multimorbidity

CRAN_Status_Badge CRAN_Download_Counter

The goal of multimorbidity is to create a single package which can take original claims data, with the ability to clean and organize them to a simple format, which can then be immediately used to obtain any of a number of comorbidity, frailty, and multimorbidity measures. This package is meant to be a simple and transparent one-stop-shop for those working with claims or other administrative health care data. The measures included in this package have been developed by other researchers and are detailed in depth below.

This package is meant to be both user-friendly and transparent. An individual should feel comfortable understanding what’s under the hood with these various metrics. Given this, this function has been written in a way that makes the code, including specific diagnosis codes, accessible.

For examples of the various functions included in this package, please see the documentation or the vignette.

Installation

There is currently no released version of multimorbidity on CRAN.

The development version of multimorbidity can be downloaded from GitHub with:

# install.packages("devtools")
devtools::install_github("WYATTBENSKEN/multimorbidity")

Citation Information

In addition to citing this R Package, we ask that you please cite the original manuscripts which developed these algorithms. Portions of this package, specifically the Elixhauser and Charlson diagnoses codes, were adapted from another package, comorbidity.

Data Cleaning Functions

There are two data cleaning functions in this package. The first, prepare_data() should be run first as it prepares the data to a uniform format, which the various measures rely on. The end-goal is to have a dataset that has 1 column with a patient ID, 1 column which contains the diagnosis code, and 1 column which will note if it’s ICD-9 (9), ICD-10 (10), or HCPCS/CPT (1). There are other variables that may be of interest depending on the specification including type (inpatient or outpatient) and date.

The second function, comorbidity_window(), is not highly suggest as the prepare_data() function is but may be useful to some investigators. Oftentimes, we may be interested in limiting our claims to a specific window, such as the 1-year before diagnosis. To accommodate this, comorbidity_window() will merge your prepared diagnosis dataset with an ID dataset and limit the claims/diagnoses to a specific time window relative to a date of interest.

Included Comorbidities and Indices

Index Citation(s)
Elixhauser Comorbidities and Index Elixhauser (1998), Moore et al. (2017)
Charlson Comorbidities and Index Charlson et al. (1987), Deyo et al. (1992), Quan et al. (2005)
Claims Frailty Index Kim et al. (2018)
Multimorbidity Weighted Index Wei et al. (2020)
Nicholson and Fortin Conditions Nicholson et al. (2015), Fortin et al. (2017)

Elixhauser Comorbidities and Index

The Elixhauser Comorbidities and Comorbidity Index are a widely-used set of comorbidities originally developed in 1998 by Elixhauser, with two indices for mortality and readmission created in 2017 by Moore et al.

In this package, we used the codes provided in the format programs by the Agency for Healthcare Research and Quality for ICD-9, ICD-10 Beta, and ICD-10. The ICD-10 data contain a larger set of comorbidities and, as of this writing, no calculator for the indices has been released, and thus when data contain both ICD-9 and ICD-10, we will use the ICD-9 comorbidities with the Beta code. Finally, the original algorithm takes into account DRG, which this package currently does not accommodate.

You can obtain the Elixhauser comorbidities and index by running the elixhauer() function.

Charlson Comorbidities and Index

The Charlson Comorbidities and Index are, similarly, a widely-used set of comorbidities. First developed in 1987 by Charlson et al., they’ve been modified a number of times. This algorithm employs the Deyo et al. list of 17 comorbidities, with the adaptations included in Quan et al.

You can obtain the Charlson comorbidities and index by running the charlson() function.

Claims Frailty Index

The Claims Frailty Index (CFI) is based off of work by Kim et al. in 2018. This algorithm uses ICD-9, ICD-10, and procedure codes to establish the frailty score for each patient. The code included in this package is largely developed from publicly-available code which can be found on the Harvard dataverse. As the original algorithms included HCPCS/CPT procedure codes, so does this.

You can obtain the CFI by running the cfi() function.

Multimorbidity Weighted Index

The Multimorbidity Weighted Index (MWI) was created by Wei et al. in 2020. This uses ICD-9 codes (note: ICD-10 is not yet available for MWI) to establish a multimorbidity index for each individual. The R code used in this package was developed based on the supplement included in the previously linked manuscript.

You can obtain the MWI by running the mwi() function.

Nicholson and Fortin Conditions

The Nicholson and Fortin Conditions were first published in 2015 and then updated to ICD-10 in 2017. These 20 chronic conditions are a standardized list used for multimorbidity research, and developed from a community-based primary healthcare project.

You can obtain the Nicholson and Fortin Conditions by running the nicholsonfortin() function.