gen.SORT: A function to compute SORT scores

In dannyjnwong/HSRC: Useful functions for perioperative research

Description

This function will parse a dataframe and compute morbidity and mortality based on the Surgical Outcome Risk Tool (SORT). To use the function, you will need to manipulate your dataframe to have columns with the structure detailed below.

Usage

gen.SORT(x)

Arguments

x

A dataframe or tbl where each row is a patient observation, and the columns are SORT predictor variables. x must contain the following column names (not necessarily in order): Age continuous variable, numeric or integer ASA categorical variable, can be "I"; "II"; "III"; "IV"; "V" OpUrgency categorical variable, NCEPOD classifications, can be "Ele" = Elective; "Exp" = Expedited; "U" = Urgent; "I" = Immediate Specialty categorical variable, can be "Ortho" = Orthopaedics; "Colorectal" = Colorectal; "UpperGI" = Upper Gastrointestinal; "Vasc" = Vascular; "Bariatric" = Bariatric; "Other" = All else OpSeverity categorical variable, can be "Min" = Minor; "Int" = Intermediate; "Maj" = Major; "Xma" = Xmajor; "Com" = Complex Malignancy categorical variable, can be "NM" = Not malignant; "PM" = Primary malignancy only; "MNM" = Malignancy + nodal metastases; "MDM" = Malignancy + distal metastases; Or "Y" / "N"; Or "1" / "0"

Value

A dataframe (or tbl), which you can assign to an object, with the following variables:

SORT_mortLogit

The log-odds for mortality as calculated by SORT

SORT_morbLogit

The log-odds for morbidity as calculated by SORT

Converting to probability scale

The function will produce SORT_mortLogit and SORT_morbLogit values which are on the log-odds scale To convert to probabilities (0 to 1 scale), use arm::invlogit(). See: invlogit.

References

• Protopapa KL, Simpson JC, Smith NCE, Moonesinghe SR. Development and validation of the Surgical Outcome Risk Tool (SORT). Br J Surg. 2014 Dec;101(13):1774–83. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4240514/.

• Wong DJN, Oliver CM, Moonesinghe SR. Predicting postoperative morbidity in adult elective surgical patients using the Surgical Outcome Risk Tool (SORT). BJA: British Journal of Anaesthesia. 2017;119(1):95–105. https://doi.org/10.1093/bja/aex117

Examples

## Not run: 
#Example of pre-processing to rename data variables to match expected column names
library(tidyverse)

test_data <- raw_data %>%
select(Age = S01Age,
        ASA = S03AsaPsClass,
        OpUrgency = S02OperativeUrgency,
        Specialty = Specialty,
        OpSeverity = S02PlannedProcSeverity,
        Malignancy = S04Malignancy)

## End(Not run)

test_data <- patients
test_output <- gen.SORT(test_data)
head(test_output)

dannyjnwong/HSRC documentation built on May 18, 2019, 2:37 a.m.