Lecture 1 - Study design Flashcards
What is unique to a descriptive study
no comparison groups
What does an observational study not have?
it has a comparison group but no intervention
What are the features of a clinical trial
- randomisation
- concealed allocation
- placebo
- blinding
- intention to treat
What is randomisation?
allocation of treatments to patients using some random process
What is the purpose of randomisation?
it will produce study groups which are comparable with respect to known and unknown factors which are likely to influence outcomes
What are the benefits of randomisation?
- observed difference in the two groups will be due to effects of treatment
- fair comparison between groups
- statistical analysis only valid if this takes place
What is a confounding factor
factors that influence treatment and outcome measures
-include demographic characteristics / prognostic factors
What is restricted randomisation and give an example
number of patients per treatment chosen in advance
- block randomisation
benefits of stratification?
useful for small sample sizes as simple random allocation would not mean risk factors would be distributed equally
what is the aim of a clinical trial
minimise bias and confounding
what is selection bias and is it removed
enrolling a certain patient is influenced by which treatment they might receive
-removed using sealed envelopes
What can knowledge of placebo effect do?
influence subject QoL/ pain levels
What is intention to treat?
analysing data if they received treatment or if they did not
Why is intention to treat useful?
accurately reflects ‘real life’
What is per protocol analysis
reported as secondary outcomes measures the effect on who actually received the treatment
What is the pinnacle of evidence-based medicine
- systematic reviews and meta-analyses
2. randomised controlled double blind studies
How can results from a trial of 2 treatment with 2 categories be summarised
2 x 2 frequency table
how is risk of failure calculated?
failures of A / failures of A + success of A
- failure / total of A
How is odds of failure calculated?
failure of A / success of A
define risk reduction
difference in risk between the two treatment groups
define number needed to treat
reciprocal of the risk reduction
- number of patients needed to be treated to prevent a single end point
what is the odds ratio
a/b divided by c/d
useful to describe relative difference between treatment groups
-logistic regression produces odds ration
What can’t be guaranteed by observational studies?
-other important factors causally related ti both disease and exposure
Can causation be inferred from an observational study
no due to bias and confounding variables
what is a cohort study?
- prospective in nature
- population subgroup is defined and followed up over time
- exposure status is measured at the beginning of the study
- disease status is observed and recorded during follow-up
what is a case-control study?
-retrospective
-subjects w disease (cases)
subjects wo disease (controls)
How do case-control studies draw their conclusions?
- frequency of exposure ascertained in each group
- higher prevalence of exposure among controls suggest exposure might be a risk factor
- selection /recall bias present
Cohort vs case-control
- cohort is more expensive and results can take a long time to materialise
- cohort is unsuitable for rare diseases
- cohort can confirm findings in case-control
what is the risk ratio
risk of the disease in the exposed group relative to the unexposed group
why must odds ratio be used in case-control studies
odds of disease cannot be calculated directly
what can a small sample size mean?
- lack of statistical power
- if there is a real difference you won’t detect it