Chapter 1 - Epidemiology and clinical analysis Flashcards
Define: type II error
Failure to reject null hypothesis when it is false
Beta = probability of making type II error
Level of evidence: systemic reviews of case-control studies with homogeneity
3a
Level of evidence: individual cohort studies
2b
Level of evidence: All or non trials
1c
Level of evidence: systematic reviews of cohort studies with homogeneity
2a
Define: Kaplan-Meier method
Each event recorded at time of occurrence without the need for fixed time frames
Definition: clinical equipoise
Situation in which clinical experts professionally disagree on the preferred treatment method
Name: weighted composite dependent variable graph used in meta analyses
Forest plot
Define: selection bias
Effect tested differs among patients who participate to those who do not
Can be due to physicians or patients (self-selection)
Level of evidence: clinical outcome studies
2c
Define: type I error
Rejection of null hypothesis when it is true
Alpha = probability of making type I error
Define: censorship variable
Data about the event of interest are unknown because of withdrawal or failure to follow up
Level of evidence: individual RCT with narrow confidence intervals
1b
Define: Bayesian inference
Probabilities conditional on observations and with degrees of uncertainty; the use of odds ratio reflects this
Level of evidence: case-series studies
4
Define: power
Probability of rejecting the null hypothesis when it is false
Ability to detect true difference
1 - beta
Define:
internal validity
statistical validity
construct validity
external validity
Internal validity: ability to support the causal effect being tested
Statistical validity: ability to generalize the effect to the larger population from the sample drawn
Construct validity: support the intended interpretation of the effect
External validity: support the generalization of results to external populations from which samples are not drawn
Define: VascuQol
Vascular quality of life questionnaire
5 domains: pain, symptoms, activity, social, emotional
Epidemiology is greek for
Upon - the people - study
Define: confounding factor
Second variable correlates with a primary independent variable and its associated dependent variable
Level of evidence: individual case-control studies
3b
Define: propensity scoring
Control confounding factors to reflect conditional probability of being in a treatment group based on known variables. Carefully chosen to reflect potential predictors of treatment assignment (not outcomes of assignment)
Define: information bias
Erroneous information collected by study because of poor design
Define: validity
Degree to which an instrument measures what it is intended to measure
= accuracy
Define: carryover effect
Influence of time and previous exposure of the other treatment
Definition: prevalence and incidence
Prevalence = ratio of persons affected for the population at risk
Incidence = ratio of persons in whom the disease develops within a specified period for the population at risk
(frequency of the disease at the measured time point regardless of the time of disease development)
Table: Grade of recommendation, recommendation and basis
List: Limitations of RTC’s
1) costly to conduct
2) lack of clinical equipoise
3) bias from physician or patient even with clinical equipoise
4) risk aversion or randomization aversion
5) intent to treat analysis with high crossover or dropout
6) conducted at high volume centers and therefore not representative
7) designed and powered to test only one hypothesis
Define: block randomization
Randomizing patients in a study in groups as they are enrolled
Define: SF-36
Short form health survey for measurement of health-related quality of life instrument
8 domains
physical: physical functioning, physical role, body pain, general health
mental: vitality, social functioning, emotional role, mental health
Define: recall bias
Cannot recall events accurately or have bias in recall
Define: Koch’s postulate (Robert Koch)
Postulate linking microbe and disease
Level of evidence: systematic reviews of RCT study with homogeneity
1a
Who is John Snow
Recognized for ameliorating cholera epidemic in 1854 London
Define: reliability
Consistency of the instrument if it were to be repeated on the same participant under identical conditions
= precision
Level of evidence: expert opinion without critical appraisal or based on bench research
5
Define: EQ-5D
European quality of life group instrument for quality of life assessment
5 dimentions of health: mobility, self-care, usual activities, pain/discomfort, anxiety/depression
Define: counterfactual ideal
Impracticality of conducting study to expose people to certain things
Define: WIQ
Walking impairment questionnaire
4 components: walking distance, walking speed, stair climbing, other symptoms that limib walking
Define: survival analysis
Event status between fixed periods of measurement
Statistical tests based on research question and data characteristics
1 group, 2 independent, 2 related, 3+ independent, 3+ related, 2 comparable variable, 1 dependent variable and 2+ independent variables
One group parametric: one sample t-test
nonparametric: sign test, Wilcoxon signed rank, transform data for t-test
proportions: exact binomial, z approximation to exact test
Two independent parametric: t-test
nonparametric: Wilcoxon rank-sum
proportions: chi squared, Fisher’s exact
Two related parametric: paired t-test
nonparametric: sign, Wilcoxon signed rank
proportions: McNemar or kappa statistic
Three or more independent parametric: ANOVA
nonparametric: Kruskal-Wallis
proportions: chi square, Fischer’s exact
Three or more related parametric: repeated-measures ANOVA
nonparametric: ANOVA by ranks
Two comparable variable
nominal: relative risk
ordinal: Spearman’s rank correlation
continuous: linear regression
One dependent and two or more independent variables
binary dependent: logistic regression
categorical dependent: ANCOVA
continuous dependent: multiple linear regression
censored observation: COX proportional hazards model
clustered or hierarchic parametric data: linear mixed models
clustered or hierarchic semi-parametric data: generalized estimating equations
Define: life-tables
Measure events at fixed intervals
