R package for cleaning data from Electronic Health Record systems, focused on cleaning height and weight measurements.
Carrie Daymont, Michelle E Ross, A Russell Localio, Alexander G Fiks, Richard C Wasserman, Robert W Grundmeier, Automated identification of implausible values in growth data from pediatric electronic health records, Journal of the American Medical Informatics Association, Volume 24, Issue 6, November 2017, Pages 1080–1087, https://doi.org/10.1093/jamia/ocx037
This package also includes an R version of the SAS macro published by the CDC for calculating percentiles and Z-scores of pediatric growth observations and utilities for working with both functions. As of summer 2021, it also supports cleaning anthropometric measurements for adults up to age 65. The adult algorithm has not yet been published in a peer-reviewed publication, but is described in detail at Adult algorithm.
To install the stable version from CRAN:
growthcleanr package processes data prepared in a specific format to identify biologically implausible height and weight measurements. It bases these evaluations on techniques which use patient-specific longitudinal analysis and variations from published growth trajectory charts for pediatric subjects. These techniques are performed in a specific order which refines and improves results throughout the process.
growthcleanr include a flag for each measurement indicating whether it is to be included or excluded based on plausibility, with a variety of specific types of exclusions identified distinctly. These flags can be analyzed further by researchers studying anthropometric EHR data to determine which measurements to include or exclude in their own studies. No values are deleted or otherwise removed; each is only flagged in a new column.
To start running
growthcleanr, an R installation with a variety of additional packages is required, as is a growth measurement dataset prepared for use in
The rest of this documentation includes:
growthcleanrwith large data sources