README.md

In nyilin/QcDM: Quality Care for Diabetes Mellitus

QcDM: Quality care for Diabetes Mellitus

Overview

QcDM is an R package developed for generating glucometrics measures from point-of-care blood glucose data at three different units of analysis: patient-sample, patient-day, and patient-stay. This package is distributed under the MIT license.

This package is a part of the QcDM Project, which provides a user-friendly R-Shiny user interface for convenient and flexible data processing, and generates a detailed and well-formatted report on the glucometrics measurements.

Users are strongly recommended to use QcDM via the R-Shiny interface. Sections below describes the installation of the package and its basic usage.

Citation

Ying Chen, Yilin Ning, Prem Thomas, Mark Salloway, Maudrene Luor Shyuan Tan, E-Shyong Tai, Shih Ling Kao, Chuen Seng Tan, An open source tool to compute measures of inpatient glycemic control: translating from healthcare analytics research to clinical quality improvement, JAMIA Open, Volume 4, Issue 2, April 2021, ooab033, https://doi.org/10.1093/jamiaopen/ooab033

Installation

Package devtools is needed to install the QcDM package from Github:

# First, install devtools if it is not already installed:
install.packages("devtools")
# Then, load devtools to install QcDM from Github:
library(devtools)
install_github("nyilin/QcDM")

Basic usage

The QcDM package includes a simulated example data to illustrate its basic usage, which was generated for 4 hypothetical ward locations (indicated by letters A to D) in July 2020. The following is a preview of the first 5 rows of the example data:

library(QcDM)
data("gluDat")
head(gluDat, 5)
##   ADMISSION.ID RESULT    RESULT.DATE LOCATION
## 1            1   12.6 7/1/2020 03:13        A
## 2            2   14.1 7/1/2020 03:15        B
## 3            2    8.6 7/1/2020 06:43        B
## 4            1   17.3 7/1/2020 11:56        A
## 5            1   14.1 7/1/2020 12:26        A

The unit of measurement of glucose readings in the example data is mmol/L. To illustrate how the software program developed in the QcDM Project handles non-numerical values, three entries had non-numerical values:

gluDat[is.na(as.numeric(gluDat$RESULT)), ]
## Warning in `[.data.frame`(gluDat, is.na(as.numeric(gluDat$RESULT)), ): NAs
## introduced by coercion
##      ADMISSION.ID RESULT     RESULT.DATE LOCATION
## 2300           60   c7.7 7/16/2020 14:04        D
## 4605           70   d9.6 7/24/2020 13:53        A
## 5143           47   a4.4 7/26/2020 22:58        B

Before generating glucometrics measurements, first process the date-time stamps of glucose readings, and identify glucose monitoring episodes based on the admission information in ADMISSION.ID. In this example, we focus on ward A.

gluDat2 <- FormatDate(dat = gluDat[gluDat$LOCATION == "A", ], yy = 2020, mm = 7)
gluDat3 <- GenEpisode(dat = gluDat2, epiMethod = "Admininfo")
## 1 rows with non-numeric glucose readings are removed.

Then, generate basic glucometrics measurements using the following cutoffs as an example:

• Hyperglycemia: consider three cutoff values, i.e., higher than or equal to 14 mmol/L, 20 mmol/L or 24 mmol/L.
• Hypoglycemia: consider three cutoff values, i.e., below 4 mmol/L, 3 mmol/L or 2.5 mmol/L.
• Normal range of glucose readings: above or equal to 4 mmol/L and below 10 mmol/L.
• Hyperglycemia index (HGI): above 10 mmol/L per hour of hospital stay.

metricList <- GenGluM(dat = gluDat3, hypocutoffs = c(4, 3, 2.5),
                      hypercutoffs = c(14, 20, 24), normalrange = c(4, 10),
                      hgicutoff = 10, unitVal = 1)
knitr::kable(ProGluTable(metricList = metricList, unitVal = 1)[[1]])

| | Patient-sample | Patient-day | Patient-stay | |:-----------------------------------------------------------|:---------------|:------------|:-------------| | Number (count) | 1610 | 316 | 29 | | Percent with glucose >= hyper-cutoff1 | 697 (43.3%) | 292 (92.4%) | 29 (100%) | | Percent with glucose >= hyper-cutoff2 | 61 (3.8%) | 56 (17.7%) | 23 (79.3%) | | Percent with glucose >= hyper-cutoff3 | 5 (0.3%) | 5 (1.6%) | 4 (13.8%) | | Median HGI | | | 3.5 (0.7) | | Mean HGI | | | 3.6 (0.5) | | Percent with glucose in normal range | 247 (15.3%) | 6 (1.9%) | 0 (0%) | | Median glucose | 13.2 (0.4) | 13.6 (2) | 13.5 (0.5) | | Mean glucose | 13.5 (0.5) | 13.5 (1.7) | 13.5 (0.5) | | Patient-day weighted median glucose | | | 13.5 (0.4) | | Patient-day weighted mean glucose | | | 13.5 (0.5) | | Percent with glucose < hypo-cutoff1 | 1 (0.1%) | 1 (0.3%) | 1 (3.4%) | | Percent with glucose < hypo-cutoff2 | 0 (0%) | 0 (0%) | 0 (0%) | | Percent with glucose < hypo-cutoff3 | 0 (0%) | 0 (0%) | 0 (0%) | | Percent of patient-stays with a recurrent hypoglycemia day | | | 0 (0%) | | Median SD | | 3.2 (1.6) | 3.4 (0.5) | | Mean SD | | 3.3 (1.3) | 3.5 (0.5) | | Median J-index | | 90.8 (31) | 92.6 (10.5) | | Mean J-index | | 92.9 (24.2) | 92.8 (7.3) |

nyilin/QcDM documentation built on June 29, 2021, 1:14 a.m.