Biostatistics
Epidemiology
Genetics
Investigations
Treatment
Tests:
Univariate analysis: single variable, descriptive e.g. age percentiles
Bivariate analysis: 2 variables, relationships, e.g. prevalence by age
Multivariant analysis: >2 variables
Type I error (false positive) = rejection of (true) null hypothesis i.e. thinking there is an effect
Type II error (false neg) = acceptance of (false) null hypothesis i.e. thinking there is no effect
Nominal
- absolute numbers > plot standard distribution > compare means and SD > t test
- - paired v unpaired
- percentages > chi squared / fishers exact test
Ordinal
- rank (non parametric) e.g. wilcoxon
Study designs
Case-control
RCT
Cohort
Cross-over (own-control): each participant receives multiple interventions sequentially eliminating covariant bias
Cross-sectional (Prevelance)
Prospect v retrospective
Meta-analysis
SD 68-95-99.7 rule (% within 1,2 and 3 SDs) (i.e. 16th/84th, 2nd/98th, 1st/99th)
Phases of Clinical Trials
0: microdose for pharmacokinetics
1: Pharmacokinetics, pharmacodynamics, safety (what it does to body)
2: Efficacy at different doses (what it does to disease)
3: RCT v gold standard (how it compares)
4: Surveillance, cost effectiveness
(5: post release comparative studies by others)
Genetics
Investigations
Treatment
Tests:
Univariate analysis: single variable, descriptive e.g. age percentiles
Bivariate analysis: 2 variables, relationships, e.g. prevalence by age
Multivariant analysis: >2 variables
Type I error (false positive) = rejection of (true) null hypothesis i.e. thinking there is an effect
Type II error (false neg) = acceptance of (false) null hypothesis i.e. thinking there is no effect
Nominal
- absolute numbers > plot standard distribution > compare means and SD > t test
- - paired v unpaired
- percentages > chi squared / fishers exact test
Ordinal
- rank (non parametric) e.g. wilcoxon
Study designs
Case-control
RCT
Cohort
Cross-over (own-control): each participant receives multiple interventions sequentially eliminating covariant bias
Cross-sectional (Prevelance)
Prospect v retrospective
Meta-analysis
SD 68-95-99.7 rule (% within 1,2 and 3 SDs) (i.e. 16th/84th, 2nd/98th, 1st/99th)
Phases of Clinical Trials
0: microdose for pharmacokinetics
1: Pharmacokinetics, pharmacodynamics, safety (what it does to body)
2: Efficacy at different doses (what it does to disease)
3: RCT v gold standard (how it compares)
4: Surveillance, cost effectiveness
(5: post release comparative studies by others)