Prevention, Ethics and Biostatistics Flashcards
What are the USPSTF recommendations for cholesterol screening in patients WITHOUT CAD risk factors?
- Men begin at 35
- women at 45yo
- screen Q5years
USPSTF recommendations for screening mammograms?
- women >50yo
- q2 years
Legal obligation to conform to a reasonable standard of care?
Duty of care
Failure to conform to the standard of care?
Breach of duty
The breach of duty is determined to be the cause of injury/harm?
Causation
what are the 2 ways that the standard if care can be established in a malpractice case?
- expert testimony
- Res ipsa loquitur (“it speaks for itself”)
Supervisors are responsible for the actions of their subordinates?
Vicarious liability
Incidence
new cases/total # of people at risk
Prevalence
of people with disease/total population
Sensitivity
A/(A+C)
Specificity
D/(D+B)
PPV
A/(A+B)
true positives that tested positive
NPV
D/(C+D)
How many true negatives our of all of the negative tested
Do high-prevalence diseases have a high PPV or a high NPV?
High PPV
Relative Risk
RR= (A/(A+B))/(C/(C+D))
- RR>1 –> exposure increases risk of disease
- RR
What study is used to develop relative risk?
Cohort study
Odds Ratio
(AxD)/(CxB)
used to estimate RR in low0-prevalence
What study is used to develop odds ratio?
Case-control
Attributable Risk
(A/(A+B)) - (C/(C+D))
Absolute Risk Reduction
ARR= (C/(C+D)) -(A/(A+B))
Number needed to treat
1/ARR
What is a statistically significant p-value?
What does a lower p-value indicate?
the more statistically significant the finding it (the less likely it was just due to chance)
Confidence interval
CI=mean +/- (Zx SEM)
- For 90%, Z=1.645
- for 95%, Z=1.96
- for 99% CI, Z=2.58
If the CI for an odds ratio or relative risk crosses one, what does this indicate?
No association between risk and disease
If the CI for a study between 2 treatment groups crosses 0, what does this indicate?
No actual difference between the 2 treatments
what percentage of the study population falls within 1 SD of the mean? 2 SD? 3 SD?
1=68%
2=95%
3=99.7%
What can you identify in case-control studies? What can you calculate?
- potential risk factors (these are ALWAYS retrospective)
- Odds ratio
What is a cohort study? What can you calculate?
- Follow a cohort that is exposed to a given risk factor to see how many develop a disease (prospective)
- Relative risk
Type of bias? When studying the effects that infection control education has on physicians, both the experimental and the control groups improve their hand hygiene
Hawthorne effect (chase behavior when you know you’re being observed)
Bias? Sending out a survey to people diagnosed with a fatal illness 5 years after a diagnosis will preferentially sample those will low-grade disease or fewer comorbidities
Late look bias
Bias? Sicker patents are preferentially encouraged to enroll in a study
Selection bias
