normalizing risk of mortality scores

Question

Hi,

I need to normalize risk of mortality scores. One is called RePEAT or PRISM and the other is called PIM. Repeat/ Prism takes place in the ER, and PIM is measured in the Peds ICU on admission.

No idea now to normalize these.

any help appreciated!

Thanks, .

Byron_JMP · Answer

So just looking at the data, it looks like there are some pretty big outliers, and those outliers aren't really handled by the standardization. This means that in all your "delta" columns there are big outliers too. The outliers are going to affect statistical tests that assume the data is normally distributed.  This isn't a show stopper, but something that needs to be considered before making comparisons between the groups.

Speaking of groups, there is a crossover between the levels in consult assigned and consult received. This means that consult assigned needs to be nested in consult received.

One way to see the structure of the data is to use the Variability chart. The script to replicate the figure below is included at the bottom. Download the script, make the table: 11.16data.analysis.standardized.delta.calc.jmp the primary table (click on it) then run the script.

or, make a new script window, past the JSL below into the script window and run it from there.

Without doing any calculations you can begin to see the differences between telemedicine and telephone based on the source. Standardizing and Normalizing the data might show you a different picture, but just looking at raw data is often helpful.

Variability Chart(
    Y( :sPIM2ROM ),
    X( :consult_received, :consult_assigned ),
    Model( "Nested" ),
    Std Dev Chart( 0 ),
    Mean Plots( 1 ),
    Column Switcher(
        :sPIM2ROM,
        {:calc.prisa2_score, :sPRISA, :calc.RePEAT_score_817, :sRePEAT, :vps.PIM.2.ROM, :sPIM2ROM,
        :Name( "Box Cox[sPIM2ROM]" ), :vps.PIM.3.ROM, :sPIM3ROM, :Name( "delta sPRISA -sPIM2ROM" ),
        :Name( "delta sPRISA- sPIM3ROM" ), :Name( "Delta sRePEAT-PIM2ROM" ), :Name( "delta sRePEAT- PIM3ROM" )}
    )
)

example variability chart.jsl

Ellabluebird · Answer

hopefully these help...not sure how to attatch .jrp files

GSD workbook 8.15 redo data.xlsx

11.16standardized delta.jmp

11.16data.analysis.standardized.delta.calc.jmp

Ellabluebird · Answer

thanks! So I have ER risk of mortality scores, and PICU risk of mortality scores. the ER scores are called RePEAT and PRISM. I'm trying to see if theres an association between any change in risk of mortality scores from kiddos in the ER who were then admitted to the PICU. I'm looking at whether there differences  between the type of consult used ( telemedicine vs. telephone). my hypothesis is that perhaps telemedicine led to a greater decrease in risk of mortality scores.

so. I Have the scores for 345 patients. I standardized them in excel using the standardize function because thats what Dr. Google said to do, and then subtracted them ( PIM - PRISM) and (PIM- RePEAT).

now I'm going to do a wilcoxon test to see whether there are significant differences between the telemed or telephone cohort.

am i on the right track? I'm kinda doing this flying blind.