Social Flashcards
Difference between modifier and confounder
Modifier is only associated to disease, not risk factor
Confounder is linked to both disease and exposure
Case-control studies have what measurement?
Odds ratio (exposure)
Prevalence odds ratio is from cross-sectional study comparing prevalence between two populations
Relative risk is incidence and is measured in cohort studies
Number needed to treat
1/ARR
Absolute risk reduction = % event in placebo - % event with treatment
When does odds ratio approximate relative risk?
When incidence/outcome is low (RR = (a/a+b)/(c/c+d), a«b and c«d)
Also when cases and controls are representative of the population (with regard to exposure of interest)
Attributable risk percent
Excess risk in exposed population that can be attributed to risk factor
ARP = (Risk in exposed - risk in unexposed)/Risk in exposed
or
ARP = (RR - 1)/RR
What must be performed after sentinel events?
Root cause analysis
What kind of event is wrong-patient error?
Sentinel event
WPE can be improved through interprofessional rounds
Admission rate bias
When distortion in risk ratio exists due to hospitals’ different admission rates for certain cases
Response bias
When participant purposely gives slanted response for whatever reason (e.g. med student not wanting to admit to smoking)
Type I vs II error
Type I (alpha): false positive; significance level
Type II (beta): false negative; power = 1-beta
When is informed consent not needed for research?
Minimal risk and no interaction with patients (i.e. secondary analysis of existing data)
However, informed consent is generally needed if it is protected health info (which helps identify the patient)
Absolute vs relative difference in a clinical trial with odds ratio of response
Absolute = treatment response rate - control response rate
Relative = (treatment response rate - control response rate)/control response rate
What is a measure commonly used in cross-sectional studies?
Prevalence ratio
When is hazard ratio used?
Measure of association for time-to-event (survival) outcomes
Commonly used in cohort studies and experimental designs
What is a receiver operating characteristic (ROC) curve?
Plot of true-positive rate (sensitivity on y-axis) against false-positive rate (1-specificity on x-axis)
Compare same cut point on two different curves to assess changes in sensitivity and specificity
Hawthorne effect
When subjects change their behavior because they are aware they are being studied
Per-protocol vs intention-to-treat analysis - what’s the difference?
Per-protocol: excludes dropouts; estimates “true effect” assuming a perfect scenario but overestimates effect in realistic, practical, clinical setting
Intention-to-treat: includes all participants as initially allocated; conservative estimate but reflects “expected effect”
Psychological safety
Comfort with taking actions for the purpose of patient safety
Change management
Engaging personnel to make initiatives to improve safety
Parallel study
One treatment to one group, different to another group; no other variables measured
Factorial design study
2+ experimental interventions, within each are 2+ variables studied independently (same ones across the interventions)
What are “forcing functions” or “hard stops”?
Physical designs that prohibit incorrect action (e.g. gas containers with different connectors to ensure connection to correct tubing)
Calculating vaccine efficacy
VE = (risk in unvaccinated - risk in vaccinated)/risk in unvaccinated
Who should perform oversight of data from clinical trial?
Independent data monitoring committee, not IRB
