Study designs Flashcards
Design can be described on the basis of:
- Identification (or grouping) of study subjects (exposed vs non-exposed or cases vs controls)
- Time course of intervention i.e. prospective, retrospective, cross-sectional
Descriptive studies
Describe the characteristics present in members of the group
Case reports / series
Cross sectional studies (prevalence studies)
Cross sectional studies
Single point in time
Census - whole population
Prospective longitudinal studies - observations repeated in the same population over a prolonged period of time, costly and time consuming
Analytical studies
Comparative analysis of two groups
- Case-control studies
- Cohort study
- Ecological study
Case control study
Group of cases and group of controls
Hypothesis of putative causal relationship supported if larger amongst cases than controls
Cohort study
Observational study of a group of people with specified characteristics or exposure, followed over a period of time to detect events / outcome
Ecological study
Data describes what happens in a group rather than an individual
Drawing subject level conclusions based on group level observations in an ecological study is termed ecological fallacy
Experimental studies
Generally prospective cohort studies where the exposure is experimentally assigned - deliberate manipulation
- Controlled trials
- Uncontrolled trials
Best study designs for clinical enquiries
- Treatment effectiveness - Pragmatic RCT / SR
- Treatment efficacy - Experimental RCT / SR
- Causation - Cohort / case control
- Prognosis - Cohort
- Diagnostic assessment - Cross sectional comparison to gold standard
- Health economics - cost-effectiveness study
- Meaning or health experience - Qualitative studies
Case control design
Retrospective
Both exposure and disease have occurred before the onset of the study
Often first mode of studying a suspected association of causality
Case control advantages
Easy to carry out
Less time-consuming
Less expensive
Suitable for investigating rare diseases
Subjects not exposed to any new risks
Several different etiological factors can be studies
No attrition problems
Case control disadvantages
Highly prone to selection + recall bias
Control group selection may be difficult
Incidence cannot be measured - so odds ratio only, not relative risk can e measured
Cannot prove causality
Temporality is difficult to determine
Cohort studies
Birth cohort e.g. all those born on 13/09/1981
Exposure cohort - all those exposed to a risk factor
Inception cohort - group of patients who are assembled at a single (or narrow) point of time based on a common factor
Cohort study disadvantages
Difficult to carry out as more time / resource consuming
Not suitable for investigating rare diseases
Only one etiological factor can be studied at a time
Attrition / drop out is a major issue
