Introduction to sbcdata

Source: vignettes/sbcdata.Rmd

Authors: Sebastian Gibb [aut, cre], Paul Ahrens [aut, dtc], Daniel Steinbach [aut, dtc], Maria Schmidt [aut, dtc], Thorsten Kaiser [aut, dtc], Mark Wernsdorfer [dtc], Matthias Nauck [dtc], Stefan Bollmann [dtc], Thomas Hildebrandt [dtc]
Last modified: 2024-03-07 08:43:33.649335
Compiled: Thu Mar 7 08:45:51 2024

Introduction

The data were collected as part of the AMPEL project (Analysis and Reporting System for the Improvement of Patient Safety through Real-Time Integration of Laboratory Findings, ). They include administration data, sepsis labels based on ICD10 codes and laboratory diagnostics from patients admitted to the University Hospital Leipzig between Januar 2014 and December 2019 and to the University Hospital Greifswald between Januar 2015 and December 2020, respectively. Additionally it provides functions to generate a similar dataset from MIMIC-IV (Johnson et al. 2021).

Installation

You can install the released version of sbcdata from directly from github:

# install.packages("remotes")
remotes::install_github("ampel-leipzig/sbcdata")

License

The datasets are available under the CC-BY license.

Citation

To cite the sbcdata package, please use:

citation("sbcdata")
## To cite the sbcdata package in publications use:
## 
##   D. Steinbach, et al. 2024. Applying Machine Learning to Blood Count
##   Data Predicts Sepsis with ICU Admission Clinical Chemistry 70,
##   506-515.
## 
## A BibTeX entry for LaTeX users is
## 
##   @Article{sbcdata,
##     title = {Applying Machine Learning to Blood Count Data Predicts Sepsis with {ICU} Admission},
##     author = {Daniel Steinbach and Paul C. Ahrens and Maria Schmidt and Martin Federbusch and Lara Heuft and Christoph Lübbert and Matthias Nauck and Matthias Gründling and Berend Isermann and Sebastian Gibb and Thorsten Kaiser},
##     volume = {70},
##     number = {3},
##     pages = {506--515},
##     year = {2024},
##     doi = {10.1093/clinchem/hvae001},
##     journal = {Clinical Chemistry},
##     publisher = {Oxford University Press},
##   }

The sbcdata package

Overview

The sbcdata dataset contains ncol(sbcdata) variables for nrow(sbcdata) laboratory diagnostics.

str(sbcdata)
## Classes 'data.table' and 'data.frame':   2709568 obs. of  18 variables:
##  $ Id       : int  1 2 3 3 3 3 3 3 3 3 ...
##  $ Age      : int  25 75 77 77 77 77 77 77 77 77 ...
##  $ Sex      : chr  "W" "M" "W" "W" ...
##  $ Diagnosis: chr  "Control" "Control" "Sepsis" "Sepsis" ...
##  $ Center   : chr  "Greifswald" "Greifswald" "Greifswald" "Greifswald" ...
##  $ Set      : chr  "Validation" "Validation" "Validation" "Validation" ...
##  $ Sender   : chr  "AMB" "GEN" "OTHER" "OTHER" ...
##  $ Episode  : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Time     : num  0 0 0 318840 578640 ...
##  $ TargetIcu: chr  NA NA NA NA ...
##  $ SecToIcu : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ CRP      : num  15.5 7.4 96.1 57 93.4 74.3 51.7 27 48 95.7 ...
##  $ HGB      : num  7 8.4 4.8 4.4 5.7 5.7 5.8 5.6 6.2 5.9 ...
##  $ MCV      : num  80.5 87.9 81.7 82.2 82 81.7 85.1 84.7 85.6 83.3 ...
##  $ PCT      : num  NA NA NA NA 0.22 NA NA NA NA 0.35 ...
##  $ PLT      : int  264 260 385 416 437 439 391 311 333 285 ...
##  $ RBC      : num  4.2 4.8 3 2.8 3.5 3.5 3.4 3.3 3.8 3.5 ...
##  $ WBC      : num  8.4 8.47 13.2 14.2 13.8 10.9 9.65 8.98 8.77 8.85 ...
##  - attr(*, ".internal.selfref")=<externalptr>

