Research/Stats/Ethics Flashcards
Sensitivity
- Good screeNing test but can have false positives
- Negative results are really reliable, positive results are not as reliable
- seNsitivity have reliable Negatives (CAN RULE OUT DISEASE!)
- Probability that the test will produce a true positive when used on a population with the disease
= TP / (TP + FN) —– true positives divided by everyone with the disease
Specificity
- Positive results are really reliable, negative results are not as reliable
- Good Confirmatory test (to determine treatment) but can have false negatives
- sPecificity have reliable Positives
- Probability that the test will produce a true negative when used on a population without the disease
= TN / (TN + FP) —- true negatives divided by everyone without the disease
Positive Predictive Value
- Probability that someone who tests positive actually has the disease and isn’t a false positive
- Most useful when the prevalence is high (PPV increases as prevalence increases)
= TP / (TP + FP)
Negative Predictive Value
- Probability that someone who tests negative actually does not have the disease and isn’t a false negative
- Most useful when prevalence is low
= TN / (TN + FN)
P value definition
- Chance that the null hypothesis was rejected in error
- Odds that the null was correct and the results were due to chance
Type I error
- Rejected the null hypothesis in error
- Probability of a type I error is the same as the p value
Type II error
- Accepted the null hypothesis in error
Validity
- Whether a test actually measures what it’s intended to measure
- Internal validity reflects accuracy
- External validity reflects generalizability
Reliability
- Consistence or repeatability of scores
Intention to treat
- All patients remain in the original groups to which they were initially randomly assigned in the study
Number needed to treat
- Number needed to treat to prevent one adverse effect
- NNT = 1 / absolute risk reduction
- Absolute risk reduction = rate in untreated group - rate in treated group
Incidence
Newly diagnosed cases in a given period of time
Prevalence
Total number of cases of disease existing in a population at a given time
Standard error
Describes how accurate the sample mean value used in the analysis is compared to the true population mean value
Confidence interval
Measure of the reliability of your result
Pre-test probability
Best estimate of the probability that a condition is present before you start diagnostic testing
Likelihood ratio
Likelihood that a person who has the condition will have a positive test result
- LR > 1 increases the probability that the target condition is present
- LR < 1 makes condition less likely
Absolute risk
Dividing number of patients who develop disease by total patients exposed
= A / (A+B)
Relative risk
Probability of an outcome in the exposed group to the probability of the outcome in an unexposed group
= (A / (A+B)) / (C / (C+D))
Odds ratio
Odds of an outcome in one group compared to the odds of that outcome in another group
= (A/B) / (C/D)
Order of best study types
- Systematic reviews, meta analysis
- Randomized, controlled trials
- Cohort studies
- Case control studies
- Cross sectional studies
- Case studies/case reports
- Expert opnions/editorials
Randomized control trial benefits
- Blind study, randomly assigned
- Baseline groups are equal, no confounding variables
- Limits expectation bias of patient and reduces bias of interpreting results
Cohort studies
- Compare a group with an exposure to a control group without an exposure
- Limitations: need large sample size and long time to do they study
Case control studies
- Compare patients with the disease and patients without disease and then check for risk factors
- Risk factor has to come before the outcome
- Useful only with a small sample size, a long latency period, or a rare outcome
Cross sectional studies
- Observational, good for looking at associations between two measured factors at one point in time
- Limitations are risk of confounding variables and results can’t be used to infer causality
Systematic reviews
Require: specific clinical question, thorough search for studies, clear explanation for which studies were considered and which were not, descriptive results section summarizing the findings
Definition of brain death
Patients in a coma lack wakefulness and lack awareness of self environment. Patients in a vegetative state lack awareness but do have wakefulness.
Sibling donor rules
Both donor and recipient are to benefit (mostly psychological to donor so if sibling has severe cognitive disabilities they are not going to benefit), risk to donor is minimal, absence of coercion
Maternal-fetal conflicts
- Autonomy and bodily integrity of mother is critical and must be honored
- Can’t legally force a mom into a C-section for any reason
Assent
- Willingness to accept treatment
- It is preferred for kids to agree to the treatment but since they’re minors, it’s not required like consent
Newborn screening guidelines to include disease
- Disease should have a significant deleterious effect on the child, have effective treatment available if diagnosed early, have a screen that is acceptable to the population
Newborn screening positives/negatives
High rate of false positives so need more definitive testing (but can also have some false negatives)
Ethics for correcting ambiguous genitalia
- Decision should be multidisciplinary approach
- Possibility of gender change in adulthood should be factored in and surgery should be reversible if possible, fertility should be preserved if able
Ethics for HIV positive patients
- Disclose slowly over time to child
- Participation in close contact sports where blood and bodily fluid transmission is likely is not recommended
- It is NOT mandatory for parents to disclose a child’s positive HIV status to the school
Children exposed to domestice violence are at risk for
- Physical injury
- Aggression with their peers
- Separation anxiety as toddlers
- School failure as adolescent as well as risky sexual behavior
Foster care health rules
- Health screen should be done within 72 hours of placement into foster care
- Comprehensive health assessment within 30 days
Most common type of medical error
Medication error
Disclosing medical errors to family
- Be totally transparent, why it happened, how it will impact them, steps being taken to prevent it from happening again
- Shown to reduce litigation and decrease settlement amounts
