Outpatient Management of Deep Venous Thrombosis: 2 Models of Integrated Care

Objective: To describe and evaluate the impact, effectiveness, and safety of 2 models of outpatient deep venous thrombosis (DVT) management.

Study Design: Observational health-records survey conducted in 2 community hospitals using a comparative design.

Methods: The model centered around the emergency department (ED) was studied from April 15, 1997, through December 31, 1999; the decentralized model was studied from July 15, 2002, through March 15, 2004. The effect of the decentralized model on processes of care was determined by comparing the percentages of patients who were managed by office-based primary care providers under the 2 models. Two-week clinical outcomes also were compared, including symptomatic pulmonary embolism, symptomatic progression of index DVT, new DVT, major bleeding, minor bleeding, and death.

P

Results: During the study periods, 187 and 254 outpatients were enrolled in the ED-centered model and the decentralized model, respectively. Under the ED-centered model, 17 (22.1%) of 77 patients who presented to the primary care providers' offices on weekdays before 1:30 PM (on-hours) were managed exclusively within the clinic setting. Under the decentralized model, 78 (91.8%) of 85 patients who presented on-hours were managed within the clinics—a difference of 69.7 absolute percentage points (95% confidence intervals, 58.7%, 80.7%; < .001). The models had comparable 2-week clinical outcomes.

Conclusions: The decentralized model for outpatient DVT management allowed more primary care clinic patients to receive their initial care exclusively in the clinic setting. There was no significant difference between the models in short-term effectiveness of therapy and patient safety.

(Am J Manag Care. 2006;12:405-410)

Stable patients with uncomplicated, proximal lower-extremity deep venous thrombosis (DVT) are commonly managed as outpatients.^1,2 This shift in the site of care followed the introduction of low-molecular-weight heparins (LMWHs).^3,4 Compared with hospitalization for intravenous unfractionated heparin, outpatient use of LMWHs is equally effective, safe, and less costly.^5-9 Regardless of the site of care, the impact of DVT on quality-of-life measures can be significant.¹⁰ Outpatient DVT management provides more patient satisfaction than inpatient treatment and is associated with higher levels of physical activity, better social functioning, and a more rapid return to premorbid levels of activity.^11,12

Integrating the many necessary steps that contribute to successful outpatient DVT management is not a simple process.^13-15 Multiple elements of care must be efficiently coordinated, including radiology, laboratory, pharmacy, patient education, and (if necessary) home healthcare. Coordination of these services is complex and time consuming for the primary care provider, who often refers such patients to the emergency department (ED), where the integration of time-intensive care can be managed.

Managed care organizations can more easily integrate healthcare services.¹⁶ Our community hospitals have an integrated model of healthcare, and its ability to coordinate multiple services is well demonstrated in the management of outpatients with DVT. In early 1997, we began managing select patients with lower-extremity DVT using an ED-centered model, in which the large majority of outpatients with DVT were referred to the ED for coordination of their initial care. A description of this model, as well as its safety and effectiveness, has been published.¹⁷ We modified the treatment pathway in 2002 by enabling the primary care providers to initiate outpatient DVT management in their offices without an ED visit, depending on the day and time of presentation. This decentralized model allowed patients with DVT to be enrolled in the outpatient care pathway from the office of the primary care provider. This study describes the development of the decentralized model and its impact on processes of care. We also compare the effectiveness and safety of the decentralized model with the effectiveness and safety of the prior ED-centered model for DVT treatment.

METHODS

Setting

This observational, retrospectively defined comparative study was performed in 2 neighboring suburban community hospitals that are part of a large group-model health maintenance organization. The initial ED-centered model was in effect from April 15, 1997, through July 14, 2002; and its effectiveness and safety were studied from April 15, 1997, through December 31, 1999.¹⁷ The decentralized model was instituted July 15, 2002, and was studied from its inception through March 15, 2004. The study was approved by the institutional review board of Kaiser Foundation Research Institute.

