Abstract
Introduction
Thrombophilia diagnostics are frequently ordered in the inpatient hospital setting, but their impact on patient care is often equivocal. Thrombophilia testing is expensive, and many results are subject to confounding when ordered in the context of an acute hospitalization. Furthermore, these tests are frequently lost to follow-up or wastefully repeated after the patient is discharged. In this study, we conducted a retrospective chart review to determine the rate and financial impact of inappropriate thrombophilia test ordering across all inpatient services at Stanford Hospital over one calendar year.
Methods
Utilizing data from our finance department, we obtained a list of all inpatient thrombophilia testing ordered at Stanford Hospital from June 2014 through June 2015. Thrombophilia testing was defined as ordering any of the following: factor V Leiden, prothrombin G20210A mutation, antithrombin III, lupus anticoagulant, beta-2 glycoprotein 1 IgM/IgG, anticardiolipin IgM/IgG, dilute Russell viper venom time, protein C or protein S levels, and JAK2 V167F mutation.
The criteria for defining a test as 'inappropriate' were guided by utilizing major society guidelines and current evidence, placing an emphasis upon the ordered tests' clinical relevance and reliability in the context of the patient's admission diagnosis. The criteria were formulated by a senior hematologist with specific expertise in thrombophilia evaluations.
Two internal medicine resident physician data reviewers independently evaluated the ordered tests to determine their appropriateness. To ensure consistency between reviewers, identical test datasets were evaluated and compared, demonstrating satisfactory concordance (>0.85). When the appropriateness of a test was unclear, joint evaluation was performed with the entirety of the study team to arrive at a final conclusion. Each test was linked to the ordering primary service. Charge data for each individual test was obtained through our financial department. Aggregate data were evaluated manually.
Results
In total, we reviewed 889 individual orders involving 167 patients across 20 ordering specialties. Of the 889 total orders, 331 were deemed inappropriate (37.2%), translating into a cumulative hospital charge of $152,923 (Figure 1). The tests most frequently inappropriately ordered included antithrombin III (94.4%), factor V Leiden (93.2%), protein C (92.7%), protein S (92.2%), and the prothrombin G20210A mutation (89.3%). Ordering individual tests in the setting of clearly provoked thrombotic events, during the acute thrombotic period, while patients were on concurrent anticoagulation, or when results failed to impact management represented the most common reasons testing was deemed inappropriate.
Ordering practices were then stratified across the hospital's different primary services. Of services with the highest volume of test ordering, General Medicine (38.1%) and Neurology (34.9%) ordered testing inappropriately at the highest rates, while Rheumatology (12.8%) and Hematology (15.9%) ordered inappropriately at the lowest rates. Notably, the non-teaching services ordered testing inappropriately at one of the highest rates (62.2%), though their volume of ordering was lower in comparison with the aforementioned groups.
Discussion
Our results illustrate the high prevalence and significant financial impact of inappropriate or unnecessary thrombophilia testing conducted in the inpatient setting at our institution. Factors confounding test validity were frequently present at the time of ordering. Furthermore, stratifying ordering practices by specialty illustrated the differential rates of inappropriate ordering between services.
Even when thrombophilia testing results fail to impact short term decision-making, misappropriated labeling of patients as 'thrombophilic' can have a lasting negative impact on future anticoagulation decisions. Combined with the high cost of errant ordering, these serve as a strong impetus to reduce the rate of thrombophilia testing during inpatient hospitalizations. Our baseline data demonstrate a need for institution-wide changes such as implementing electronic best practice advisories or potential ordering restrictions, and of tantamount importance, service-specific educational interventions in order to reduce unnecessary expenditures and improve patient care.
Figure 1 Figure 1.
Disclosures
No relevant conflicts of interest to declare.
