the investigation of cause and effect Flashcards
What three types of studies come under observational study?
Cross-sectional studies
Case-control studies
Cohort studies
what is a cross sectional study?
try to determine the cause and effect when measuring exposure and disease at one point in time to for a hypothesis
What is a case control study?
investigate the cause and effect by comparing exposure in
- cases (with disease or outcome)
- controls (without disease or outcome)
usually retrieve
hypothesis generating or testing
What is a cohort study?
Cohort followed over a period of time
Investigate cause and effect by comparing outcome in :
- Exposed people
- Non - exposed people
usually prospective
- Looking forward in time over a log period
Hypothesis generating or testing
What is an experimental study?
2 types
- randomised controlled clinical trials
- non-randomised controlled clinical trials
researchers introduce an intervention and study effects
usually randomised, subject grouped by chance
explain randomised controlled clinical trials
gold standard
- ability to minimise bias obtained in information obtained from participants
random assignment of participants to experimental treatment or a control group (placebo)
patients kept unaware of which treatment received in trial, Known as blind study or single blinded
When would a non-randomised controlled clinical trial be used instead of a randomised one?
only used when random allocation is problematic
More prone to bias
explain a crossover design of a randomised controlled clinical trial
All participants enter both groups, in random order, often separated by a washout period in which they receive no treatment
Cancels many forms of bias as patients are exposed to all protocols
What is cause and effect?
a cause is anything that produces an effect or result
The knowledge of cause and effect relationship can lead to successful prevention strategies
What 3 steps can causation be summarised as?
- investigation of statistical association
- investigation of temporal relationships
- Elimination of alternative explanations
Explain the investigation of statistical association in terms of cause and effect
must first find an association between the risk and the disease
statistical association can be:
1. strength
2. consistency
3. specific
4. plausible
5. Dos-response relationship
Explain investigation of temporal relationships in terms of cause and effect
causal relationship depends on the temporal relationship
to show causation the suspected causal factor must have occurred before the effect or disease developing
Explain the elimination of alternative explanations in terms of cause and effect
need to show there is no other reason for the effect other than the causal factor
What is internal validity?
How well the study answers the clinical question for
the sample.
Measure and measurements.
What is External validity?
How well the study answer the clinical question for
the population.
Internal validity and accurate sampling
What are the sources of bias within study design?
Measurement bias
Recall bias
selection bias
systematic bias
explain measurement bias ang how it can be controlled
can occur during collection of baseline or follow up data
can be differential or non differential
can compromise internal or external validity
can be controlled by using the same instruments for all measurements and double blinding the study
What is a differential in terms of study bias?
miss classification
occurs when a measurement device leads to incomparable measurements between groups
What is recall bias?
can take many forms, may occur if people who have experienced an adverse event such as a disease, they are more likely to recall previous risk factors than those who have never experienced the event
compromises internal and external validity
What is selection bias and how is it controlled?
happens when participant chosen for the study influences the outcome of the study
also happens when participants are allowed to select the group they want to join
this can be controlled using randomisation sampling methods from the target population
name 4 more types of systematic bias
Performance bias
exclusion bias
detection bias
publication bias
What is performance bias?
refers to systematic differences in the care provided or in exposure to factors other than intervention of interest
What is exclusion bias?
Systematic differences in withdrawals from the trial
common reasons for exclusion of patients : ineligibility, protocol violation, early outcome, loss to follow up after study
What is detection bias?
Systematic differences in outcome assessment
cause an over or under estimate of the cause of the effect
Neyman bias (preferential detection of mild, slowly progressive cases of disease)
What is publication bias?
Selective publication of positive results
reluctance to publish papers with negative results
distortion of literature
What is random error and how does it influence the study design?
Random error refers to distortion of data in any direction.
Produces finding that are too high and too low in
approximately equal amounts.
Decreases the probability of finding a real association by
reducing the statistical power of a study.
less damaging than bias
What is cofounding effect and how does it effect the study design?
Confounding is the confusion between two factors that lead to an
outcome.
Can lead to a false identification or obscuration of a cause-and-effect
relationship.
Compromises internal and external validity.
What is Effect modification and how does it effect study design?
in effect modification the directoion or strength of an association between two variables differ according to the value of a third variable
known as interaction
occurs when a cause and effect relationship is modified by using another variable
compromises internal and external validity
can be controlled using statistical modelling
What is a qualitative study
Seeks a deeper truth, form of social enquiry where the
researcher seeks to understand human behaviour.
Based on:
Action
Observation, Interview
Classification
Inductive
Theoretical
Validity
What are the ttributes of a qualitative study?
People focused
use of textual data
exploration in natural settings
inductive reasoning
systematic and flexible
small sample size
rich descriptions
identifies patterns
builds theories
What is a questionnaire based study?
Objective means of collecting information about people’s
knowledge.
can be qualitative or quantitative depending on questions asked
Based on:
Questions
Multiple options
Simple options (true/false, yes/no)
What are the requirements for a questionnaire based study?
Appropriate questions
valid and reliable
appropriate for target pop
clear instruction
questionnaire adequately piloted
appropriate sampling
maximise response rate
data analysis
What is secondary literature analysis?
Includes meta-analysis and decision analysis.
Meta-analysis:
Exposure and outcome examined from the literature.
Useful when previous studies have sample size.
Decision Analysis:
Improve clinical decisions.
Decision trees.
What are the steps for a meta analysis?
- The research question
- inclusion and exclusion criteria
- objective method for article selection
- blind article review and decide on eligibility
- effect across studies is calculated
- homogeneity across study is calculated
- weighting method applied
- average combined methods is calculated
What are the steps to decision analysis?
- improve clinical decisions
- decision tree
2.1 Decision nodes
2.2 chance node
2.3 probability values
2.4 utility values
2.3.1 adopting a given management strat
2.3.2 any possible outcome