Week 8 Experimental Study Design Flashcards
Difference between observational and experimental studies
Observational = no manipulation of stud variables. Purely descriptive and analytical. E.g. cohort (retro and pro spective), case-control studies
Experimental = aka trials
Study factors are manipulated by investigator. E.g. participants randomly allocated to a treatment and non treatment group
What are the 3 types of trials?
Field Trial = used to evaluate interventions intended to prevent disease in healthy people. E.g. field trial of a vaccine
Community Trials = used to evaluate community-wide interventions (e.g. a community trial on the effects of fluoridation of public water supply)
Clinical Trials = used to evaluate efficacy of treatments on ill people (e.g. drug trials, surgery trials, etc)
Benefits of Expiermental Study Designs?
Provide better evidence of effect and outcome than can be obtained with observational studies
Afford greater internal validity due to greater control over exposure, etc, and thus greater confidence and assurance of truth of results
Fastest and safest way to find treatments that work in people, to improve health
Selection Criteria for Experimental Studies
There is an inclusion criteria = the condition being studied may be restricted to specific manifestations of the condition
Exlusion criteria = may be based on presence of other co-morbidities, patients within whom the treatment is directly contraindicated, patients not expected to live long enough to experience the outcome of interest, if refuse to participate
What is internal and external validity?
Internal validity = the extend to which the results hold true in terms of the people being studied. Based on reduction of bias within study design
External validity = how accurately the findings of the study can be applied to the greater population - “generalizability”
How to increase external validity - generalizability
Internal validity still imprtant
Describe the study sample carefully
Avoid using unusual patients
Simplify the trial: reformulate your study questions and exclusion criteria to include more participants
Control Group
You cannot assess the efficacy of the treatment in the treatment group without an appropriately matched control gorpu
Improvements unrelated tot he treatment
Controls are important, because there are other reasons why improvements may occur, that are unrelated to treatment.
E.g. the natural evoltion of the disease may change over time
Placebo effect
Observing the control group in the same way as the treatment group may migrate the problem
Benefits of randomisation?
Are greateset with large sample sizes
Simple randomisation
Completely random.
Stratified random
E.g. sort of random, but balanced so that the severity of disease is equally distributed between groups, that there is even distribution of age, and co-morbidities
Outcome assessment
Outcome assessment MUST be standardized and uniformly applied
This is to minimalise bias
Can be ensured via:
Explicitly stated criteria for outcome assessment
Ensure comparibility of study groups at study entry
Blinding of participants and observers
Blinding
Important in experimentation. Blinding both participants and observers.
What differences can arise after randomisation, during the study?
Groups that were comparable at the onset of the study may become less so as time progresses
Hospitalised patients may be easier to follow
Co-interventions - other treatments may be started in participants, posing potential for bias if they affect the oucome
Unplanned crossover
can occur when participants are originally randomly allocated to a surgery or medical group. Surgery-allocated participants may refuse surgery and thus have to join the medical group. Participants assigned to the medical group may end up requiring surgery.
Large numbers of crossover can pose a real problem. It is rarely evenly distributed between the groups
