idxpr

idxpr provides R functions to transform and standardize electronic health record (EHR) diagnosis and procedure data into analysis-ready formats for clinical and epidemiological research.

Key Features:

• ICD Code Transformation: Standardizes International Classification of Diseases (ICD) codes into a uniform format.
• Data Grouping Strategies: Supports multiple grouping strategies, including four strategies for diagnoses and two for procedures, to aggregate clinical codes into clinically meaningful groups.
• Flexible Query Functions: Enables dividing EHRs into case and control cohorts based on specified criteria and splitting records temporally around events using record dates.
• Data Integration and Wrangling: Provides mechanisms to integrate and wrangle heterogeneous and sparse clinical data into consolidated long-format datasets.
• Data Structure Conversion: Converts integrated long-format clinical data into wide, analysis-ready formats suitable for statistical analysis and visualization.

Scientific Applications:

• Clinical cohort and comorbidity studies: Prepares enriched EHR datasets for studying disease progression and comorbidities in clinical and epidemiological research.
• MIMIC-III newborn cohort example: Used to define patent ductus arteriosus (PDA) cases using ICD-9-CM codes and compare them to controls without PDA, grouping diagnoses into comorbidities and revealing statistically significant differences between groups.

Methodology:

Transforms ICD codes, applies specified grouping strategies (four diagnosis, two procedure), performs flexible queries to define case/control cohorts and temporal splits by record date, integrates and wrangles long-format clinical data, and converts resulting datasets to wide, analysis-ready formats.