ncats / CTSA-Metrics

17 stars 16 forks source link

C1: Time constraints #24

Closed mgkahn closed 2 years ago

mgkahn commented 3 years ago

C1: Document header says "at any time" but spreadsheet says "in time frame". Should the search for a lab test be limited to those tests done within one-year/five-year windows or not?

Currently I am only limiting the cohort of patients or encounters by the one-year/five-year windows but I am NOT limiting the search for labs by the same time windows. Should I?

This would change the meaning of C1 to be: Patients seen in one year who have at least one lab test (from any coding system) that occurred in one year. My current logic is patients seen in one year who have at least one lab test (from any coding system) that occurred at any time. (limiting the "one-year" only to the patient cohort, not the lab test).

ams219 commented 3 years ago

Q1: It should be in the "time frame" one-year and five-year.

Q2: I'm guessing that we should be limiting the search by the same time windows. I'm not the final word.

mgkahn commented 3 years ago

The question is what is "it" in your answer: the patients/encounters in the time frame? The lab test in the time frame? Both?

Thinking as an investigator: I agree with you that the most useful answer is both -- for a study using patients seen in 2020 (cohort time restriction), how much 2020 lab data can I get from the database (lab domain time restriction).

ams219 commented 3 years ago

Whenever possible, time frames should be both one-year and five-years for the measures.

mim18 commented 3 years ago

Use the time frames to constrain both your numerator and denominator.