### Step-wise Procedure to Calculate Screening Cut Point (Parametric and Robust Parametric Method) for Anti-Drug Antibody Assay

The cut-point of the assay is the level of response of the assay that defines the sample response
as positive or negative. Establishing the appropriate cut-point is critical to
minimizing the risk of false-negative results. The screening cut points are determined by assaying a set of approximately 50 drug-naive individual serum samples over three days by two analysts for a total of six runs. Here, we describe the stepwise procedure to calculate the statistical cut point using the parametric and robust parametric method.

*Steps for estimating Screening cut points*

1) Compile the data from 50 drug-naive individuals assayed by two analysts over 3 days for a total of six days. (See template below)

2) Compute the log transformation of the assay signal and NC-normalized values for the data sets.

*Shankar et al 2008 recommend stepwise evaluation of non transformed and transformed data for establishing cut points. Based on our experience, the NC-normalized data are mostly suitable for establishing the cut point because these data sets are less susceptible to plate-to-plate and day-to-day variability.*

3) Use the box plot analysis in JMP to identify and exclude outliers.

4) Use the Shapiro-Wilk Test for the test of homogeneity or normality. The data is normally distributed when

*the p-value*for the Shapiro-Wilk Test is less than 0.05.*Use the nonparametric method for the data that are not normally distributed*.5) Perform ANOVA analysis (Oneway ANOVA in JMP software) to compare the mean and variances among six runs.

6) Use the Levene test and F test for comparing means and variances among six runs. The means and variance were considered when P<0.05.

7) Use the

**Parametric method**for the data sets -when means and variances are equal variances or

-when the variances are equal but means are different.

The formula for parametric cut point with a 5% false-positive rate

*Screening Cut Point = Mean (Normalized) + 1.65 X Standard Deviation*8) Evaluate the cut point using

**Robust Parametric Method**for the data sets with different means and variances or high outliers

*Screening Cut Point = Median (Normalized) + 1.65 X (1.45 X Median Absolute Deviation)***Template for Data Complilation**

Cutpoint Run | Day | Analyst A/B | Plate NC-signal | %CV | Drug Naive Individual ID | Sample Signal(S) | Log (S) | S/NC |

**Reference**

Shankar et al 2008

FDA Guidance on ADA assay 2019