Yield of Familial Hypercholesterolemia Genetic and Phenotypic Diagnoses After Electronic Health Record and Genomic Data Screening.

Background Data mining of electronic health records to identify patients suspected of familial hypercholesterolemia (FH) has been limited by absence of both phenotypic and genomic data in the same cohort. Methods and Results Using the Geisinger MyCode Community Health Initiative cohort (n=130 257), we ran 2 screening algorithms (Mayo Clinic [Mayo] and flag, identify, network, deliver [FIND] FH) to determine FH genetic and phenotypic diagnostic yields. With 29 243 excluded by Mayo (for secondary causes of hypercholesterolemia, no lipid value in electronic health records), 52 034 excluded by FIND FH (insufficient data to run the model), and 187 excluded for prior FH diagnosis, a final cohort of 59 729 participants was created. Genetic diagnosis was based on presence of a pathogenic or likely pathogenic variant in FH genes. Charts from 180 variant-negative participants (60 controls, 120 identified by FIND FH and Mayo) were reviewed to calculate Dutch Lipid Clinic Network scores; a score ≥5 defined probable phenotypic FH. Mayo flagged 10 415 subjects; 194 (1.9%) had a pathogenic or likely pathogenic FH variant. FIND FH flagged 573; 34 (5.9%) had a pathogenic or likely pathogenic variant, giving a net yield from both of 197 out of 280 (70%). Confirmation of a phenotypic diagnosis was constrained by lack of electronic health record data on physical findings or family history. Phenotypic FH by chart review was present by Mayo and/or FIND FH in 13 out of 120 versus 2 out of 60 not flagged by either (P<0.09). Conclusions Applying 2 recognized FH screening algorithms to the Geisinger MyCode Community Health Initiative identified 70% of those with a pathogenic or likely pathogenic FH variant. Phenotypic diagnosis was rarely achievable due to missing data.

Developing and Optimizing Innovative Tools to Address Familial Hypercholesterolemia Underdiagnosis: Identification Methods, Patient Activation, and Cascade Testing for Familial Hypercholesterolemia.

BACKGROUND: Familial hypercholesterolemia (FH) is the most common cardiovascular genetic disorder and, if left untreated, is associated with increased risk of premature atherosclerotic cardiovascular disease, the leading cause of preventable death in the United States. Although FH is common, fatal, and treatable, it is underdiagnosed and undertreated due to a lack of systematic methods to identify individuals with FH and limited uptake of cascade testing. METHODS AND RESULTS: This mixed-method, multi-stage study will optimize, test, and implement innovative approaches for both FH identification and cascade testing in 3 aims. To improve identification of individuals with FH, in Aim 1, we will compare and refine automated phenotype-based and genomic approaches to identify individuals likely to have FH. To improve cascade testing uptake for at-risk individuals, in Aim 2, we will use a patient-centered design thinking process to optimize and develop novel, active family communication methods. Using a prospective, observational pragmatic trial, we will assess uptake and effectiveness of each family communication method on cascade testing. Guided by an implementation science framework, in Aim 3, we will develop a comprehensive guide to identify individuals with FH. Using the Conceptual Model for Implementation Research, we will evaluate implementation outcomes including feasibility, acceptability, and perceived sustainability as well as health outcomes related to the optimized methods and tools developed in Aims 1 and 2. CONCLUSIONS: Data generated from this study will address barriers and gaps in care related to underdiagnosis of FH by developing and optimizing tools to improve FH identification and cascade testing.

Effect of Access to Prescribed PCSK9 Inhibitors on Cardiovascular Outcomes.

BACKGROUND: Atherosclerotic cardiovascular disease remains a major cause of death and disability, especially for high-risk familial hypercholesterolemia individuals. PCSK9i (proprotein convertase subtilisin kexin type 9 inhibitors) reduce low-density lipoprotein cholesterol levels and cardiovascular event rates. However, PCSK9i prescriptions are rejected at high rates by payers, and use is often delayed or eventually abandoned as a treatment option. We tested the hypothesis that acute coronary syndromes, coronary interventions, stroke, and cardiac arrest are more prevalent in patients with rejected or abandoned PCSK9i prescriptions than for those with paid PCSK9i prescriptions. METHODS AND RESULTS: We identified 139 036 individuals aged ≥18 years who met the following 3 criteria: prescribed PCSK9i between August 2015 and December 2017, had claims history, and had an established date of exposure for paid, rejected, or abandoned status. To compare the effects of rejected versus paid and abandoned versus paid status, propensity score matching was performed to minimize confounding because of baseline differences in patient groups. Cox regression analyses and incidence density rates for cardiovascular events were estimated on the propensity score-matched cohorts. Patients who received 168 or more days of paid PCSK9i medication within a 12-month period were defined as paid. The hazard ratios for composite cardiovascular events outcome in propensity score-matched analyses were 1.10 (95% CI, 1.01-1.19; P=0.02) for rejected versus paid and 1.12 (95% CI, 1.01-1.24; P=0.03) for abandoned versus paid. In a stricter analysis where paid patients were defined by receiving 338 or more days of therapy within 12-months, hazard ratio was 1.16 (95% CI, 1.02-1.30; P=0.04) for rejected versus paid and 1.21 (95% CI, 1.04-1.38; P=0.03) for the abandoned versus paid status. Higher PCSK9i rejection rates were observed with women, racial minorities, and lower-income groups. CONCLUSIONS: Individuals in the rejected and abandoned cohorts had significantly increased risk of cardiovascular events compared with those in the paid cohort. Rejection, abandonment, and disparities related to PCSK9i prescriptions are related to higher cardiovascular outcome rates.

Finding missed cases of familial hypercholesterolemia in health systems using machine learning.

