--- title: "Getting started with SLOS" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{Getting started with SLOS} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8} --- ```{r setup, include = FALSE} # The code below needs an internet connection (pre-trained model) and/or several # minutes of model training, so it is shown but not evaluated when the vignette # is built. knitr::opts_chunk$set(collapse = TRUE, comment = "#>", eval = FALSE) ``` ## Overview The **standardized length of stay (SLOS)** of an intensive care unit (ICU) is the ratio between the length of stay observed in the unit and the length of stay expected for its patients: $$\mathrm{SLOS} = \frac{\sum \text{observed LoS}}{\sum \text{expected LoS}}$$ A SLOS below 1 means that patients stay in the ICU for less time than expected given their case mix; above 1, for longer. The expected length of stay comes from a prediction model. **SLOS** offers two ways of getting one: 1. use the **pre-trained model** (Section 2), or 2. **train a new model** on your own data (Section 3). In both cases, `SLOS()` then turns patient-level predictions into unit-level results (Section 4). ```{r load} library(SLOS) data(SampledData) # loads `sampled_data`, 1000 anonymized admissions ``` ## Using the pre-trained model `SLOS()` downloads the pre-trained model on first use (it needs an internet connection; the file is large, about 1.1 GB, so it can take a long time; the model is kept for the rest of the R session). `sampled_data` identifies each ICU in the column `UnitCode` and has the observed length of stay in `UnitLengthStay_trunc`. ```{r pretrained} res <- SLOS( data = sampled_data, icu_column = "UnitCode", outcome = "UnitLengthStay_trunc" ) head(res$df_unit_slos) # one row per ICU res$plot_SLOS_obs_prev # observed vs. predicted LoS per ICU plot(res$funnel_plot) # funnel plot ``` Patient-level predictions and error metrics are also available directly: ```{r predict} pred <- slos_predict_pretrained(sampled_data, outcome = "UnitLengthStay_trunc") pred$RMSE ``` ## Training a model with your own data The pre-trained model reflects the case mix and the data of the hospitals it was trained on. For other settings, build a new model. The workflow has four steps; a complete script is available in `system.file("examples", "retrain_workflow.R", package = "SLOS")`. **Step 1 – Pre-processing.** Removes near-zero variance predictors and highly correlated numeric or categorical predictors, splits the data into training and test sets and imputes missing values with *mice*. ```{r preprocess} outcome <- "UnitLengthStay_trunc" non_predictors <- c("UnitLengthStay_days", "admissionid", "UnitCode") pp <- slos_pre_processing( data = sampled_data, outcome = outcome, corr_cutoff_num = 0.8, corr_cutoff_cat = 0.5, do_impute = TRUE, include_outcome = TRUE, m = 1, maxit = 5, seed = 123 ) pp$removed # what was removed and why train_imp <- pp$imputation$train_completed test_imp <- pp$imputation$test_completed ``` **Step 2 – Feature selection** (recursive feature elimination). ```{r rfe} res_fs <- slos_feature_select_rfe( data = train_imp[, setdiff(names(train_imp), non_predictors)], outcome = outcome, method = "treebag", cv_folds = 5, verbose = FALSE ) res_fs$selected_vars ``` **Step 3 – Training.** A linear model and a random forest are stacked with a random-forest meta-learner. ```{r train} keep_cols <- union(res_fs$selected_vars, outcome) fit <- slos_train_new_model( train = train_imp[, keep_cols], test = test_imp, outcome = outcome, icu_column = "UnitCode" ) fit$metrics # patient-level and ICU-level RMSE, MAE and R2 ``` **Step 4 – Evaluate the units with the new model.** ```{r custom} res <- SLOS(test_imp, icu_column = "UnitCode", outcome = outcome, model = fit) ``` ## Interpreting the results `res$df_unit_slos` contains, for each ICU, the number of admissions, the sums of observed and expected length of stay and the SLOS. Summaries of the distribution across units are straightforward to compute: ```{r summary} slos_values <- res$df_unit_slos$SLOS sum(res$df_unit_slos$soma_los_obs) / sum(res$df_unit_slos$soma_los_esp) # overall median(slos_values) quantile(slos_values, c(0.25, 0.75)) ``` The funnel plot shows each unit's SLOS against its number of admissions with control limits around the overall SLOS; units outside the limits differ from the overall performance by more than chance would explain. ## References Peres et al. (2022) . Peres et al. (2023) . .