OHDSI Home | Forums | Wiki | Github

Are there any data files from the HowOften or Flouroquinolone (or other) studies publicly available?

General
#1

I’m currently setting up a study modeled after the HowOften study from the 2023 gloabal symposium using Strategus and the HowOftenApp for viewing the data.

It would be useful to have some known output to use for testing and to validate output from my study (for example: Are all of the data files being generated? Are all of the data files that should have data populated? What screens in the HowOftenApp should be populated and have useful information? etc.).

  • Are the data files from the HowOften study availble to be used as testing and validation tools (I’m particularly interested in the “azza” study as this is the one I’ve been using as a starting-point/template)?
  • Same question for the Flouroquinolone study from the SOS series.

Specifically, I’m looking for the .csv files that are generated by running the StrategusCodeToRun.R script in the HowOften project.

More specifically, my ci_incidence_summary.csv file has zero rows and I’m thinking I’ve misconfigured my study somewhere along the way that none of my subjects (in either the treatment or control group) are falling into the outcome cohort.

image
image1307×402 32.5 KB

Thanks,
John

#2

If you go into your results folder CohortGeneratorModule_1, you will find a file of cohort counts. If you’ve only run the azza study, you should see rows for the azza cohort IDs in there, and if everything is zero, then you had no people to study.

If you have some people in cohorts, you can look into your ci results again and see what it says for the ‘persons at risk’, which is the number of people that had time at risk in your study population. if there were 0, then you won’t get any rates.

If there are data files from your own CDM, then I think you would be able to use it for some sort of test/validation of your own local env. If you’re asking for data files from other sources, then that would be up to the other source maintainers if they want to share it directly with you, although i’m not sure what having a results file would do for you to test your own system…it’s just different data, so what would you be validating?

The HowOften analysis was performed on about 17 databases, and most did return results. There’s not a lot of configuration that goes into your execution settings: database connection info, and folders that store the results. So unless you have different cdms where one might be empty and you accidently pointed to the wrong cdm, my guess is that there’s some ETL problem in your CDM that leads to your issue (a common one is how OBSERVATION_PERIOD table is constructed).

t