Participants

Outpatients were eligible for enrollment if they had an acute, symptomatic, proximal lower-extremity DVT that was documented by compression ultrasonography, as interpreted by a board-certified staff radiologist. Isolated distal DVTs below the popliteal vein were not included in the treatment pathways. Exclusion criteria were described in a prior publication.¹⁷

Emergency Department-centered Model of Outpatient DVT Management

In early 1997, we pioneered an ED-centered outpatient pathway for the treatment of proximal lower-extremity DVT. Patients in the outlying clinics with suspected DVT were referred to the ED, where the radiologic diagnosis was established and laboratory analyses and patient education were undertaken. After discharge from the ED, patients were evaluated daily by home health nurses until the achievement of serial therapeutic international normalized ratios (INRs). Our Integrated Quality Services department monitored the safety and effectiveness of the outpatient pathway. The details of this ED-centered model and its short-term patient outcomes have been reported.¹⁷

Though the ED-centered model was shown to be safe and effective, it had several disadvantages. This model was disruptive to continuity of care, because patients whose DVT was suspected or diagnosed by their primary care provider had to be transferred to the ED for complete evaluation, treatment, and education. It also failed to standardize the education given to the DVT patients. Finally, this model placed a small added burden on the ED to care for nearly all patients with DVT. We designed the decentralized model to address these shortcomings.

Decentralized Model of Outpatient DVT Management

The 2 greatest impediments to having office-based primary care providers manage outpatient DVT cases were (1) their access to radiology and laboratory and (2) the extensive time requirements for patient education. To address the first need, primary care providers' patients were given priority access to urgent lower-extremity compression ultrasound exams, allowing the patients to go from the office to the ultrasound suite and return with an official reading in a timely fashion. Similar priority access was given for laboratory services. As a result, patients could obtain an urgent ultrasound exam and laboratory studies, allowing their primary care provider to initiate appropriate outpatient DVT management that day from the office. Our solution to the second need, that of patient education, is discussed below.

Patients Presenting to the Office Before 1:30 ^PM on

Weekdays.

Once DVT was suspected, these patients were sent for an urgent sonographic evaluation and immediate laboratory evaluation. The ultrasound and laboratory findings were reviewed by the clinician to determine the patient's eligibility for the outpatient DVT pathway. Eligible patients would receive their first subcutaneous injection of enoxaparin (1.5 mg/kg up to 150 mg daily, beyond which the dosing was changed to 1 mg/kg twice daily). The first dose of warfarin (7.5 mg) also was administered. Patients were given a brief informational handout, as well as a follow-up appointment in 5 days with their primary care provider and a next-day outpatient appointment with a nurse educator.

Patients Presenting to the Office After 1:30 ^PM on

Weekdays, and Anytime on Weekends.

If a patient with a probable lower-extremity DVT presented to the primary care provider's office during these hours, the clinician would order the ultrasound exam and laboratory evaluation as described above. However, because time constraints wouldn't allow the primary care provider to complete the initial assessment and treatment, the patient with a DVT would be directed by the radiology department to the ED for further management. Patients with probable DVT seen in nonprimary care clinics (eg, orthopedics, rheumatology, gynecology) also were referred to the ED after their radiologic diagnosis.

Appointment With Nurse Educator

Patients with DVTs then were given an appointment with the outpatient nurse educator on the morning after their index visit. This expedited next-day follow-up (available 365 days a year) allowed for standardized patient education before the second LMWH dose at 24 hours was indicated.^18-20 If patients were unwilling or unable to self-administer the subcutaneous injection, they were asked to bring someone with them to the education appointment who was able to do so. During this appointment with the nurse, patients were taken through a well-defined course curriculum, which included an educational video on DVT and enoxaparin administration. The following issues were addressed: the reasons for anticoagulation therapy; the importance of compliance; the need for monitoring; the process of warfarin dosage adjustment; potential side effects, complications, and risks of anticoagulation therapy; potential drug-drug and drug-food interactions; the use of over-the-counter medications (particularly those containing aspirin products); the storage and handling of enoxaparin and warfarin; new prescriptions; and refill requests. Compression stockings were prescribed to help prevent postphlebitic syndrome.

