+ Reply to Thread
Results 1 to 3 of 3

Thread: Intention to treat analysis and estimating loss to follow up

  1. #1

    Smile Intention to treat analysis and estimating loss to follow up




    Hi guys,

    Doing my MSc in Clinical Trials and after looking at a research paper I am stumped about something...

    The authors have stated that the sample size needed for a detection of the primary outcome measure was 36 per group (parallel RCT) assuming a loss to follow up of 15%, with a 2 tailed α of 0.05 and a β of 0.2.
    The true loss to follow up was actually 35%... They used intention to treat analysis..
    Does this now make the study underpowered.. Although the original recruitment number was reached and all participants are being included in analysis??

    Quite new to this and would love the help
    Thank you so much

  2. #2
    Points: 12,976, Level: 74
    Level completed: 32%, Points required for next Level: 274

    Posts
    154
    Thanks
    10
    Thanked 2 Times in 2 Posts

    Re: Intention to treat analysis and estimating loss to follow up

    ITT takes all subjects that were randomized. However, if subjects are lost to follow-up, you may have missing values of the primary endpoint. From now on it depends on how you handle missing values. Basically, if you need n subjects to get your power, and you have less, for whatever reason, you are under-powered.

  3. The Following User Says Thank You to Yankel For This Useful Post:

    clinicaltrialsnewb (01-24-2016)

  4. #3
    Omega Contributor
    Points: 38,423, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    7,005
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Intention to treat analysis and estimating loss to follow up


    Side note, the sample size is based on a conjectured effect size. If for some reason the conjecture effect was smaller than the truth, then the study may be underpiwered for the hypothesis but still reveals statisticallly significant results (differences) .

    Also, the study needs to have nondifferential loss to be unbiased.
    Stop cowardice, ban guns!

  5. The Following User Says Thank You to hlsmith For This Useful Post:

    clinicaltrialsnewb (01-24-2016)

+ 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