Familial hypercholesterolemia (FH) is an underdiagnosed dominant genetic condition affecting approximately 0.4% of the population and has up to a 20-fold increased risk of coronary artery disease if untreated. Simple screening strategies have false positive rates greater than 95%. As part of the FH Foundation's FIND FH initiative, we developed a classifier to identify potential FH patients using electronic health record (EHR) data at Stanford Health Care. We trained a random forest classifier using data from known patients (n = 197) and matched non-cases (n = 6590). Our classifier obtained a positive predictive value (PPV) of 0.88 and sensitivity of 0.75 on a held-out test-set. We evaluated the accuracy of the classifier's predictions by chart review of 100 patients at risk of FH not included in the original dataset. The classifier correctly flagged 84% of patients at the highest probability threshold, with decreasing performance as the threshold lowers. In external validation on 466 FH patients (236 with genetically proven FH) and 5000 matched non-cases from the Geisinger Healthcare System our FH classifier achieved a PPV of 0.85. Our EHR-derived FH classifier is effective in finding candidate patients for further FH screening. Such machine learning guided strategies can lead to effective identification of the highest risk patients for enhanced management strategies.

Precision screening for familial hypercholesterolaemia: a machine learning study applied to electronic health encounter data.

BACKGROUND: Cardiovascular outcomes for people with familial hypercholesterolaemia can be improved with diagnosis and medical management. However, 90% of individuals with familial hypercholesterolaemia remain undiagnosed in the USA. We aimed to accelerate early diagnosis and timely intervention for more than 1·3 million undiagnosed individuals with familial hypercholesterolaemia at high risk for early heart attacks and strokes by applying machine learning to large health-care encounter datasets. METHODS: We trained the FIND FH machine learning model using deidentified health-care encounter data, including procedure and diagnostic codes, prescriptions, and laboratory findings, from 939 clinically diagnosed individuals with familial hypercholesterolaemia (395 of whom had a molecular diagnosis) and 83 136 individuals presumed free of familial hypercholesterolaemia, sampled from four US institutions. The model was then applied to a national health-care encounter database (170 million individuals) and an integrated health-care delivery system dataset (174 000 individuals). Individuals used in model training and those evaluated by the model were required to have at least one cardiovascular disease risk factor (eg, hypertension, hypercholesterolaemia, or hyperlipidemia). A Health Insurance Portability and Accountability Act of 1996-compliant programme was developed to allow providers to receive identification of individuals likely to have familial hypercholesterolaemia in their practice. FINDINGS: Using a model with a measured precision (positive predictive value) of 0·85, recall (sensitivity) of 0·45, area under the precision-recall curve of 0·55, and area under the receiver operating characteristic curve of 0·89, we flagged 1 331 759 of 170 416 201 patients in the national database and 866 of 173 733 individuals in the health-care delivery system dataset as likely to have familial hypercholesterolaemia. Familial hypercholesterolaemia experts reviewed a sample of flagged individuals (45 from the national database and 103 from the health-care delivery system dataset) and applied clinical familial hypercholesterolaemia diagnostic criteria. Of those reviewed, 87% (95% Cl 73-100) in the national database and 77% (68-86) in the health-care delivery system dataset were categorised as having a high enough clinical suspicion of familial hypercholesterolaemia to warrant guideline-based clinical evaluation and treatment. INTERPRETATION: The FIND FH model successfully scans large, diverse, and disparate health-care encounter databases to identify individuals with familial hypercholesterolaemia. FUNDING: The FH Foundation funded this study. Support was received from Amgen, Sanofi, and Regeneron.

A nationwide 2010-2012 analysis of U.S. health care utilization in inflammatory bowel diseases.

BACKGROUND: Implementation of the 2010 Affordable Care Act (ACA) calls for a collaborative effort to transform the U.S. health care system toward patient-centered and value-based care. To identify how specialty care can be improved, we mapped current U.S. health care utilization in patients with inflammatory bowel diseases (IBD) using a national insurance claims database. METHODS: We performed a cross-sectional study analyzing U.S. health care utilization in 964,633 patients with IBD between 2010 and 2012 using insurance claims data, including pharmacy and medical claims. Frequency of IBD-related care utilization (medication, tests, and treatments) and their charges were evaluated. Subsequently, outcomes were put into the framework of current U.S. guidelines to identify areas of improvement. RESULTS: A disproportionate usage of aminosalicylates in Crohn's disease (42%), frequent corticosteroid use (46%, with 9% long-term users), and low rates of corticosteroid-sparing drugs (thiopurines 15%; methotrexate 2.7%) were observed. Markers for inflammatory activity, such as C-reactive protein or fecal calprotectin were not commonly used (8.8% and 0.13%, respectively). Although infrequently used (11%), anti-TNF antibody therapy represents a major part of observed IBD charges. CONCLUSIONS: This analysis shows 2010-2012 utilization and medication patterns of IBD health care in the United States and suggests that improvement can be obtained through enhanced guidelines adherence.

Challenges in designing a national surveillance program for inflammatory bowel disease in the United States.

This review describes the history of U.S. government funding for surveillance programs in inflammatory bowel diseases (IBD), provides current estimates of the incidence and prevalence of IBD in the United States, and enumerates a number of challenges faced by current and future IBD surveillance programs. A rationale for expanding the focus of IBD surveillance beyond counts of incidence and prevalence, to provide a greater understanding of the burden of IBD, disease etiology, and pathogenesis, is provided. Lessons learned from other countries are summarized, in addition to potential resources that may be used to optimize a new form of IBD surveillance in the United States. A consensus recommendation on the goals and available resources for a new model for disease surveillance are provided. This new model should focus on "surveillance of the burden of disease," including (1) natural history of disease and (2) outcomes and complications of the disease and/or treatments.