Addressing the Use of Direct Anticoagulants in Antiphospholipid Syndrome - Quality Improvement Initiative

Background The vitamin K antagonist, warfarin, is the preferred anticoagulant in patients with antiphospholipid syndrome (APS). Direct oral anticoagulant (DOAC) use in APS is associated with inferior outcomes, including significantly increased odds of arterial thrombotic events, particularly cerebrovascular accidents (CVA). Despite growing evidence supporting warfarin as the preferred first-line agent, DOACs continue to be utilized by providers in the management of APS. The primary aim of this analysis was to scrutinize the patients with a diagnosis of APS and their anticoagulant choices, to identify quality gaps such as unconfirmed diagnosis (with the need for further testing), lack of awareness of the evidence by the prescribing provider or inadequate documentation of risk benefit conversation, versus patient preference and then systematically intervene to reduce the utilization of DOACS in this patient population. Methods We retrospectively reviewed 100 patients with an ICD code diagnosis of APS documented in the electronic health record (EHR) who were recently prescribed a DOAC within Mayo Clinic enterprise. These cases were assessed for appropriateness of the diagnosis (defined as having at least one clinical manifestation of APS, including venous/arterial thrombosis and/or adverse pregnancy outcome in the setting of persistent antiphospholipid antibody (aPL) positivity by Sapporo criteria). aPL positivity was defined as having positive tests on two occasions with a minimum of 12 weeks apart. Antiphospholipid antibody testing included anticardiolipin (aCL) antibodies (IgG and IgM), anti-beta2 glycoprotein I antibodies (IgG and IgM) and lupus anticoagulant (LA). If the patient met diagnostic criteria, they were then risk stratified to identify high-risk cases. High-risk profile was defined as persistently positive levels of all three aPL (triple-positive). Following review for the appropriate diagnosis, intervention gaps were identified, including the need for additional diagnostic testing, provider education, documenting risk-benefits discussion, and removal of APS from the patient's problem list in the EHR. Clinical intervention was done by sending an EHR message to the prescribing provider which included reason for communication, intervention and recommendation. Reminders were sent every 2 weeks to providers who did not respond, with maximum of 3 messages per provider. Results Of the 100 patients, 60/100 (60%) did not meet criteria for APS, as either initial or repeat aPL testing was negative, and therefore were excluded from this project. Within the remaining 40 patients, 17/100 (17%) had thrombotic APS and 23/100 (23%) had unconfirmed diagnosis Reasons for no intervention (20/40) included not meeting criteria for APS, no longer on a DOAC, and risk versus benefits discussion documented. After adjudication only 20 patients qualified for intervention, 10/20 (50%) patients required additional testing due to lack of repeat testing or referral to a specialist, 7/20 (35%) required removal of APS from problem list, and 3/20 (15%) needed provider. Only 12/20 (65%) of providers accepted the recommendation, 7/20 (35%) did not, and 1/20 (5%) did not respond. Conclusions Despite syndrome-defining criteria, APS remains a commonly misdiagnosed condition. Our QI analysis found that more than half of patients in our cohort with HER documented diagnosis of APS did not meet the Sapporo diagnostic criteria. The tested intervention was time-intensive and prospective, which may have contributed to limited provider acceptance. We propose a robust system-based intervention to remedy this quality issue, which will leverage the EHR to create smart tools in the form of order sets and a passive Our Best Practice Advisory alert to highlight incomplete testing or suggest follow up lab testing at 12 weeks. We will also integrate just in time provider education resources and material embedded in the orders. Our future goal is to assess the impact of this intervention with a similarly sampled cohort.