+ Reply to Thread
Results 1 to 6 of 6

Thread: Design Issue: Retrospective Cohort or Case Control?

  1. #1

    Lightbulb Design Issue: Retrospective Cohort or Case Control?




    Good Morning!
    I am currently trying to decide which approach to my research question is best. I am using a large data set from a state agency. The central hypothesis is that adults with drug susceptible pulmonary tuberculosis receiving directly observed therapy will have a faster culture conversion time and shorter completion of therapy time (comparison of mean time by group) over those on self-administered therapy.

    My research aims are:
    •Compare the impact of directly observed therapy (DOT) with self-administered therapy on culture conversion time and completion of therapy.
    •Determine the impact of directly observed therapy (DOT) and self-administered therapy on culture conversion time and completion of therapy while controlling for intervening variables (Human Immune Deficiency Virus, Diabetes, Place of Birth, Age, Race).

    There has been much debate to if this is a retrospective cohort or case control study. I would appreciate any feedback! Thank you for your time. Christina

  2. #2
    Pirate
    Points: 15,159, Level: 79
    Level completed: 62%, Points required for next Level: 191
    victorxstc's Avatar
    Posts
    875
    Thanks
    229
    Thanked 332 Times in 297 Posts

    Re: Design Issue: Retrospective Cohort or Case Control?

    Its interesting, since a similar question was the reason I joined this forum. In case-control setups, we split the sample according to the outcome, while in retrospective cohorts, we split it according to the exposure.

    The outcome in your case is healing time, while the exposure is the type of therapy. So your design is not case-control, when you are splitting your patients according to their therapy type. But you can still split them according to their healing time, and it becomes case-control.

    Is that true? I am afraid not fully. Your design is none of the above, in essence. If we disregard the retrospective nature of your study, we see that you are comparing treatments, which are "interventions". So basically you are experimenting on your patients [or comparing the effects of experimentations]. So it is a sort of clinical trial. I know well that a clinical trial has to be prospective. But what you are doing, even while it is retro, is neither cohort, nor case control, because there had been an intervention. The best study design it fits is clinical trial.

    Trust me on this, it took about 6 months of me to finally verify that.

  3. #3
    Omega Contributor
    Points: 38,406, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    7,002
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Design Issue: Retrospective Cohort or Case Control?

    Depends, was the use of two treatments part of a study or not? Did two treatments occur, now you are going back and comparing them? Or did two treatment occur per intention to conduct this study.

    Also, may want to watch out for confounding by indication, if treatments were not randomized.
    Stop cowardice, ban guns!

  4. #4
    RoboStataRaptor
    Points: 11,861, Level: 71
    Level completed: 53%, Points required for next Level: 189
    bukharin's Avatar
    Location
    Sydney, Australia
    Posts
    1,312
    Thanks
    12
    Thanked 326 Times in 315 Posts

    Re: Design Issue: Retrospective Cohort or Case Control?

    It's a retrospective cohort study (or more fully it's a retrospective cohort study of prospectively collected administrative data).

    "Clinical trial" is a loose term that doesn't describe a specific study type. However, almost nobody would call this a clinical trial - the term is used almost exclusively for randomised studies, and occasionally used (misleadingly) for prospective non-randomised intervention studies. Your study is neither of these, so shouldn't be called a clinical trial. I work in the areas of medicine, epidemiology, and evidence-based medicine; victorxstc works in dentistry - so it's possible that the term is used differently in his area. However, you're working in medical research so don't call it a "clinical trial" unless you want to get shot down by reviewers.

  5. #5
    Pirate
    Points: 15,159, Level: 79
    Level completed: 62%, Points required for next Level: 191
    victorxstc's Avatar
    Posts
    875
    Thanks
    229
    Thanked 332 Times in 297 Posts

    Re: Design Issue: Retrospective Cohort or Case Control?

    Quote Originally Posted by bukharin View Post
    It's a retrospective cohort study (or more fully it's a retrospective cohort study of prospectively collected administrative data).

    "Clinical trial" is a loose term that doesn't describe a specific study type. However, almost nobody would call this a clinical trial - the term is used almost exclusively for randomised studies, and occasionally used (misleadingly) for prospective non-randomised intervention studies. Your study is neither of these, so shouldn't be called a clinical trial. I work in the areas of medicine, epidemiology, and evidence-based medicine; victorxstc works in dentistry - so it's possible that the term is used differently in his area. However, you're working in medical research so don't call it a "clinical trial" unless you want to get shot down by reviewers.
    Hey thanks Bukharin

    I had the similar question once. Last year (or two years ago, god two years passed this fast?) I first thought such a design should be a retrospective cohort and discussed it with Link and you. Ultimately we (Link and me at least) agreed upon it being somehow a clinical trial. I don't know about your final opinion on that design though (not updated).

    I am aware of the exclusive meaning of clinical trial. Also the design definitions in dentistry and medicine are exactly the same (dentists too treat people, almost always by the means of medication and surgery). But I am somehow confident this design that assesses the effect of a Human Intervention (therapy) cannot be a retrospective cohort which deals with non-intervention exposures.
    "victor is the reviewer from hell" -Jake
    "victor is a machine! a publication machine!" -Vinux

  6. #6
    RoboStataRaptor
    Points: 11,861, Level: 71
    Level completed: 53%, Points required for next Level: 189
    bukharin's Avatar
    Location
    Sydney, Australia
    Posts
    1,312
    Thanks
    12
    Thanked 326 Times in 315 Posts

    Re: Design Issue: Retrospective Cohort or Case Control?


    A few points:
    - There is nothing in the definition of cohort studies that says the "exposure" can't be a human intervention. The key element of a cohort study is that you have a group of people who may have one or more exposures at baseline, and they are followed up over time and assessed for one or more outcomes of interest. This is what christinarn2000 has described, with the exposure being DOT and the outcomes being culture conversion time and completion of therapy
    - Having said that, I can't find any universally agreed definitions for study designs in general - so you are free to disagree with me, and we probably won't reach consensus since there doesn't seem to be any citation that would "prove" anyone right
    - Your study cited in your original link is a cross-sectional study. I am 100% sure of this. I did not reply to that thread further because I had already clearly stated my views and did not see any value in just repeating them (I didn't want to be like this guy!)

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats