Template Clinical Study Reporting repository

This project template contains the folder structure and templated code (ADaM & TFL) to populate clinical study project.

This repo project is copied by the Clinical Programming team to instantiate a new study specfic reporting effort GitHub repo and Domino Project.

Each reporting effort for a study will be a new branch from this single overall clinical study repository.

Directory structure

The programming is created in a typical clinical trial folder structure, where the production (prod) and qc programs have independent directory trees.

Reporting effort level standard code (e.g. SAS macros) should be stored in the share/macros folder.

The global domino.sas autoexec progam is also included in the repository to appropriately set up the SAS environment.

repo
│   domino.sas
├───prod
│   ├───adam
    ├───adam_flows
    ├───tfl
│   └───tfl_flows
├───qc
│   ├───adam
│   │       compare_adam.sas
    ├───adam_flows
    ├───tfl
│   └───tfl_flows
├───utilities
│       init_datasets_re.py
│       import_metadata.sas
├───flows
│       flow_1.py
│       flow_2.py
│       flow_3.py
│       flow_4.py
│       example_flow.py
└───share
    └───macros

Setup

1. Create a new project, named CDISC01_RE_YOURNAME, from the project template. This will create a new project and a new study specfic GitHub repo.

2. Run utilities/dataset_init_re.py as a job to create the appropriate analysis domino datasets (ADAM, TFL, ADAMQC, TFLQC and METADATA). As well as import SDTM datasets from an existing project following the same naming convention (CDISC01_SDTM for CDISC01_RE_XXXXX).

3. Add the external data volume (EDV) metadata-repository to your project.

   a. Ask your Domino contact on how to set up this example EDV within your Domino deployment.

4. Set the DCUTDTC project environment variable to the same value as the latest SDTM Snapshot e.g. MAR122024.

5. Set the SDTM_DATASET project environment variable to SDTMBLIND.

6. Run utilities/import_metadata.sas as a job (on the SAS environment!) to move and transform the metadata Excel file stored in the metadata-repository EDV to sas7bdat files in your local METADATA project dataset.

7. Run each of your prod ADaM and TFL programs as seperate jobs to produce your outputs.

8. To run all the ADaM and TFL programs using the multijob.py utility, first you need to add your SAS enviroment ID pipeplines/jobs_example.cfg.

9. Once added, input the following into your job command to run the full, dependancy aware study pipeline pipelines/multijob.py pipelines/jobs_example.cfg

   a. Ensure you select a Python enviroment to run this pipeline head jobs within.

Naming convention

The programs follow a typical clinical trial naming convention, where the ADaM programs are named using the dataset name (e.g. ADSL.sas, etc.) and the TFL programs have a t_ prefix to indicate tables, etc.

QC programming and reporting

The QC programming is all in SAS, and there is a compare.sas program which uses SAS PROC COMPARE to create a summary report of all differences between the prod and qc datasets. This program also generates the dominostats.json files which Domino uses to display a dashboard in the jobs screen.

compare.sas references a read-only SAS macro stored in the SCE_STANDARD_LIB repo so ensure this is imported as a secondary repo in order to run it.

Support

Programming was created by Veramed Ltd. on behalf of Domino Data Lab, Inc.