Missing data 

centers <- c("Greifswald", "Leipzig")

labs <- sbcdata[, labcodes$Code, with = FALSE]
labs <- as.matrix(labs)
labsna <- is.na(labs)
mode(labsna) <- "numeric"

nna <- rowsum(labsna, group = sbcdata$Center)
nna <- rbind(nna, Total = colSums(nna))
knitr::kable(t(nna), caption = "Missing Data")
Missing Data
Greifswald Leipzig Total
CRP 344944 737450 1082394
HGB 19715 109956 129671
MCV 19726 109978 129704
PCT 620732 1885699 2506431
PLT 20059 110912 130971
RBC 19725 109973 129698
WBC 19725 109955 129680 
pna <- round(nna / c(table(sbcdata$Center), nrow(sbcdata)) * 100, 2)

knitr::kable(t(pna), caption = "Missing Data [%]")
Missing Data [%]
Greifswald Leipzig Total
CRP 51.83 36.08 39.95
HGB 2.96 5.38 4.79
MCV 2.96 5.38 4.79
PCT 93.26 92.26 92.50
PLT 3.01 5.43 4.83
RBC 2.96 5.38 4.79
WBC 2.96 5.38 4.79

Densities 

.plotdens <- function(l, main,
                      col = viridisLite::cividis(length(l) + 1)[-1L]) {
    old.par <- par(no.readonly = TRUE)
    on.exit(par(old.par))

    d <- lapply(l, density, na.rm = TRUE)
    xlim <- range(vapply(d, function(x)range(x$x), double(2)))
    ylim <- c(0, max(vapply(d, function(x)max(x$y), double(1))))

    plot(
        NA, xlim = xlim, ylim = ylim,
        axes = FALSE, xlab = "", ylab = "", main = ""
    )
    for (i in seq_along(d)) {
        lines(d[[i]], col = col[i])
        polygon(d[[i]], col = sub("FF$", "55", col[i]))
    }
    title(main = main, adj = 0L)
    title(ylab = "Density", adj = 1L)
    legend(
        "topright",
        legend = paste0(
            names(l), ": ",
            "N = ", vapply(d, function(x)x$n, integer(1)), "; ",
            "BW = ", vapply(d, function(x)formatC(x$bw), character(1))
        ),
        col = col, lty = 1, lwd = 1, cex = 0.75, bty = "n"
    )
    axis(1, lwd.ticks = 0L, col = "#808080")
    axis(2, lwd.ticks = 0L, col = "#808080")
}

for (v in colnames(labs))
    .plotdens(split(labs[, v], sbcdata$Center), v)

Boxplot HGB 

boxplot(HGB ~ Sex + Center, data = sbcdata, col = viridisLite::cividis(2))

Prevalence 

dup <- duplicated(sbcdata[, .(Id, Center)])
tbl <- table(sbcdata[!dup, .(Center, Diagnosis)])
addmargins(tbl)
##             Diagnosis
## Center       Control Sepsis   SIRS    Sum
##   Greifswald  161160   6995    986 169141
##   Leipzig     732763  10231   1211 744205
##   Sum         893923  17226   2197 913346
proportions(tbl, margin = 1)
##             Diagnosis
## Center           Control      Sepsis        SIRS
##   Greifswald 0.952814516 0.041356028 0.005829456
##   Leipzig    0.984625204 0.013747556 0.001627240

Session Information 

## R version 4.3.3 (2024-02-29)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 22.04.4 LTS
## 
## Matrix products: default
## BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3 
## LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.20.so;  LAPACK version 3.10.0
## 
## locale:
##  [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8       
##  [4] LC_COLLATE=C.UTF-8     LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8   
##  [7] LC_PAPER=C.UTF-8       LC_NAME=C              LC_ADDRESS=C          
## [10] LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C   
## 
## time zone: UTC
## tzcode source: system (glibc)
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
## [1] sbcdata_1.0.2     data.table_1.15.2
## 
## loaded via a namespace (and not attached):
##  [1] vctrs_0.6.5       cli_3.6.2         knitr_1.45        rlang_1.1.3      
##  [5] xfun_0.42         highr_0.10        purrr_1.0.2       textshaping_0.3.7
##  [9] jsonlite_1.8.8    htmltools_0.5.7   ragg_1.2.7        sass_0.4.8       
## [13] rmarkdown_2.26    evaluate_0.23     jquerylib_0.1.4   fastmap_1.1.1    
## [17] yaml_2.3.8        lifecycle_1.0.4   memoise_2.0.1     compiler_4.3.3   
## [21] fs_1.6.3          systemfonts_1.0.5 digest_0.6.34     viridisLite_0.4.2
## [25] R6_2.5.1          magrittr_2.0.3    bslib_0.6.1       tools_4.3.3      
## [29] pkgdown_2.0.7     cachem_1.0.8      desc_1.4.3

References

Johnson, Alistair E W, Lucas Bulgarelli, Tom Pollard, Steven Horng, Leo Anthony Celi, and Roger Mark. 2021. “MIMIC-IV.” PhysioNet. https://doi.org/10.13026/S6N6-XD98.