Statistics Flashcards
A systemic error where a study group is not representative of a population.
Selection/Sampling Bias
People who volunteer are not the same as the general population.
Healthy User Effect
When the sample is taken not from the general population, but from a subpopulation.
Berkson Bias
People who choose to be in a study are different from people who do not; research only makes conclusions about those who choose to take part in the study.
Non-Respondent Bias
Sicker people are more likely to be in studies; biased toward sicker cases.
Ascertainment Bias
Those with more severe sickness are least likely to be included in the study; harder to access or more likely to be decreased; sample biased to less sick cases.
Late-Look Bias
People who choose to be in a study are different from people who do not > study participants who dropped out/died were not accounted for analysis > drop-outs/deaths may be different from those who stay/survive.
Attrition Bias
Statistical analysis that keeps drop-outs in the same groups; drop-outs counted as non-responders.
Intent-To-Treat Analysis
What are examples of selection/sampling bias?
- Healthy User Effect
- Berkson Bias
- Non-respondent Bias
- Ascertainment Bias
- Late-Look Bias
- Attrition Bias
Information is gathered in a way that distorts the association between exposure and outcome; how the data are collected will affect the data that are obtained. For example, leading questions and Hawthorne Effect.
Measurement Bias
A subject’s behavior is altered because they are aware they are being observed.
Hawthorne Effect
What are some examples of measurement bias?
- Recall Bias
- Observer Bias
- Placebo Effect
People do not accurately remember what happened in the past; they make things up or base things on the present circumstances.
Recall Bias
Person making the assessment/measurement assumes an outcome based on their previous experience.
Observer Bias
Patient in the comparison/control/placebo group believes that they are in the treatment/intervention/drug group, and reports accordingly.
Placebo Effect
Researcher unintentionally acts in a way that influences the outcome.
Expectancy Bias (Pygmalion Effect)
Tendency to only publish results with “positive” or significant outcomes.
Publication Bias (File Drawer Effect)
The relationship between exposure and outcome is not a perfectly linear one; often there are other things that influence the association between the exposure and the outcome.
Or when the association between an exposure and an outcome is altered by the presence of an extraneous variable.
Confounding Bias
Overestimation of survival duration due to earlier detection by screening (ex. over-diagnosis/over-detection).
Lead-Time Bias
What are some features that may be included in a Randomized Control Trial?
- Placebo
- Standard of (Usual) Care
- Wait-List Control
When a treatment of benefit already exists, and we are looking for a better treatment - “Does this work better than what already exists?”
Standard of (Usual) Care
Is a group of participants included in an outcome study that is assigned to a waiting list and receives intervention after the active treatment group. This control group serves as an untreated comparison group during the study, but eventually goes on to receive treatment later.
Wait-List Control
What are the phases of a clinical trial?
SWIM
1 - Safety
2 - Work (Does it work?)
3 - Improve
4 - Marketing
Prevent the onset of illness or injury before the disease process begins. Prevents incidence.
Primary (Universal) Prevention
Lead to early diagnosis and prompt treatment of a disease, illness or injury; to prevent more severe problems developing.
Secondary Prevention
Aimed at rehabilitation following significant illness. Reducing negative impacts, increasing quality of life.
Tertiary Prevention
Collects all possible studies related to a given topic and design, and reviews and analyzes their results.
Systematic Review
Go a step further and statistically analyze the results from all the studies.
Meta-Analysis
If the diamond on a forest plot is sitting on or touches the line, it is considered?
Not Significant
If the diamond on a forest plot is to the left of the line, it is considered?
Lower Risk
If the diamond on a forest plot is sitting to the right of the line, it is considered?
Higher Risk
A measure of how different the studies are, and if that difference is outside that of expected chance.
Heterogeneity
If a p value is > 0.05, what does that mean for heterogeneity?
Low Heterogeneity; it is the opposite if less than
On a funnel plot, if there are more symmetric dots clustered around the line, that means there is?
Less Publication Bias
What is the eligibility criteria for a systematic review or meta-analysis?
PICOS
P - Population I - Interventions C - Comparator O - Outcomes S - Study Designs