Data and analysis
What kind of data are available in orthopaedics ?
1.Non-continuous, qualitative or categorical. Gu 1, 2, 3 -Analysed using Chi-squared or Fischer's exact test
2.Continuous or variable data e.g. height, age, knee valgus angle
These data are often not always -Gaussian distribution e.g. symmetrical bell-shaped curve.
Continuous data described using Mean, Median and Mode.
With Normal distribution, mean/median and mode are equal.
What is dispersion ?
It is a variability of a data set.If all values are zero, dispersion is zero.
Measured using Quartiles, SD and Variance.
Variance- how much it deviates from mean.
SD- square root of variance (to give same original dimension as data )
Standard error of mean (SEM) - measures how closely sample mean of data set approximates to population mean.
SEM = SD / √n
In a normal distribution data - comparison done using parametric tests such as student T-test.
Non normal distribution, student's T-test is used.
Sensitivity
ability of a test to detect cases that are positive.
all test with positive results/all cases truly positive * 100
Sensitivity
ability of a test to detect cases that are negative
all negative tests results/cases that are negative *100
What factors are need to be looked for in outcome scoring system ?
accuracy = how accurate it is compared with the gold standard.
validity = extent to which an experiment value represents a true value
reliability = ability to repeat the test
ease of use = assessment method should be appropriate and not too long ,complex or cumbersome.
Comments
Post a Comment