obstinit

Background

Obstinit is our OBI "tools" package, meant to provide a series of tools we regularly use in data processing and visualization and to save time used on repetitive coding tasks.

Installation

Install the development version directly from GitHub:

# install.packages("devtools")
devtools::install_github("obianalysts-umich/obstinit")

Functions

Reading data

• read_current_data() reads the most recent (today's) OBI registry data from the Turbo drive. This function does not take any arguments and is built to work with both Mac and PC operating systems.
• read_crosswalk() reads in the current crosswalk file from Turbo and cleans the column names. This function does not take any arguments.

Creating OBI cohort

• create_obi_cohort() limits an OBI nightly export dataset to locked, complete cases with a delivery date after 2019. Note that case_lock_dt is programmed to match what is in the AMx workstation and represents the LAST DAY a case can be edited by a hospital; the case LOCKS at midnight that night. Therefore, cases are considered locked when case_lock_dt < today().
• create_90_day_lock_dt() doesn't take any arguments but provides the date at which cases pulled on a given day will lock. Can be used with create_obi_cohort().

OBI measure specs

• The create_*_cohort() family of functions limits the dataset to denominator criteria based on the measure of interest (Cat II, IA)
• The calculate_*_comp() family of functions calculates compliance with OBI measures (usually P4P scorecard measures) for Cat II, IA, TeamBirth, and Labor Dystocia. For dystocia compliance, individual types can be added by using include_types == T.

Plotting

• create_ctrl_df() creates a data frame containing the numerator, denominator, rate, CL (center line) UCL (upper control limit) and LCL (lower control limit) for a variable of interest (ex. Cesarean). The dataset is limited to complete cases; additional date or other constraints must be added manually. To use this function with plot_ctrl_hchart(), set for_highchart == T. This function used to be called structure_data().
• plot_ctrl_chart() takes the data frame constructed by structure_data() and plots it using ggplot2. Plot title, axis titles, and caption must be added manually.
• plot_ctrl_hchart() is identical to the plot_ctrl_chart() function, except that it formats everything to be used with the highchart rather than ggplot2 package.
• create_caterpillar_df() structures data for caterpillar (ranking) plots; each row of the data frame is a site name by default and the data frame includes a numerator, denominator, rate, and (0-1 bounded) 95% CI.
• plot_caterpillar_chart() plots a caterpillar chart, presumably using the dataframe created by create_caterpillar_df(). Plot title, axis titles, and caption must be added manually.
• theme_obi() is a ggplot2 theme that standardizes formatting of plots for semi-annual meetings, presentations, and more. It uses theme_minimal() as a base and additionally centers the plot title and subtitle, left-justifies the plot caption, standardizes text sizes across the plot, and makes all text OBI's dark blue. All OBI plots should use this theme.

Other data processing

• sort_muti_selection_var() takes in variable name and sort multiple selection values. For example sort {10,5,3,4} to {3,4,5,10}. use ?sort_muti_selection_var to see examples to run.
• add_CI_values: This function creates upper and lower CI for an observed rate. Two new variables (LC and UC) are added. use ?add_CI_values to see examples to run.

Example workflow - reading in data, structuring data, and plotting control chart in ggplot

library(obstinit)
library(tidyverse)

obi = read_current_data()

obi_cohort = obi %>% 
  create_obi_cohort()

ctrl_cohort = obi_cohort %>% 
  create_ctrl_df(infant_dob_dt,
                 overall_dystocia_compliance_num, 
                 overall_dystocia_den_all, 
                 date_gran = "quarter",
                 increase_is_bad = F)

ctrl_chart = ctrl_cohort %>% 
  plot_ctrl_chart()

Status updates

• 12/6/2023 - obstinit is fully functional and is considered to be in a finalized format. New functions are added as needed. Package is in maintenance phase and status updates will no longer be posted unless major changes are made.
• 9/19/2023 - added read_current_data
• 8/7/2023 - added calculate_dys_comp()
• 5/23/2023 - create_90_day_lock_dt() and create_obi_cohort() are live
• 5/19/2023 - plot_ctrl_hchart() is live
• 5/10/2023 - color legend fix added to plot_ctrl_chart
• 4/5/2023 - Added color gradient and new OBI colors to plot_ctrl_chart(). Note that there is still some debugging that needs to be done, primarily in regards to the color legend which is often incorrect.
• 3/29/2023 - plot_ctrl_chart() function is usable.
• 3/28/2023 - structure_data() function is usable.
• 3/23/2023 - under construction