• Statistics
    • Sensitivity
      • Ability to correctly identify those with the disease
      • TP/(TP+FN)
    • Specificity
      • Ability to identify those without the disease
      • TN/(TN+FP)
    • PPV
      • How likely is it that this patient who tests positive has the disease?
      • TP/(TP+FP)
    • NPV
      • How likely is it that a patient who tests negative does not have the disease?
      • TN/(TN+FN)
    • Types of data
      • Interval
        • Continuous and quantitative eg weight, height
      • Categorical
        • Nominal
          • Eg gender
        • Ordinal
          • Eg VRS for pain
          • Have an implied 'order' to data
    • Correlations
      • If normally distributed
        • Pearson correlation coefficient
      • Not normally distributed
        • Spearman or Kendal coefficient
      • If comparing 2 methods of measuring the same variable then a Bland-Altman plot may be used
    • Errors
      • Type 1 error (α)
        • A difference is found where one really didn't exist
      • Type 2 error (β)
        • Found no difference when a difference actually exists
        • Usually due to inadequate sample size
        • Power = 1-β
        • Should be 80%
    • Confidence intervals
      • A range of sample data that will contain an unknown population parameter (eg the mean) 95% of the time
    • Relative risk
      • Ratio of observed risk after first intervention divided by the observed risk after the second intervention
    • Normal distribution
      • Can be defined by the SD and the mean alone
      • Also known as parametric data
      • 2/3 of the data lies within 1SD of the mean, 95% within 2SD and 99.7% within 3SD of the mean
      • Variance = SD²
      • SEM = SD/√'n'
        • It is used to describe the precision of the sample mean versus the population mean
    • Measurement of spread
      • Range
        • Upper and lower values
      • Percentiles
        • Quartiles commonly used ie lower (25th percentile), middle (50th percentile) and upper (75th percentile)
      • The inter quartile range (IQR) is often quoted if the data is not normally distributed
    • Central tendency
      • Mode
        • Most common or frequent observation
      • Median
        • Central datum when all the data are ranked in numerical order
      • Mean
        • Adding all the numbers in the dataset and then dividing by the number
    • CEACCP
      Link:ceaccp.oxfordjournals.org/content/7/5/167.full.pdf
    • CEACCP
      Link:ceaccp.oxfordjournals.org/content/7/4/127.full.pdf
    • CEACCP
      Link:ceaccp.oxfordjournals.org/content/7/3/95.full.pdf
    • CEACCP
      Link:ceaccp.oxfordjournals.org/content/7/6/208.full.pdf
    • CEACCP
      Link:ceaccp.oxfordjournals.org/content/8/4/143.full.pdf
    • CEACCP
      Link:ceaccp.oxfordjournals.org/content/8/6/221.full.pdf