Patients also received prescriptions for their subsequent doses of both enoxaparin and warfarin. After the initial warfarin dose, 5 mg daily was begun on the day after the diagnostic visit. Subcutaneous enoxaparin was continued until the INR was maintained at a therapeutic level (2.0-3.0) for 2 consecutive days. The serial INR results were monitored by our anticoagulation clinic, which contacted the patient by telephone to adjust the warfarin dose as indicated and to assess for complications.

Home health visits were arranged to provide education and therapy for home-bound patients who were unable to return to the hospital for their next-day follow-up with a nurse educator.

Study Subject Identification and Data Abstraction

International Classification of Diseases, Ninth Revision

The method of data collection for the ED-centered model was published previously.¹⁷ All patients with DVT during the second portion of the study period (July 15, 2002, to March 15, 2004) were identified by using codes. Clinical data were obtained by the 4 investigators from an explicit review of each patient's medical record, in conjunction with the organization's comprehensive computerized database. Charts of patients with acute DVT who had been admitted to the hospital were reviewed for the presence of criteria that excluded them from outpatient management. Those DVT patients who were enrolled in the outpatient DVT pathway under the decentralized model of care were identified by using a database managed by the Integrated Quality Services department, which has tracked all outpatient DVT cases since 1997. All abstractors received training on the content and coding of each data element, procedures for data handling and data transmission, and protocols to handle possible questions or problems during the study. A structured data-abstraction tool was used.

Clinical Variables

Clinical variables were abstracted for enrollees, including demographics (age, sex); the site, day of week, and time of day of the initial assessment; medical comorbid conditions; thromboembolic risk factors²¹; and exclusion criteria.

Impact of the Decentralized Model on Processes of Care

The impact of the new model was determined by evaluating processes of care, including where the patients first presented (primary care clinic, specialty clinic, or ED), what day of the week and time of day they presented, and whether they were transferred to the ED for continuation of care. If the patient was transferred from the primary care provider's office to the ED, the primary reason for transfer was determined (off-hour, comorbidity, or unknown).

Outcome Measures

Major

minor

The outcome measures of this study addressed failure of therapy (effectiveness) and complications of therapy (safety) within 2 weeks of enrollment, which included progression of thromboembolic disease (symptomatic pulmonary embolism, symptomatic progression of index DVT, or new DVT), bleeding of any kind, and death. Pulmonary embolism was detected by using standard diagnostic criteria with ventilation/perfusion imaging or spiral computed tomography, and pulmonary arteriography if indicated. Symptomatic progression of index DVT was detected clinically by the evaluating physician in the presence of increasing pain or swelling in the index leg. Confirmatory ultrasonography was not required, as this measure focused on the progression of symptoms. New contralateral or upper-extremity DVT subsequent to the index visit was detected by compression ultrasonography. Bleeding was detected only if clinically overt. bleeding was diagnosed in the following circumstances: a fall in the hemoglobin level of at least 2.0 g/dL or a transfusion of packed red blood cells, retroperitoneal or intracranial bleeding, or bleeding that warranted the permanent discontinuation of treatment. ²² Bleeding that did not meet the aforementioned criteria was considered . Two-week outcomes with the decentralized model were compared with 2-week outcomes with the ED-centered model.

Statistical Analysis

P

Chi-square tests were used to evaluate the differences between the 2 models of care. A value of less than .05 was considered to indicate statistical significance. The normal approximation to the binomial distribution was used to calculate the confidence intervals. Analyses were performed with the Statistical Package for the Social Sciences, version 9 (SPSS, Inc, Chicago, Ill).

RESULTS

During the study periods, 187 and 254 patients with acute lower-extremity DVT were enrolled in the ED-centered model and the decentralized model, respectively, of the outpatient pathway. The clinical characteristics of the patients treated under each model are described in Table 1. During the study period for the decentralized model, 189 (42.7% of a total 443) patients with DVT were admitted to the hospital for initiation of therapy. The reasons for their exclusion from the outpatient pathway are described in Table 2.

