Racial Disparity in the Prescription of Anticoagulants and Risk of Stroke and Bleeding in Atrial Fibrillation Patients.

BACKGROUND: Oral anticoagulant (OAC) therapy is proven to be effective at reducing risk of stroke in patients with atrial fibrillation (AF). However, racial minorities with AF are less likely to be prescribed vitamin K anticoagulants (VKA). There is little information on the racial disparity in the prescription of the non-vitamin K oral anticoagulants (NOACs) and the associated risks of stroke and bleeding. METHODS: We used data from the Northwestern Medicine Enterprise Data Warehouse - a joint initiative across 11 Northwestern Medicine affiliated healthcare centers within metropolitan Chicago, Illinois. Newly diagnosed AF patients between Jan, 2011 and Dec, 2017 with CHA2DS2VASc (congestive heart failure, hypertension, age ≥75 years, diabetes, stroke/transient ischemic attack, vascular disease, age 65 to 74 years, female sex) score of 2 or more and no prior history of stroke or major bleeding were eligible. Logistic regression was used to examine differences in the prescription of any OAC and NOACs by race. Racial differences in the associations of NOACs prescription with incident stroke (a composite of ischemic and hemorrhagic stroke and cerebral embolism) and major bleeding were evaluated using Cox regression. RESULTS: Among 11,575 newly diagnosed AF patients with CHA2DS2VASc score of 2 or more, 48.7% (47.8-49.6) were on any OAC and among those 40.1% (38.8.3-41.4) received any NOACs. After adjusting for age, gender, income, insurance status, and stroke risk factors, the odds of receiving any OAC was .69 (95% CI: .58-.83) in blacks, .74 (.53-1.903) in Hispanics, and .75 (.58-.95) in Asians compared to whites. Among anticoagulated patients, blacks and Hispanics had significantly lower odds of receiving NOACs: .72 (.53-.97) and .53 (.29-.99), respectively. Use of NOACs, as compared to VKAs, was associated with significantly lower risk of stroke [.52(.31-.85)] and bleeding [.72(.54-.95)] in whites but not in non-whites [stroke: .71 (.22-2.31); bleeding .83(.43-1.57)] independent of other risk factors. CONCLUSIONS: Racial minorities with AF who are at risk of stroke were less likely to receive any OAC and NOACs specifically compared to whites even after accounting for insurance status, income, and stroke risk factors. Independent of other risk factors, use of NOACs as compared to VKA was associated with significantly lower risk of stroke and bleeding only in whites but not in non-whites.

Risk Differences in Secondary Prevention Patients Who Present With Acute Coronary Syndrome and Implications of Guideline-Directed Cholesterol Management.

The 2018 American College of Cardiology/American Heart Association cholesterol guidelines for secondary prevention identified a group of "very high risk" (VHR) patients, those with multiple major atherosclerotic cardiovascular disease (ASCVD) events or 1 major ASCVD event with multiple high-risk features. A second group, "high risk" (HR), was defined as patients without any of the risk features in the VHR group. The incidence and relative risk differences of these 2 groups in a nontrial population has not been well characterized. Using the Northwestern Medicine Enterprise Data Warehouse, we compared the incidence of VHR and HR patients as well as their relative risk for cardiovascular morbidity and mortality in a single-center, large, academic, retrospective cohort study. Total 1,483 patients with acute coronary events from January 2014 to December 2016 were risk stratified into VHR and HR groups. International Classification of Diseases versions 9 and 10 were used to assess for composite events of unstable angina pectoris, non-ST elevation myocardial infarction, or ST-elevation myocardial infarction, ischemic stroke, or all-cause death with a median follow-up of 3.3 years. VHR patients were found to have 87 ± 5.4 composite events per 1,000 patient-years compared with HR patients who had 33 ± 5.1 events per 1,000 patient-years (p <0.001). VHR group had increased risk of future events as compared to the HR group (multivariable adjusted hazard ratio 1.66 [1.01 to 2.74], p = 0.047). In conclusion, these results support the stratification of patients into the VHR and HR risk groups for secondary prevention.

Automated Extraction of Grade, Stage, and Quality Information From Transurethral Resection of Bladder Tumor Pathology Reports Using Natural Language Processing.

PURPOSE: Bladder cancer is initially diagnosed and staged with a transurethral resection of bladder tumor (TURBT). Patient survival is dependent on appropriate sampling of layers of the bladder, but pathology reports are dictated as free text, making large-scale data extraction for quality improvement challenging. We sought to automate extraction of stage, grade, and quality information from TURBT pathology reports using natural language processing (NLP). METHODS: Patients undergoing TURBT were retrospectively identified using the Northwestern Enterprise Data Warehouse. An NLP algorithm was then created to extract information from free-text pathology reports and was iteratively improved using a training set of manually reviewed TURBTs. NLP accuracy was then validated using another set of manually reviewed TURBTs, and reliability was calculated using Cohen's κ. RESULTS: Of 3,042 TURBTs identified from 2006 to 2016, 39% were classified as benign, 35% as Ta, 11% as T1, 4% as T2, and 10% as isolated carcinoma in situ. Of 500 randomly selected manually reviewed TURBTs, NLP correctly staged 88% of specimens (κ = 0.82; 95% CI, 0.78 to 0.86). Of 272 manually reviewed T1 tumors, NLP correctly categorized grade in 100% of tumors (κ = 1), correctly categorized if muscularis propria was reported by the pathologist in 98% of tumors (κ = 0.81; 95% CI, 0.62 to 0.99), and correctly categorized if muscularis propria was present or absent in the resection specimen in 82% of tumors (κ = 0.62; 95% CI, 0.55 to 0.73). Discrepancy analysis revealed pathologist notes and deeper resection specimens as frequent reasons for NLP misclassifications. CONCLUSION: We developed an NLP algorithm that demonstrates a high degree of reliability in extracting stage, grade, and presence of muscularis propria from TURBT pathology reports. Future iterations can continue to improve performance, but automated extraction of oncologic information is promising in improving quality and assisting physicians in delivery of care.