Statistics

Data types, outcome measures and epidemiology

Types of data, significance tests for associations, outcome measurement, incidence and prevalence.

16 questions 4 source pages

Images appear with the first question taken from each source page — tap a question to open it.

Q1What types of data do you know?▸

Categorical: nominal (e.g. eye colour) or ordinal (e.g. mild/moderate/severe)
Numerical: discrete or continuous

Q2What is a normal (parametric) distribution?▸

Q3How do you formally test for normality and homogeneity of variance?▸

Q4How can non-parametric data be transformed to parametric?▸

Q5How do you measure variability?▸

Q6What are the measures of central tendency?▸

Q7What is statistical inference used for?▸

Used to test specific hypotheses about associations or differences among groups of subjects
Applied to sample data

Q8The Harris hip score is shown - what type of outcome measurement is it?▸

Q9What is a PROM?▸

Q10What is a CBOM (clinician-based outcome measure)?▸

Q11How do you choose a scoring system?▸

Reliability: consistency of results on repeated measurement (Kappa for categorical data), internal consistency, reproducibility (intra/inter-observer)
Validity: does it measure what it is supposed to - construct/content/criterion/concurrent
Clinical utility: patient and clinician friendliness

Q12Define prevalence and incidence.▸

Prevalence: cross section measure of the people having the event / whole population
Incidence: number of new case in a period of time/ population at risk

Q13What population and disease factors make a good screening test (Wilson's criteria)?▸

Q14What test and economic factors make a good screening test?▸

A well-accepted test
High sensitivity essential, high specificity desirable
Cost of case-finding (diagnosis + treatment) economically balanced against overall medical care expenditure

Q15Why is scoliosis suitable for screening?▸

Known natural history with an early detectable phase (>20 degrees in Risser stage 1 or before high chance of progress)
Early treatment (bracing) may alter the course of disease
Test: Adam forward bending with scoliometer - well accepted
Sensitivity uncertain

Q16What are the four categories of Wilson's criteria for a screening test?▸

Population factor: important health problem with high enough prevalence
Disease factor: well-understood natural history, early/latent stage, accepted treatment, agreed policy
Test factor: well-accepted test - high sensitivity essential, high specificity desirable
Economic factor: cost of case-finding balanced against overall medical care expenditure

This site is private