• Follow us on Twitter - #MDMLG
  • Midwest Chapter MLA / MHSLA 2017 Joint Conference
  • Announcements & Information
    Contributors
    Syndication
    Atom Feed

    Subscribe with Bloglines

    MDMLG is a regional member of
    MHSLA logo
    Credits
    MDMLG Webmaster: Valerie Reid
    Hosted by: Provide Net

    Powered by:
    Powered by Blogger

    Creative Commons License
    This work is licensed under a
    Creative Common License

    Wednesday, October 18, 2006
    The hard part of information mastery
    Evaluating information:

    particpant allocation - one of 7 types of blinding, one of two absolute types of blinding.

    Why is allocation concealment important?

    Could change the outcome of the results.

    Mamography screening
    Lancet Jan 8, 2000; Oct 20, 2001

    "Mundus Vult Decipi"
    "The world wished to be deceived"
    People would rather be deceived than have the truth cause anxiety

    Caleb Carr, "Killing Time"

    Second absolute type of blinding - judicial assessor blind.

    Also look for intention to treat

    Levels of Evidence:

    SR with homogeneity = 1a
    RCT: LOE = 1b
    Cohort: LOE = 2b
    Case Control: LOE = 3b
    Case Series: LOE = 4
    Expert Opinion: LOE = 5

    Statistics you need to understand to evaluate research:

    Probability Level (P-Value) - likelihood that the deifference observed between two interventions could have arisen by chance.

    Number Needed to Treat - 1) the number of patients that need to be treated for one additional patient to receive benefit, 2) the number of paitents that need to be treated to prevent one additional outcome, 3) takes into account the relative risk as well as the absulte risk of not treatment.

    NNT = 100 / % in treatment group - % in control group

    Relative risk - risk of harm (or benefit) of one treatment as compared with another. Does not take into account the risk of NO treatment (absolute risk).

    relative risk tells part, but not all of the story; NNT does better

    confidence interval - upper and lower possibilities of our statistical estimates. If CI crosses 1.0, the difference is not significant.

    Statistical signiificance is a requirement for determining clinical significance, but is not enough to signify a clinical difference.

    Confidence Intervals helop us to understand how close our answer is to the truth.

    What to look for in validity:
    randomization
    allocation concealment
    blinded judicial assessors
    intention to treat
    follow-up > 80%
    Narrow confidence intervals

    Time for lunch...

    Labels:

    Posted by Alexia Estabrook @ 10:54 AM   0 comments

    0 Comments:

    Post a Comment

    << Home