Practice and Correlates of Cancer Associated Thrombosis: The Impact of Socioeconomic Status

Introduction: Venous thromboembolism (VTE) is a leading cause of death among cancer patients receiving chemotherapy. Many patients receive non-guideline concordant care for cancer-associated thrombosis (CAT). The factors associated with anticoagulant selection for CAT are not well defined. We hypothesized that three months after an episode of VTE, patients with lower socioeconomic status would be more likely to receive warfarin, compared to the low molecular weight heparin (LMWH) suggested by guidelines during the study period.

Methods: We used Optum© Clinformatics® Datamart, a large, de-identified claims database, to identify patients with active cancer over 18 years of age who had at least twelve months of follow-up data with at least one ICD-9-CM code for a diagnosis of cancer between 2007-2015, with a simultaneous or subsequent cancer directed- treatment identified by Healthcare Common Procedure Coding System Codes (HCPCS) code or National Drug Code (NDC) from inpatient or outpatient claims data. This cohort was then limited to patients who later had an ICD-9-CM code for VTE, without a claim for VTE in the preceding 12 months and who were newly started on an outpatient anticoagulant (warfarin, LMWH, or a direct oral anticoagulant, DOAC) by NDC codes. Patients were censored if they had a gap in anticoagulation supply of ≥30 days (for LMWH or DOACs) or ≥60 days for warfarin, without an INR measured at least every 42 days.

The primary outcome was the most recent anticoagulant prescription three months after the VTE event. We analyzed the data using multinomial logistic regression with patient- and prescription-level covariates including demographics, recent hospitalization, comorbidities, concomitant medications, type of thrombosis and cancer, education, household income, region, and insurance type. We also compared the co-payments of the most recent anticoagulant prescription filled.

Results: A total of 12,622 patients met the inclusion criteria, 1,485 (12%) on LMWH, 1,546 (12%) on DOACs, and 9,591 (76%) on warfarin. The most common cancer groups represented were breast (19%), genitourinary (17%), and gastrointestinal (14%). The type of CAT was most often lower extremity deep vein thrombosis (DVT, 22%), DVT and pulmonary embolism (PE, 22%), and isolated PE (16%). Approximately 17% of patients were hospitalized within the 28 days prior to the episode of CAT.

Regression analysis showed that, when controlling for other factors, patients with neurologic (p<0.001), gastrointestinal (p=0.002), gynecologic (p=0.017), or lung cancers (p=0.04) were more likely to receive LMWH while patients with breast (p<0.001) and genitourinary (p<0.001) cancers were more likely to receive warfarin. Patients with isolated lower extremity DVT were less likely to receive LMWH (p=0.005). Comorbidities, as reflected by the Charlson comorbidity index, were not significantly associated with which anticoagulant a patient received. Geographic variation was also evident. Compared to the Central Northeast US, patients in the Central Southeast (p<0.001) and Mountain regions (p<0.001) were less likely to receive LMWH, while the Mid-Atlantic (p<0.001) and New England (p<0.001) regions were more likely to receive LMWH. Patients with a Bachelor's degree or greater education had a 22% greater probability of receiving LMWH (p=0.004) and a 22% (p=0.02) greater probability of receiving a DOAC compared to having less than a Bachelor's degree. Additionally, having a household income >$100,000 was associated with a significantly higher probability of receiving LMWH than for all income levels <$60,000 (p=0.002-<0.001). Average co-pays were lowest for warfarin, followed by the DOACs, and then LMWH (mean $8 for warfarin, $45 for DOACs, $76 for LMWH, median $6 for warfarin, $35 for DOACs, and $25 for LMWH, standard deviation $8 for warfarin, $55 for DOACs, and $230 for LMWH, p<0.001).

Conclusions: Patients of lower socioeconomic status who developed CAT, as reflected by a lower household income or lower education, were less likely to receive LMWH than warfarin at three months after their VTE event. Significant regional variation in CAT management was also evident. CAT is associated with significant morbidity. Further attention to the impact of socioeconomic variables on prescribing practices is needed to ensure equitable and optimal anticoagulant access for all CAT patients.