Welcome to the PRO-CRM App

Summary:

This web application allows you to conduct the patient-reported outcomes continual reassessment method (PRO-CRM) to estimate the maximum tolerated dose (MTD) by incorporating both clinician and patient perspective, using the marginal approach as proposed in Lee, Lu, and Cheng 2020. PRO-CRM uses both toxicity outcomes captured by clinicians and reported by patients and allows different toxicity thresholds for these. The app can be used to obtain dose assignment based on available toxicity data and finial estimation for the MTD.

Disclaimer:

We do not take responsibility for any misuse or misinterpretation of this application or the results produced.

References:

[1]Lee SM, Lu X, Cheng B. Incorporating patient-reported outcomes in dose-finding clinical trials. Stat Med. 2020 Feb 10;39(3):310-325.

[2]Lee and Cheung (2011). Calibration of prior variance in the bayesian continual reassessment method, Statistics in Medicine; 30 (17): 2081-2089.

[3]Ken Cheung (2019). dfcrm: Dose-Finding by the Continual Reassessment Method. R package version 0.2-2.1.

Funding:

This app was supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through Columbia University Irving Medical Center CTSA Grant Number UL1TR001873. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Privacy & Security:

BERD web apps are hosted on Amazon Web Services’ (AWS) HIPAA compliant infrastructure. Data are not permanently stored.

In the Data Upload tab, the user will upload trial data similar to the format from the sample data set on the right. This data can be downloaded and used as a sample for PRO-CRM trial conduct. The data should be a CSV file and have the following format:

  • ID Variable: this variable is the ID of each participant patient, which will be used as a unique identifier.
  • Dose level Variable: The assigned dose level for each patient. This variable should be ordered and positive integer. Higher value indicates higher dose level with the lowest dose level equal to 1.
  • C_DLT Variable: Clinician reported outcome of DLT. It should be 1 or 0 only. 1 indicates the occurrence of DLT, 0 otherwise.
  • P_DLT Variable: Patient reported outcome of DLT. It should be 1 or 0 only. 1 indicates the occurrence of DLT, 0 otherwise.
  • Start Time Variable: A time variable in date format indicating the order of patient's recruitment. It should be the date of dose assignment, which is used to order patients.
  • Evaluable: Indicates if a patient is evaluable or not. It should be 1 if the patient is evaluable, and 0 otherwise.

Sample Trial Data Set

Download sample data

A statistician file specifying model parameters also must be uploaded. Download the template file on the right, to fill out trial details.

  • Target DLT rate: a pre-specified threshold for DLT rate. Both clinician target DLT rate and patient target DLT rate have to be specified.
  • Skeleton: the skeleton based on an empiric working model. Different skeletons should be specified for clinician and patient working models.
  • Number of dose levels: the number of designed dose levels.
  • Method: the method used to estimate the model parameter. Options for the Method are 'bayes' and 'mle'(case sensitive). If bayes is used, please specify prior variance of the model parameter as described below.

Sample Statistician Parameter File

Download parameter file

Summary of enrollment

Summary for patient and clinician outcomes by dose level

Prefix C indicates clinician; prefix P indicates patient

Estimated MTD/Next Assigned Dose

DLT plot:

The points on the graph represent toxicity result for the assigned dose level based on the start date. "Both" indicates the patient experienced both P-DLT and C-DLT; "P-DLT only" indicates the patient experienced P-DLT, but not C-DLT; "C-DLT only" indicates the patient experienced C-DLT, but not P-DLT; "No DLT" indicates the patient did not experience DLT.


Please cite this web application wherever used in published work:

BERD Apps Group. (2023) Trial Design & Conduct Apps [PRO-CRM]. Irving Institute for Clinical and Translational Research. Retrieved from https://www.irvinginstitute.columbia.edu/services/visualization-apps [Accessed 2025-04-16 ]