Is this question worth taking to the IRB
Check the question holds before you start the paperwork.
Where it jams
Research starts from one question, but to see whether it holds you need data, and to see data you need an approved topic and a full procedure. Sometimes you only learn the question never held after all of it is done.
For residents it is worse. In-house data opens only through an advisor and review, and records at other hospitals are out of reach entirely. At the stage of picking a topic there is nothing to check against.
We solve only the step in front. Apply your criteria, see whether the cohort actually exists and the question stands — then start the procedure.
How long the records run
Not just counts — one person continues across years.
- 10.1 years
- mean span we follow one patientmedian 6.7 years
- 1,138
- patients with more than five years of record
- 610
- patients whose records span two or more hospitals
How this relates to ethics review
Checking a patient count does not require review. What we hand over is an aggregate, not patient-level data.
The procedures for conducting the study itself do not go away. Those follow your institution's rules and we do not handle them for you.
What we remove is not the procedure — it is the time spent stalled because the number was out of reach.
Common questions
- Can a research grant pay for this?
- Grant line items for data purchases are often ambiguous. In practice a sponsor funds the researcher and the work is paid from there. Tell us your situation and we will work it out.
- Can we try it free first?
- Patient counts need no account. Record excerpts need a work email once.
- Can the result go into a paper?
- The output carries the counting rule, the denominator and the limits, in a citable form. Submission and review responses are not ours to do.
- Can it be merged with our own institutional data?
- Not at patient level. Only aggregates leave our side.
