Research Methodology 3: RCTs Flashcards
Define a RCT
RCT: a study in which participants are allocated randomly between an intervention
(e.g. treatment) and a control group (e.g. no treatment or standard treatment)
why are trials conducted?
- Safety
• Ascertain the safe dose of a
new drug. - Efficacy/ Effectiveness
• Demonstrate efficacy of new
drug
why do we randomise ?
When looking at cause-effect relationships, randomisation allows all random factors (confounders) apart from the proposed cause to be held constant between groups.
define confounder
A confounder is a variable that influences both the dependent
variable and independent variable causing a spurious association.
what does randomisation do
• Randomisation ensures all potential confounding variables (known
and unknown) will be distributed by chance across all study groups
Define equipoise
Clinical equipoise means that there is genuine uncertainty in the expert
medical community over whether one treatment will be more beneficial
than another.
what makes a good RCT?
Internal validity – is the IV causing the DV in this study?
External validity – to what extent can these findings be generalised to
other people, situations and times?
Define bias
In statistics - ‘a tendency of an estimate to deviate in one direction
from a true value’
list 4 types of bias
selection
performance
attrition
observer/detection
define selection bias
Systematic differences between baseline characteristics of groups that
are compared
Study sample does not represent target population
define performance bias
Systematic differences between groups in the care that is provided, or in
exposure to factors other than the interventions of interest
define attrition bias
Systematic differences between groups in withdrawals from a study
• Can cause systematic differences between groups – e.g. one
treatment has more side effects than another.
define observer/detection bias
Outcome measure does not adequately capture outcome of interest
• Systematic differences between groups in how outcomes are determined
Not adequately capturing the outcome of interest
Define intention to treat
– analysed in treatment group they were
randomised to, whatever happens later.
Define on-treatment analysis or per protocol analysis
only analyse
patients who finish the treatment according to the study protocol.
how many types of trials can you blind and describe who is blinded in each?
open: no one
single: patient is blinded
souble: pt and dr blinded
triple: pt/dr/investigators
what is the significance level?
0.05
what is the power of the study accepted at ?
80-90%
If i observe a difference and there is one what is this?
well powered
if i observe no difference and there isn’t one what is this ?
well designed
what is a type 1 error
observe a diff but no diff
what is a type 2 error
observe no diff but there is a diff
define the p value
The P-value is the probability of observing the result you got, or more
extreme, if the null hypothesis were true,
where the result is a statistic estimated on the data you have collected in your
study.
why do we calculate sample size?
Too few participants... If you get a null result, you don’t know if you have: • Evidence of no effect • or simply… • No evidence of an effect too many = not ethical
what does parametric data assume
normally distributed
define positive and negative skew
look at the tables
what do you use if your data is skewed? descriptive analysis
median and IQR
if data is normally distributed which statistical tests would you use?
t test (scale) chi sqaured (categories) anova (scale)
what assumptions are made for a t test
random sample
independent observation
normally distributed
variances for each group are equal
what test is used if data is non-parametric
Mann-Whitney U/Wilcoxon Rank Sum replace t-test
