Terminology Flashcards
Define representative sample
sample of the pop whose characteristics are representative of those in the larger pop
Define random sample
sample not representative of the pop
Define incidence
new cases
is a rate
units of person-years or per person per year
Define prevalence
total number of cases
is a proportion
has no units
Define ratio vs proportion
ratio = one number divided by another
proportion = fraction where the top part has to be part of the bottom
Define rate vs ratio
rate = number of events per population per time
ratio = comparison (rates or odds)
Define ratio of benefit vs ratio of harm
benefit = 1/ratio of harm
harm = 1/ratio of benefit
Define confounding
becomes an issue when 2 groups are being compared
influences the measure of association = misinterpretation
associated with both factors
Define bias
samples being studied are diff or are being treated diff
e.g. selection bias, info/measurement bias
Define 95% confidence interval
range in which we can be 95% certain that the true value of underlying tendency really lies
Define p-value
probability of getting an observation as extreme as, or more extreme than one observed assuming that the stated hypothesis is true
Define prospective study
watches for outcomes
cohort
Define retrospective study
looks backwards and examines exposures in relation to an outcome that is established at the start of the study
cohort
case-control
Define random selection
recruitment into a clinical trial and affects external validity
Define random allocation
allocation between treatment groups in a clinical trial and affects internal validity
Define as-treated
analysis according to what the trial participants did
this assesses the physiological effects of treatments studied
Define intention-to-treat
analysis according to the original allocation of the participants
this preserves the original random allocation, reduces bias and reflects what would happen in practise
Define forest plot
graphical representation of a meta-analysis
each study shown by varying sized square
horizontal lines runs through square to represent confidence interval
diamond symbol used to represent the overall estimate from the meta-analysis
Define funnel plot
used to check existence of publication bias in a systematic review
its assumes the spread of studies is symmetrical
variation from the assumption can indicate publication bias
Define homogeneity
null hypothesis = there is no variability between the studies in the systematic review
Define heterogeneity
statistical heterogeneity = variation as a result of the observed intervention effects, being more diff than we would expect due to chance alone
Define systematic review
used to appraise, summarise and communicate large quantities of research on a similar topic
qualitative or quantitative
Define meta-analysis
stat tool used to combine evidence and quantify the results
Define fixed effect
we assume that all the included studies share a common side effect
observed effects will be distributed about the common effect size, with a variance that depends primarily on the sample size for each study
Define random effect
rather than assuming that the studies share a common effect, we assign a distribution of effect sizes
Define publication bias
bias towards publishing studies according to whether they show statistically significant results
Define statistical significance
whether the p-value is less than or greater than 0.05
if we are looking at the diff between a new intervention and a comparison then a p-value less than 0.05 will result in a conclusion that the intervention is more effective than the comparison group
Define clinical significance
determined by the minimally clinically important difference between the 2 groups.