Conversation
| scenarios.append( | ||
| Scenario(name=f"Blood_{scenario}", params=get_default_params()) | ||
| .transform(transform_initial_compliance(screening_rate)) | ||
| .transform(transform_lesion_risk_alpha(1.19)) | ||
| ) | ||
|
|
||
| scenarios.append( | ||
| Scenario(name=f"FIT_Blood_TestCombiningMethod.PARALLEL_{scenario}", params=get_default_params()) | ||
| .transform(transform_initial_compliance(screening_rate)) | ||
| .transform(transform_lesion_risk_alpha(1.19)) | ||
| ) | ||
|
|
||
| scenarios.append( | ||
| Scenario(name=f"FIT_Blood_TestCombiningMethod.SERIAL_{scenario}", params=get_default_params()) | ||
| .transform(transform_initial_compliance(screening_rate)) | ||
| .transform(transform_lesion_risk_alpha(1.19)) | ||
| ) |
There was a problem hiding this comment.
After we modify the test combining script to set the proportion for combined tests to 0, the base params will assign FIT to all agents. So these tests also need to use transform_routine_test_proportion to set the proportion for the relevant test to 1 and all others to 0 (same pattern as Colonoscopy).
|
This looks good, Kate! Reviewing the combined test parameters highlighted a couple of changes that I want to make to the test combining script before we proceed.
I will make these changes in the I also added one comment to the Finally, we need to get your development environment set up to avoid the lint-and-format check failure. I will talk you through that during our next meeting. |
|
Oh, one last thing - you'll see I changed the base branch from |
|
@korourke5 I pushed my changes to test combining to |
template' into exp-test-combining
the most recent versions of FIT_Blood_parallel and FIT_Blood_serial from updated combine_tests.py
|
@sandypreiss I have updated the branch with the changes you mentioned above. |
|
@sandypreiss I have updated with the latest run. I wasn't sure exactly which results we were looking for so I have not yet updated the README. |
|
Bottom line is that the results look as they ought to with the combined test. FIT + blood in parallel leads to less CRC and more diagnostic tests (ie, positive screenings) than either test alone. FIT + blood in serial leads to more CRC and fewer diagnostic tests than either test alone. You can copy that exactly for the readme. Also, please delete the |
|
@sandypreiss This PR should be all set for closing out. |
2 participants
CRCsim PR Template
Description
Created new branch for test combining experiment combining FIT and Blood tests in serial and parallel. Changes:
PR Type
Experiment review checklist
exp-prefixcrcsimcommit hash (typically, the latest commit in themainbranch). This ensures that the experiment uses the latest version of the model. Note: used experiment-templateparameters.jsonfileprepare.pyscript applies incidence rate ratio (IRR) adjustment if appropriate for this experimentprepare.pyand:scenarios/directory (eg, uncommitted local changes from a previous experiment)crcsim/experiment/README.mdwith detailed information about the experiment's goals, scenarios, and corresponding AWS objectsExperiment review process
All experiments must follow each step of this review process.
prepare.py, and spot-checking scenarioscrcsim/experiment/summarize.py)crcsim/experiment/summary/summarized.xlsx)Note that experiment PRs are never merged into
main! The PR is closed, and the experiment is maintained as a separate branch. That's why we keep all experiment PRs as drafts.