Bayesian Efficacy Monitoring Via Predictive Probability

J. Jack Lee, Ying-Wei Kuo, Diane Liu and Nan Chen

Department of Biostatistics, MD Anderson Cancer Center, Houston, TX 77030

PID:901; v1.1.4.0 ; Last Updated: 10/19/2020

Calculate Stopping Boundaries

Trial Name (Optional)

Design Parameters:

Probability of Response ( θ )

Declare efficacious if

\(\textit{PP} = \) \(\sum\) \(\bigl\{ \) \(\textit{Prob(future data) * [Prob}\) (\(\theta > \theta_{0}\) \(\textit{|current and future data}\)) \(\ge P_{T}\)\(\textit{]}\) \(\bigr\} \)

Reference response rate (\(\theta_{0}\))

Threshold for declaring efficacy at the end of the trial (\(P_{T}\))

Early Stopping for Futility: \(\textit{PP <}\) \(P_{L}\)

\(P_{L}\)

Early Stopping for Efficacy: \(\textit{PP }\) \(\ge P_{U}\)

\(P_{U}\)

Prior distribution for \(\theta\): \(\textit{Beta}\)(\(a_0, b_0\))

\(a_0\):

\(b_0\):

Sample size and cohort size:

Input Cohorts Manually

Maximum number of patients in the trial

Minimum number of patients before early stopping rule applies

Cohort size

Cohorts for interim analysis: (separate by a comma)

Upload Input file (.csv format): [Sample input file]

Browse...

Save Input into a csv file

Filename

Download PDF Report
Download EXCEL Report

Operating Characteristics

Trial Name (Optional)

Scenarios:

(separate by a comma)

Response Rates (in ascending order):

Graph Parameters:

Auto Generated Graph Parameters

(separate by a comma)

Color:

Density:

Angle:

Line Type:

Filename

Save Input into a csv file

Download PDF Report Download Excel Report

Trial Monitoring

Trial Name (Optional)

Number of Patients

Seq#

Patient ID

Resp. Outcome

1

2

Save Input into a csv file

Filename

Download PDF Report Download Excel Report

Bayesian Efficacy Monitoring Via Predictive Probability

J. Jack Lee, Ying-Wei Kuo, Diane Liu and Nan Chen

Department of Biostatistics, MD Anderson Cancer Center, Houston, TX 77030

PID:901; v1.1.4.0 ; Last Updated: 10/19/2020

Calculate Stopping Boundaries

Design Parameters:

Probability of Response ( θ )

Declare efficacious if

\(\textit{PP} = \) \(\sum\) \(\bigl\{ \) \(\textit{Prob(future data) * [Prob}\) (\(\theta > \theta_{0}\) \(\textit{|current and future data}\)) \(\ge P_{T}\)\(\textit{]}\) \(\bigr\} \)

Reference response rate (\(\theta_{0}\))

Threshold for declaring efficacy at the end of the trial (\(P_{T}\))

Early Stopping for Futility: \(\textit{PP <}\) \(P_{L}\)

\(P_{L}\)

Early Stopping for Efficacy: \(\textit{PP }\) \(\ge P_{U}\)

\(P_{U}\)

Prior distribution for \(\theta\): \(\textit{Beta}\)(\(a_0, b_0\))

\(a_0\):

\(b_0\):

Sample size and cohort size:

Maximum number of patients in the trial

Minimum number of patients before early stopping rule applies

Cohort size

Cohorts for interim analysis: (separate by a comma)

Upload Input file (.csv format): [Sample input file]

Go to Operating Characteristics ⇒

Operating Characteristics

Scenarios:

(separate by a comma)

Response Rates (in ascending order):

Graph Parameters:

(separate by a comma)

Color:

Density:

Angle:

Line Type:

Trial Monitoring

Number of Patients

Seq#

Patient ID

Resp. Outcome

1

2

Support Document