Leveraging electronic health records data to predict multiple sclerosis disease activity

ABSTRACTBackgroundThe current classification of pediatric CD is mainly based on cross-sectional data. The objective of this study is to identify subgroups of pediatric CD through trajectory cluster analysis of disease activity using data from electronic health records.MethodsWe conducted a retrospective study of pediatric CD patients who had been treated with infliximab. The evolution of disease over time was described using trajectory analysis of longitudinal data of C-Reactive Protein (CRP). Patterns of disease evolution were extracted through functional principal components analysis and subgroups were identified based on those patterns using the Gaussian mixture model. We compared patient characteristics, a biomarker for disease activity, received treatments, and long-term surgical outcomes across subgroups.ResultsWe identified four subgroups of pediatric CD patients with differential relapse-and-remission risk profiles. They had significantly different disease phenotype (p < 0.001), CRP (p < 0.001) and calprotectin (p = 0.037) at diagnosis, with increasing percentage of inflammatory phenotype and declining CRP and fecal calprotectin levels from Subgroup 1 through 4. The risk of colorectal surgery within 10 years after diagnosis was significantly different between groups (p < 0.001). We did not find statistical significance in gender or age at diagnosis across subgroups, but the BMI z-score was slightly smaller in subgroup 1 (p =0.055).ConclusionsReadily available longitudinal data from electronic health records can be leveraged to provide a deeper characterization of pediatric Crohn disease. The identified subgroups captured novel forms of variation in pediatric Crohn disease that were not explained by baseline measurements and treatment information.SummaryThe current classification of pediatric Crohn disease mainly relies on cross-sectional data, e.g., the Paris classification. However, the phenotypic classification may evolve over time after diagnosis. Our study utilized longitudinal measures from the electronic health records and stratified pediatric Crohn disease patients with differential relapse-and-remission risk profiles based on patterns of disease evolution. We found trajectories of well-maintained low disease activity were associated with less severe disease at baseline, early initiation of infliximab treatment, and lower risk of surgery within 10 years of diagnosis, but the difference was not fully explained by phenotype at diagnosis.

Download Full-text

The prevalence of hypertension in multiple sclerosis based on 37 million electronic health records from the United States

European Journal of Neurology ◽

10.1111/ene.14557 ◽

2020 ◽

Author(s):

F. B. S. Briggs ◽

E. Hill ◽

H. Abboud

Keyword(s):

United States ◽

Multiple Sclerosis ◽

Electronic Health Records ◽

The United States ◽

Health Records ◽

Electronic Health

Download Full-text

Integration of the Extreme Gradient Boosting model with electronic health records to enable the early diagnosis of multiple sclerosis

Multiple Sclerosis and Related Disorders ◽

10.1016/j.msard.2020.102632 ◽

2021 ◽

Vol 47 ◽

pp. 102632

Author(s):

Ruoning Wang ◽

Wenjing Luo ◽

Zifeng Liu ◽

Weilong Liu ◽

Chunxin Liu ◽

...

Keyword(s):

Multiple Sclerosis ◽

Early Diagnosis ◽

Electronic Health Records ◽

Gradient Boosting ◽

Health Records ◽

Extreme Gradient Boosting ◽

Electronic Health

Download Full-text

Estimating the prevalence of multiple sclerosis using 56.6 million electronic health records from the United States

Multiple Sclerosis Journal ◽

10.1177/1352458519864681 ◽

2019 ◽

Vol 26 (14) ◽

pp. 1948-1952 ◽

Cited By ~ 3

Author(s):

Farren BS Briggs ◽

Eddie Hill

Keyword(s):

Multiple Sclerosis ◽

Electronic Health Records ◽

Census Data ◽

Black Americans ◽

The United States ◽

Health Claims ◽

Health Records ◽

Black Race ◽

Electronic Health ◽

Age And Sex

Background/objective: In 2019, the 2010 U.S. multiple sclerosis (MS) prevalence was robustly estimated (265.1–309.2/100,000) based on large administrative health-claims datasets. Using 56.6 million electronic health records (EHRs), we sought to generate complementary age, sex, and race standardized estimates. Methods/results: Using de-identified EHRs and 2018 U.S. Census data, we estimated an age- and sex-standardized MS prevalence of 219.5/100,000 which increased to 274.5/100,000 when accounting for White and Black race alone. Women aged 50 to 69 years had the highest prevalence (>600/100,000). Among White and Black Americans, the age- and sex-standardized prevalence was 283.7 and 226.1 per 100,000, respectively. Conclusion: Using 56.6 million EHRs and standardizing for age, sex, and race (White and Black Americans only), we estimated at least 810,504 Americans were living with MS in 2018.

Download Full-text

Prevalence of multiple sclerosis within a healthcare delivery system in Northern California: a retrospective, electronic, health records-based study from 2010 to 2016

10.26226/morressier.5b75785c5aff7400151f3b3b ◽

2018 ◽

Author(s):

Robert Romanelli

Keyword(s):

Multiple Sclerosis ◽

Electronic Health Records ◽

Delivery System ◽

Healthcare Delivery ◽

Northern California ◽

Healthcare Delivery System ◽

Health Records ◽

Electronic Health

Download Full-text

Creation of a ustekinumab external control arm for Crohn's disease using electronic health records data: a pilot study

10.1101/2021.11.12.21266064 ◽

2021 ◽

Author(s):

Vivek Ashok Rudrapatna ◽

Yao-Wen Cheng ◽

Colin Feuille ◽

Arman Mosenia ◽

Jonathan Shih ◽

...

Keyword(s):

Crohn’S Disease ◽

Crohn's Disease ◽

Electronic Health Records ◽

Disease Activity ◽

Real World ◽

External Control ◽

Health Records ◽

Electronic Health ◽

Trial Efficiency ◽

Manual Review

Objectives: The use of external control arms to support claims of efficacy and safety is growing in interest among drug sponsors and regulators. However, experience with performing these kinds of studies for complex, immune-mediated diseases is limited. We sought to establish a method for creating an external control arm for Crohn's disease. Methods: We queried electronic health records databases and screened records at the University of California, San Francisco to identify patients meeting the major eligibility criteria of TRIDENT, a concurrent trial involving ustekinumab as a reference arm. Timepoints were defined to balance the tradeoff between missing disease activity and bias. We compared two imputation models by their impacts on cohort membership and outcomes. We compared the results of ascertaining disease activity using structured data algorithms against manual review. We used these data to estimate ustekinumab's real-world effectiveness. Results: Screening identified 183 patients. 30% of the cohort had missing baseline data. Two imputation models were tested and had similar effects on cohort definition and outcomes. Algorithms for ascertaining non-symptom-based elements of disease activity were similar in accuracy to manual review. The final cohort consisted of 56 patients. 34% of the cohort was in steroid-free clinical remission by week 24. Conclusions: Differences in the timing and goals of real-world encounters as compared to controlled studies directly translate into significant missing data and lost sample size. Efforts to improve real-world data capture and better align trial design with clinical practice may enable robust external control arm studies and improve trial efficiency.

Download Full-text