RM Flashcards
NNH number needed to harm
If 88, for every 88 people undergoing the intervention, one person will have a harmful outcome
Incidence
number of new cases commenting in a period of time in a given population
Prevalence
number of persons with a particular disease within a given time and population
Relative risk
The ratio of the probability of developing an outcome in the intervention group compared with the probability of developing an outcome in the control group. If it is 1, there is no difference and it could be due to chance.
Absolute risk
the probability of an event in the population
NNT Number needed to treat
the number of patients with a condition who must receive an intervention to prevent one bad outcome
Odds
ratio of probability of an event to the probability of a non-event
Index test
the test whose diagnostic accuracy is being measured against the gold standard
reference test
the gold standard - best reflects the truth
Sensitivity
The proportion of people with disease, who test positive
Specificity
The proportion of people without disease, who test negative
Pre-test probability
Background probability of having a condition before a diagnostic test is done
Post-test probability
Probability of having a condition given the result of a diagnostic test
Positive predictive value
the probability of having a condition, given a positive test result
Negative predictive value
the probability of NOT having a condition, given a negative test result
Likelihood ratio
A measure of the efficacy of a test, calculate the ratio of the true positives to the false positives (TP/FP)
Statistical power
the likelihood that a study will detect an effect when there is a true effect to be detected
Cross-sectional study - OBSERVATIONAL
examines characteristics of a population AT ONE POINT IN TIME
Cohort study - OBSERVATIONAL
Participants selected and grouped by either having or not having an exposure, and then seeing what happens over time.
Can be prospective of retrospective
Case-control study - OBSERVATIONAL
Participants selected and group by either having the disease or not and then their exposures looked at - always retrospective. For rare outcomes of interest. Most at risk of recall bias.
What does a confidence interval of 95% (2 - 10) mean?
you are 95% confident that the true result of the X lies between 2 and 10