P

The impact of the decentralized model on processes of care is quantified in Table 3. Under the ED-centered model, 17 (22.1%) of 77 patients who presented to the primary care providers' offices on weekdays before 1:30PM (on-hours) were managed within the clinic setting, whereas under the decentralized model, 78 (91.8%) of 85 patients who presented on-hours were managed without recourse to the ED, a difference of 69.7 absolute percentage points (95% confidence intervals, 58.7%, 80.7%; < .001).

Of the 254 patients enrolled in the decentralized model, 56 (22.0%) were assigned to home health services (Table 3). Short-term outcomes under the 2 models were comparable (see Table 4).

DISCUSSION

In this paper, we describe 2 models of care for the outpatient management of patients with lower-extremity DVT: the ED-centered model, in which nearly all patients were transferred to the ED, and the decentralized model, in which patients who presented before 1:30 ^PM on weekdays were managed without a transfer to the ED. Our study found that the decentralized model was feasible and easily implemented by primary care providers in their offices. When patients with DVT presented on weekdays before 1:30 ^PM, more than 90% were managed by the primary care providers without an ED visit, suggesting successful implementation. There was a 4-fold increase in the number of DVT patients managed fully in the primary care offices in the decentralized model compared with the ED-centered model (92% vs 22%). The comparable incidences of progression of disease and adverse outcomes testify to the effectiveness and safety of the decentralized approach, and compare favorably with incidences reported in other studies that have published 2-week outcomes of DVT management.¹⁷

Our prior ED-centered model of care for outpatients with acute lower-extremity DVT was modified to a decentralized model to allow the primary care providers to retain management of their own patients with DVT by not having to refer them to the ED. Management of these patients by their primary care providers has many advantages. Primary care management fosters continuity of care, improves the physician-patient relationship, is more convenient for the patient, is more cost-effective, and promotes patient satisfaction.²³ This model also standardized patient education by referring all DVT patients to a trained nurse educator who presented a well-defined course curriculum. Decentralizing DVT management also helped to reduce the burden on the ED, which in recent years has been increasingly subject to overcrowding.^24,25 This study demonstrates that shifting some of the responsibility for initial DVT management back to the primary care provider can be done without reducing the effectiveness or the safety of care.

This study has several shortcomings. Inherent in the retrospective study design are limitations of data collection. We had the advantage, however, of gathering data from a well-conducted quality improvement program. Also, the study's applicability may be limited to similar managed care organizations that are able to integrate multiple specialties and services.

In sum, we describe 2 models of outpatient DVT management facilitated by an integrated healthcare system. The ED-centered model and the decentralized model were shown to have comparable short-term outcomes with respect to effectiveness and safety. The newer, decentralized model has the advantage of allowing the office-based primary care provider to retain the responsibility for initiating the treatment pathway during regular working hours, an aspect of care strongly associated with higher patient satisfaction, fewer ED referrals, and fewer inpatient admissions.

Acknowledgments

We thank the members of the DVT Committee for their help in the creation of both pathways and the careful monitoring of their safety and impact. We also thank Kennie Stevenson-Baker from Integrated Quality Services for her conscientious assistance with data acquisition. Scott Gordinier, PhD, generously served as our statistical consultant.

From the Department of Emergency Medicine, The Permanente Medical Group, Kaiser Permanente Medical Centers, Sacramento and Roseville, Calif (DRV, DAB, PBP) and the University of California, Irvine, Calif (DOH).

Financial support for this study was provided by a community benefits grant (CN-03DVins-02-H) from Northern California Kaiser Permanente.

Address correspondence to: David A. Berman, DO, Department of Emergency Medicine, Kaiser Permanente Medical Center, 1600 Eureka Rd, Roseville, CA 95661. Email: david.a.berman@kp.org.

Download PDF: Outpatient Management of Deep Venous Thrombosis: 2 Models of Integrated Care