2 Epidemiology Study Designs Flashcards
A hypothesis usually takes the form
Exposure X is associated with an increase/decrease risk of outcome Y.
2 types of error
systematic error (referring to bias)
random error (referring to imprecision).
what is bias
systematic error or derivation from the truth of results
due to incorrect measurements or non-representative sample.
Selection bias
allocation of subjects not representative of whole population. Controlled by randomisation with concealment.
Performance bias
performing intervention different for different patients, results are not controlled. Controlled by double blinding, use of placebo.
Detection bias
Evaluation of outcomes is different between groups. Controlled by blinded outcome assessment that doesn’t pinpoint only one group.
Measurement/observation/information bias
systematic error of poor measurement of outcomes. Controlled by review of measurement tools.
Recall/responder bias
inaccurate responses/recalled. Controlled by a good study design.
Attrition bias
different flow up between groups, + drop outs. Controlled by analysis of subjects as randomised.
What is confounding
non-causal association observed due to third variable.
describes an association that is true but potentially misleading.
How can confounding be controlled
design phase:
randomising
restricting
matching parts of the population.
analysis stage:
stratification of results
adjusting by statistical modelling.
Explain reliability
Reliability tools measure accurately and consistently. = PRECISION
Test-retest reliability: results consistent by measuring twice.
Inter-rater reliability: assesses the degree of agreement between judges.
Explain validity
refers to the legitimacy of the research = ACCURACY
Internal validity: correct results for group studied
External validity: generalisable results
Construct validity: test measure what it claims to be measuring
Case study
Reports observations of single individual, case series is an aggregate.
Ecological study
looks at groups of people rather than individuals.
Compares disease frequencies between different populations at 1 time or 1 population at different times.
Cross-sectional study
Also called a prevalence study
measures prevalence and presence/absence of exposure.
only 1 point in time no past/future
Case-control study
1 group has disease other does not, look back and determine how many from each group had the risk factor.
Retrospective
Uses O.R
Cohort study
Comparison of groups of people who share common characteristics within a define time period
Prospective = see who develops
Retrospective = exposure and outcome already occurred
Uses R.R
RCT
Patients randomly allocated to treatment and control group, compare outcomes due to intervention
Randomisation controls for confounding by distributing confounders evenly
Cross over trial
normal RCT, washout period, then swap intervention and controls groups.
n of 1 trial
cross over trial with only one patient
Cluster of RCT
Unit of measure is not an individual but a group. Often used in communities where there is a risk of contamination
What is a systematic review
answers a defined research question by summarising evidence of eligibility criteria.
What is a meta analysis
statistical methods to summarise results of studies
Narrative review
broad
author can use literature to support view point
limited critical appraisal
characteristics of systematic review
minimise bias
answers question
summary of data
regularly updated.
Steps of a meta analysis
- Formulate structured clinical question
- Conduct extensive search documenting search strategy
- Select studies for inclusion, justify exclusion
- Assessment of methodological quality of including studies
- Abstraction of relevant data
- Data synthesis (meta-analysis)
Describe blinding
people are unaware of which trial group the participants are in.
minimise performance and detection bias such as the care that is provided to the comparison groups.
