Study Design and Summarising Data Flashcards
What is a randomised controlled trial?
A choice or treatment allocated to a group of people randomly and observed
Why do you randomise RCT’s?
Ensures unbiased treatment
Treatment groups balanced
Ensures patient characteristics don’t affect treatment
Outcome differences can then be attributed to treatment received
What are the types of blinding in RCT’s?
Where the treatment is concealed and it reduces bias
Double blind is where both patient and assessor are blind to what is being treated and who has what
What are case control studies?
Observation study
Where you don’t intervene and observe subjects in a natural world
Investigates causes it factors associated with disease or conditions
What are cases and controls?
Cases - groups with disease
Controls - comparator group without disease
What are the limitations of case control studies?
Choice of control group effects comparison
Data retrospective so can be biased
Often quick and inexpensive
What are cohort studies?
Observation study
Observing and not intervening in real world
Investigates causes or factors associated with disease
Selects groups of healthy individuals and follows up over time to monitor disease state or risk factors
Usually perspective
What are cross sectional studies?
Observational study
Collect data from subject at one point in time
Useful for measuring prevalence of disease or attitude or behaviour
Why summarise data?
Monitor data quality, check for invalid entries and to describe characteristics of participants in study before complex analysis
What is quantitative data?
Any data that can be measured
Continuous or discrete
What is continuous and discrete data?
Continuous lies on a continuum so it’s any value between a range like height
Discrete is data that can only take certain values like integers e.g children in family
What is categorical data?
Where individuals fall into one of a number of separate catergories
How to summarise continuous data?
If looking at the centre of data use mean or median
If looking at spread of data use range and standard deviation
What are the two types of categorical data?
Unordered categories (nominal data) - cannot be ordered or measured, used to label variables without quantitative value Ordered categories (ordinal data) - where variables have natural ordered categories
How to summarise categorical data?
Unordered categories - frequencies in each category, proportion or percentage, don’t clutter with too many decimal places
Ordered categories - frequencies in each categories, proportion or percentage, cumulative proportions
