In the Dechallenge/Rechallenge analysis that is included in the Strategus Characterization module, a condition should be added to restrict the analysis to a single observation period. In a recent analysis using Strategus, I had a single patient appear in the "challenge fails" plot which showed an initial period of exposure to the drug-of-interest with an event occurring during exposure time, after which the patient appeared to discontinue the drug-of-interest. This apparent discontinuation was followed by a long (multi-year) gap with no observation of exposure or events. After several years, the plot showed the patient re-initiating the drug (rechallenge) at which point several events occurred during time-on-treatment. Thus, this patient appears to be a compelling example of initial treatment with an event, a dechallenge success followed by a rechallenge fail. However, I inspected the patient profile for this person, I realized that for the entire period of time between discontinuing the initial treatment with the drug-of-interest and re-initiating the drug of interest the patient was not enrolled in the database.
As a solution to this problem, I would advise either 1) restricting the dechallenge/rechallenge analysis to a single observation period or 2) allow multiple observation periods as long as the gap between them is below some threshold (e.g. 30 days).
I am happy to share specific plots and data points that informed the above issue submission. If that would be helpful please reach out to me at mconove1@its.jnj.com.
In the Dechallenge/Rechallenge analysis that is included in the Strategus Characterization module, a condition should be added to restrict the analysis to a single observation period. In a recent analysis using Strategus, I had a single patient appear in the "challenge fails" plot which showed an initial period of exposure to the drug-of-interest with an event occurring during exposure time, after which the patient appeared to discontinue the drug-of-interest. This apparent discontinuation was followed by a long (multi-year) gap with no observation of exposure or events. After several years, the plot showed the patient re-initiating the drug (rechallenge) at which point several events occurred during time-on-treatment. Thus, this patient appears to be a compelling example of initial treatment with an event, a dechallenge success followed by a rechallenge fail. However, I inspected the patient profile for this person, I realized that for the entire period of time between discontinuing the initial treatment with the drug-of-interest and re-initiating the drug of interest the patient was not enrolled in the database.
As a solution to this problem, I would advise either 1) restricting the dechallenge/rechallenge analysis to a single observation period or 2) allow multiple observation periods as long as the gap between them is below some threshold (e.g. 30 days).
I am happy to share specific plots and data points that informed the above issue submission. If that would be helpful please reach out to me at mconove1@its.jnj.com.