1. Forum
  2. Usenet
  3. SCI.STAT.MATH

  • Q set margin for non-inferiority and or types of test

    From Cosine@21:1/5 to All on Mon Aug 31 09:09:26 2020
    Hi:

    In conducting a clinical trial for testing the relationship between the new object (drugs, devices, or medical procedures) and the standard/old object, we need to define first the margin indicating the magnitude of the effect observed has clinical
    effects, e.g., the drug has effects.

    But how do we properly define this margin? For a drug trial, at least we could set this margin to be larger than the effect observed when a placebo is given. Nevertheless, we do not have something equivalent to placebo if we are conducting a trial for
    new surgical procedures.

    Thus. are there general guidelines for determining the value for the margin?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rich Ulrich@21:1/5 to All on Mon Aug 31 15:00:48 2020
    On Mon, 31 Aug 2020 09:09:26 -0700 (PDT), Cosine <[email protected]>
    wrote:

    Hi:

    In conducting a clinical trial for testing the relationship between the new object (drugs, devices, or medical procedures) and the standard/old object, we need to define first the margin indicating the magnitude of the effect observed has clinical
    effects, e.g., the drug has effects.

    But how do we properly define this margin? For a drug trial, at least we could set this margin to be larger than the effect observed when a placebo is given. Nevertheless, we do not have something equivalent to placebo if we are conducting a trial for
    new surgical procedures.

    Thus. are there general guidelines for determining the value for the margin?

    Before you do a trial, you have on hand "pilot data" of
    some sort, which convinces you that there can be a
    "useful" outcome that will outweigh potential side effects.

    Or you have anecdotal reports from (rogue?) physicians.

    In any case, research institutions that seek outside funds
    will have a board that conducts a review of ethical concerns.
    Board members will have their own biases, presumably
    tempered by prior experience.

    When my own Psychiatric Institute first branched out to
    start doing major research, there was a surgeon (same
    hospital system) who refused to approve ANY application
    from us, it being his belief that all of psychiatry was BS, thus,
    no research was justifiable (early 1970s; his opinion was
    only a little out of date). He was shortly bullied to either
    modify his assessments or step down.

    --
    Rich Ulrich

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Cosine@21:1/5 to All on Mon Aug 31 15:12:13 2020
    Well, indeed we could learn something from prior studies. While this is conceptually true, it is difficult to apply this in an actual situation.

    It would be clearer if an illustration showing a practical situation would be available. For example, we want to compare the ability of a new method of imaging for identifying something like a tumor with the ability of the standard method like an X-ray.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David Duffy@21:1/5 to Cosine on Tue Sep 1 03:51:01 2020
    Cosine <[email protected]> wrote:
    Well, indeed we could learn something from prior studies. While
    this is conceptually true, it is difficult to apply this in an actual situation.
    It would be clearer if an illustration showing a practical situation
    would be available. For example, we want to compare the ability of a
    new method of imaging for identifying something like a tumor with the
    ability of the standard method like an X-ray.

    You mean like
    https://pubmed.ncbi.nlm.nih.gov/26178041/

    For clinical tests, the whole apparatus of cost-benefit analysis gets
    applied regarding what size difference in accuracy is important
    enough - consider two-stage screening designs for colon cancer -
    cheap inaccurate fecal occult blood test followed up expensive
    colonoscopy as gold standard.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)