation was applied
and the Student's t-test was conducted on the log transformed data.
In such instances, a Wilcoxon rank sum test was also used as an additional
verification procedure. Analysis of event count data (ie, the number of
patients who experienced a postsurgical bleed requiring a return to the
operating room) was conducted using Poisson regression analysis. For all tests
of statistical significance, a p value less than 0.05 was considered
statistically significant. Analyses were performed using STATA, version 9.0
(Stata Corp., College Station, TX).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
|
Clinical outcomes are displayed in Table 2. Postsurgical panic INR values before the advent of the AMS occurred in 13.4% of the patients compared with 7.2% after the AMS was launched. Although not statistically significant, a substantial difference was noted in the number of patients taken back to the operating room 24 hours or more after their initial procedure due to clinically significant bleeding. Of the 8 patients who returned to the operating room after 24 hours, 6 had valve replacement surgery and 2 had coronary artery bypass surgery. The median time to repeat surgery in this group was 6.5 days (range 4–13 days). The number of clinically significant postsurgical bleeding events (hematoma, gastrointestinal bleeds) that did not require repeat surgery also displayed a downward trend. However, postsurgical thromboembolic events were unaffected by AMS involvement, and the number of patients receiving a dose of vitamin K remained unchanged. Vitamin K was administered principally for treatment of panic INR values, although it was also administered to transiently correct INR values prior to invasive procedures. Institution of the AMS was associated with a 17% decrease in the average postsurgical length of stay. The sensitivity analysis, excluding patient information beyond 25 days postprocedure, also showed a statistically significant 9% decrease in the average length of stay between the pre- and post-AMS study groups (pre-AMS 11.7 days vs post-AMS 10.6 days; p = 0.021).
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
Several studies have documented the impact of physician- or pharmacist-managed anticoagulation consultation services on inpatient care.8,9,13-15 However, only one study has demonstrated that an inpatient AMS can improve patient outcomes and reduce length of stay. Dager et al.9 documented a significant reduction in the number of patients and patient-days with INR values greater than 6.0 and in the length of stay between patients receiving usual care and care given by an inpatient-focused, pharmacist-run AMS. The benefits identified in other trials, including a study by Tschol et al.13 that assessed physician versus pharmacist management of warfarin in cardiac valve surgery patients, were limited to qualitative improvements in anticoagulation management.13,14 A recent review of pharmacy-based AMS programs concluded that additional studies are necessary to establish the benefits of these specialized services.15 We believe that our study provides important additional evidence supporting the benefit of these programs, which we hope will encourage hospital and pharmacy administrations to create these specialized services.
Compared with previous programs, our AMS is innovative in several aspects. We have a multidisciplinary staff consisting of a pharmacist and a physician as well as a rotating group of pharmacy residents. The complementary knowledge bases and skill sets of our team members allow us to comprehensively review the factors that may influence a patient's anticoagulation therapy. We feel that this multidisciplinary approach builds and expands upon the care models used by previous AMS programs and may serve as a model for future programs. Unlike efforts previously described in the literature, we partnered with the faculty and staff who were currently caring for the target patient population to ensure that the antithrombotic protocol that we designed would enhance and standardize rather than conflict with current clinical practices. This approach has been instrumental in our early progress and success.
While the results thus far have been encouraging, we have been unable to achieve all of our stated program goals, one of which is to reduce the number of episodes of clinically significant thromboembolism. We believe that our inability to reduce event rates is due in part to the infrequent nature of these events in the cardiac surgery population. In addition, the number of thromboembolic events can be influenced not only by the adequacy of anticoagulation and deep vein thrombosis prophylaxis but also by the intensity of physician screening practices. We anticipate that implementation of risk screening measures and order sets for prophylaxis of deep vein thrombosis will help to improve these outcomes.
Our study also has several limitations, the foremost being its retrospective nature and disproportionate group sizes. We cannot exclude the possibility that our results reflect the impact of unidentified factors that may have changed during the study period. Although the demographic characteristics of the patient population were unchanged, there was a significant change in the type of procedures performed during the study period.
However, several factors suggest that this potential confounding factor is not responsible for the observed benefits. First, the faculty and staff of the Cardiac Surgery service have remained relatively stable throughout the study period. Therefore, significant changes in personnel and practice style outside of our intervention are unlikely to have taken place. Second, the observed procedural shift in the patient population is likely to have negatively—rather than positively—impacted the clinical and economic outcomes associated with our service, given that cardiac valve replacement procedures tend to be associated with longer and more complicated hospital stays.16-18 Third, the observed reduction in the number of patients with panic INR values or postoperative bleeding is consistent and biologically plausible. Therefore, we believe that the outcomes demonstrated by our program should prove sustainable and reproducible when subjected to more rigorous study.
An additional limitation is the relatively small size of the group receiving care from the AMS, compared with the pre-AMS group. We included more patients in the pre-AMS group to increase the precision of the estimates of study outcomes prior to the AMS. The relatively diminutive size of the post-AMS group limited our ability to demonstrate statistically significant differences between the study groups in several outcome measures. However, we felt that it was important to conduct an analysis of the initial experience of the AMS to justify continuing the service, and statistical significance was attained despite the limited sample on several study measures. We plan to conduct a follow-up analysis when the AMS has been in operation for 30 months.
Over the past decade, numerous advances in therapeutics and technology have produced tremendous benefits for patients and challenges for caregivers. The increasing complexity of the inpatient treatment environment has increased the risks of iatrogenic harm due to medication mismanagement. Recognizing these potential risks, hospitals, hospital accrediting agencies, and national patient safety groups have responded by developing management guidelines. Clinical pharmacy specialists can play a key role in ensuring safe medication utilization. Our experience demonstrates that combining the strengths of pharmacists with those of other healthcare professionals and enlisting the faculty and staff of specific hospital units can result in a very productive, cooperative partnership that can lead to improved patient outcomes and shorter hospital stays. This model could be easily translated to other specialized hospital units and to other institutions. Given the demonstrated clinical benefits, this feasible approach is an attractive investment for hospital administrations and should pay generous dividends in improved patient safety and use of scarce inpatient bed space.
|
Conclusions |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
65 years old. Eur J Cardiothor Surg1997;11:1133-40.[Abstract]This article has been cited by other articles:
|
|
 |
|
  A. Airee, A. B Guirguis, and R. A Mohammad Clinical Outcomes and Pharmacists' Acceptance of a Community Hospital's Anticoagulation Management Service Utilizing Decentralized Clinical Staff Pharmacists Ann. Pharmacother., April 1, 2009; 43(4): 621 - 628. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
 |
 |
 |
 |
