cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Browse apps to extend the software in the new JMP Marketplace
%3CLINGO-SUB%20id%3D%22lingo-sub-777117%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3ESo%20bestimmen%20Sie%20den%20optimalen%20Biomarker-Grenzwert%20in%20einem%20multivariablen%20logistischen%20Regressionsmodell%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-777117%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3E%3CP%3ELieber%20Kollege%3C%2FP%3E%3CP%3EMein%20Labor%20hat%20einen%20Biomarker%20entdeckt%2C%20der%20schlechte%20Ergebnisse%20f%C3%BCr%20schwangere%20M%C3%BCtter%20vorhersagt.%20Ich%20versuche%20herauszufinden%2C%20wie%20man%20einen%20optimalen%20Grenzwert%20im%20multivariablen%20Modell%20bestimmt%2C%20also%20einen%2C%20der%20alle%20klinischen%20Merkmale%20und%20den%20Biomarker%20umfasst%2C%20der%20die%20bin%C3%A4ren%20Ergebnisse%20vorhersagt%2C%20also%20schlechte%20Ergebnisse%20f%C3%BCr%20schwangere%20M%C3%BCtter.%20Ich%20verwende%20das%20Entscheidungsschwellentool%2C%20um%20die%20Sensitivit%C3%A4t%20und%20Spezifit%C3%A4t%20zu%20%C3%A4ndern%2C%20habe%20jedoch%20Schwierigkeiten%2C%20die%20Modellwahrscheinlichkeit%20in%20tats%C3%A4chliche%20Biomarkerwerte%20zu%20%C3%BCbersetzen.%20F%C3%BCr%20jede%20Hilfe%20w%C3%A4re%20ich%20sehr%20dankbar.%20Ich%20bin%20ein%20dummer%20Kliniker%2C%20also%20w%C3%A4re%20es%20noch%20hilfreicher%2C%20wenn%20Sie%20mich%20durch%20den%20Prozess%20f%C3%BChren%20w%C3%BCrden.%3C%2FP%3E%3CP%3EVielen%20Dank%3C%2FP%3E%3C%2FLINGO-BODY%3E%3CLINGO-LABS%20id%3D%22lingo-labs-777117%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3E%3CLINGO-LABEL%3EPr%C3%A4diktive%20Modellierung%20und%20maschinelles%20Lernen%3C%2FLINGO-LABEL%3E%3C%2FLINGO-LABS%3E%3CLINGO-SUB%20id%3D%22lingo-sub-777455%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3EBetreff%3A%20Wie%20bestimmt%20man%20den%20optimalen%20Biomarker-Grenzwert%20in%20einem%20multivariablen%20logistischen%20Regressionsmodell%3F%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-777455%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3E%3CP%3EAus%20dem%20roten%20Dreieckmen%C3%BC%20(oben%20links)%3C%2FP%3E%0A%3CP%3ESchalten%20Sie%20den%20Profiler%20ein.%20Er%20zeigt%20Ihnen%20eine%20grafische%20Version%20des%20Wahrscheinlichkeitsmodells.%3C%2FP%3E%0A%3CP%3EDie%20vertikalen%20gestrichelten%20Linien%20k%C3%B6nnen%20nach%20rechts%20und%20links%20verschoben%20werden%2C%20um%20die%20Auswirkung%20der%20Faktoren%20auf%20die%20Antwort%20anzuzeigen.%3C%2FP%3E%0A%3CP%3E%20%3C%2FP%3E%0A%3CP%3Edieses%20Video%20hat%20nichts%20mit%20Ihrem%20Szenario%20zu%20tun%2C%20aber%20in%20den%20letzten%2030%E2%80%9340%20Sekunden%20gibt%20es%20ein%20Beispiel%20dessen%2C%20was%20Sie%20tun%20m%C3%B6chten.%20(vielleicht)%3C%2FP%3E%0A%3CP%3E%3CA%20href%3D%22https%3A%2F%2Fcommunity.jmp.com%2Ft5%2FStatistical-Thinking-for%2FFitting-a-Logistic-Regression-Model-with-Interactions%2Fta-p%2F271976%22%20target%3D%22_blank%22%3Ehttps%3A%2F%2Fcommunity.jmp.com%2Ft5%2FStatistical-Thinking-for%2FFitting-a-Logistic-Regression-Model-with-Interactions%2Fta-p%2F271976%3C%2FA%3E%3C%2FP%3E%0A%3CP%3E%20%3C%2FP%3E%0A%3CP%3E%20%3C%2FP%3E%3C%2FLINGO-BODY%3E
Choose Language Hide Translation Bar
jambo
Level I

How do to determine optimal biomarker cutoff in a multivariable logistic regression model

Dear Colleague 

My lab has discovered a biomarker that predicts bad outcomes for pregnant moms. I am trying to figure out how to determine an optimal cutoff in the multivariable model, i.e., one that includes all the clinical characteristics and the biomarker that predicts the binary outcomes, i.e., bad outcomes for pregnant moms.  I use the decision threshold tool to change the sensitivity and specificity, but I have difficulty translating model probability to actual biomarker levels.  Any help would be most appreciated. I am a dumb clinician, so if you walk me through the process, that would be even more helpful.  

Thanks much 

1 REPLY 1
Byron_JMP
Staff

Re: How do to determine optimal biomarker cutoff in a multivariable logistic regression model

From the red triangle menu (near the top left)

Turn on the Profiler.  It will let you see a graphical version of the model for probability.

the vertical dashed lines can be moved right and left to see the effect of the factors on the response.

 

this video has nothing to do with your scenario, however in the last 30-40 sec there is an example of what you want to do. (maybe)

https://community.jmp.com/t5/Statistical-Thinking-for/Fitting-a-Logistic-Regression-Model-with-Inter...

 

 

JMP Systems Engineer, Health and Life Sciences (Pharma)