cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
The Discovery Summit 2025 Call for Content is open! Submit an abstract today to present at our premier analytics conference.
Choose Language Hide Translation Bar
View Original Published Thread

Application of Reed–Muench methodology using JMP

Reed–Muench method is the simplest and most intuitive way to determine an endpoint in a biological procedure.

 

The above methodology was published in The American Journal of Hygiene in 1937. A Simple Method of Estimating Fifty Per Cent Endpoints This is the methodology described in the paper.

 

undefined

Make the data presented in the thesis into a JMP Data Table. In the data corresponding to Table 1 of the paper, the data corresponding to dilution is quantified through log transformation.

 

undefined

To create a variable called Log(Dilution) in the red box above, you need to create a simple formula (JSL Code) as shown below.

 

Log10(
 Num( Substr( :Dilution, 3, 3 ) ) /
 Num( Substr( :Dilution, 1, 1 ) )
)

After getting the 3-digit value from the 3rd digit in the dilution variable, take the Log10 value.

 

 

After performing the above operations, use Log (Dilution) on the X axis and Percent Mortality on the Y axis to fit the curve using JMP Graph Builder. Below that, a graph of overall survival/death is also shown.

undefined

 

The endpoint we want to find here is Dilution, where the Percent Mortality value is 50%. In the actual paper, interpolation was used to calculate manually, but here, we will use JMP's Nonlinear Modeling method to find the value.

 

To use this method, select Analyze >> Specialized Modeling >> Nonlinear menu.

 

undefined

 

It goes through the following steps:

 

One. Nonlinear Fitting (Logistic or Probit or Weibull, etc.)

2. Estimating the Log(Dilution) value at which the percent mortality is 50% through inverse estimation

3. Transform to return the log transformation value.

 

in the order above

 

One. Nonlinear Fitting

 

undefined

Specify Percent Mortality for Y response and Log(Dilution) for X Predictor Formula.

undefined

Press the Hot Spot (Red Triangle) on the left side of the Fit Curve and select Sigmoid Curve >> Logistic Curves >> Fit Logistic 3P. Alternatively, fitting is possible with Probit or Weibull.

 

2.Log (Dilution) value estimation that satisfies 50% percent mortality through inverse estimation

 

undefined

Click the Hot Spot on the left side of Logistic 3P and select Custom Inverse Prediction

 

undefined

 

Enter the desired value (50%) in the red box above.

 

undefined

 

This gives you the desired value of 1.447876.

Using this to get 10^(1.447876) for the next step 3, the conversion, we get 28.05.

 

in other words, 1:28 This becomes the Dilution endpoint where the Percent Mortality becomes 50%.

This post originally written in Korean and has been translated for your convenience. When you reply, it will also be translated back to Korean.

Comments