Health Stats Definitions

fusimore's version from 2016-02-16 21:16

Section 1

Effect SizeThe magnitude of difference between two groups, regarding treatment effect (Cohen's d: 0.8 LARGE, 0.5 MEDIUM, 0.2 SMALL)
CausationCondition precedes a disease such that had it been different in a specific way, disease would not have happened
Standard DeviationThe absolute value of the average difference of an individual value from a mean (good for mathematical manipulation, ONLY normal distributions)
Coefficient of Determination (r squared)The amount of variance in one measure (y) that is explained by another measure (x) (relative, can be underestimated if data is not linear)
Systematic ReviewClearly formed question with explicit methods that identify, select and appraise relevant literature
Type II ErrorProbability of saying the new treatment is not effective, when it is (difficult to calculate, unlikely to report it: -ve study)
Correlation CoefficientStrength of association between two variables
SensitivityOf the people that have the disease, the percent that will test positive (important for ruling out something sinister -- high false positives)
p-valueRepresents risk of Type I error: the chance that you will mistakenly reject the null; the probability that the result is due to chance alone (usually p<0.05, many tests p<0.001)
Evidence Based Medicine (EBM)The conscientious, explicit, and judicious use of current best evidence
Statistical SignificanceIf a test is statistically significant, the results are unlikely to be due to chance.
ReliabilityOverall consistency of a measure with repeated attempts (intra-rater, inter-rater)
Negative Predictive Value (NPV)How often a negative finding is correct (should be high if prevalence is low -- PREVALENCE DEPENDENT!)
Minimal Clinically Important Difference (MCID)Takes into account error of instrument and clinical variance
Standard Error of the MeanThe standard deviation of all sample means (of a given size) drawn from a population
Standard ErrorEstimate of population variance
Sample SizeDepends on 1) Effect Size 2) Type I error 3) Type II error 4) Population size
UtilityIs it useful? Does it make sense? Will you use it?
SpecificityOf all of the people who do not have the disease, the percent that will test negative (used for: emotionally or financially burdensome results)
AccuracyHow close a measure is to the true/actual value (must be both accurate and reliable)
EpidemicConcentration of new cases in time in a region/country
PandemicConcentration of new cases in time in a large region/country
EndemicAn infection/disease that is maintained in a population
Positive Predictive Value (PPV)How often a positive finding is correct (should be high if prevalence is high! PREVALENCE DEPENDENT)
Type I ErrorThe probability of saying there is a difference in treatment effects when there is not.
Sufficient CauseA minimum set of conditions for an outcome to occur
Quality of CareThe degree to which health services increase likelihood of desired health outcomes
Confidence IntervalDescribes the uncertainty inherent in a point estimate and a range of values you can be certain (within reason) that the true effect lies (95% CI), to narrow CI, increase n
ValidityThe degree to which the tool measures what they intend it to measure (content, criteria, construct, face)
Odds RatioA measure of association between an exposure and an outcome (the odds that an outcome will occur given a particular exposure (vs no exposure)