Read the summary and the most important questions on Measurement in medicine a practical guide practical guides to biostatistics and epidemiology | 9780521133852 | Henrica C. W. de Vet; Caroline B. Terwee; Lidwine B. Mokkink; Dirk L. Knol

• 5 Reliability

  This is a preview. There are 21 more flashcards available for chapter 5
  Show more cards here

• What do reliability parameters mean?

  How well patients can be distinguished from each other.

• How is the measurement  error expressed in a graph?

  How far the dots are from the 45 degree line

• How is SD expressed in a graph?

  Reflected by the spread of values along the 45 degree line.

• When are raters included in the error variance?

  When they are considered a random sample of all possible raters, then this creates systematic differences between raters

• Why is in ICC agreement rater variance taken into account and in ICC consistency?

  In agreement variance of observers (O) is taken into account since we want to know differences in absolute agreement.
   
  In consistency we are just interested in ranking, so variance of observers is not taken into account.

• What are two ways of calculating icc in spss?

  Anova and then by hand
  Scale analysis-reliability analysis-> choose two way analysis

• What happens to the formula when you need to average for several measurements?

  Same formula is used with a small difference
  In consistency variance of error/number of measurements taken
   
  In agreement variance of observers+variance of residual/2  
   This results in more reliable measures since error becomes smaller

• Is pearsons R useful with the icc?

  Only when there are random errors not systematic

• How can SEM be calculated?

  Square root of variance in error(so only of residual, or in case of agreement of both observer and residual)
   
  Differences between standard deviations of the raters. Take the differences between the values of the raters and calculate the mean and standard deviation.
   SEM consistency is SD difference/sqrt(2)

• What is a bland and altman plot?

  x-axis mean score
  y-axis differences of score
    dashed line is mean systematic difference
   Two lines which represent the limits of agreement (dashed line±1,96x SD difference)
  
   Makes a plot in which systematic errors can be seen easily