Ethics & Statistics Flashcards
Elizabeth Kubler-Ross psychological stages at the end of life
Denial, anger, bargaining, depression, & acceptance
Sensitivity
the probability that a person w/ a disease will have a positive result on a given test
***high sensitivity is useful in a screening test, as the goal is to identify everyone w/ a given disease
Specificity
the probability that a person w/o a disease will have a negative result on a test
***high specificity is desirable for a confirmatory test
positive predictive value (PPV)
the probability that a person with a positive test result has the disease (true positives/all positives); if a disease has a greater prevalence, then the PPV is higher
negative predictive value (NPV)
the probability that a person w/ a negative test result is disease free (true negatives/all negatives); a test has a higher NPV when a disease has a lower prevalence
incidence
the number of NEW cases of a given disease per year
prevalence
the total number of EXISTING cases of a given disease in the entire population
absolute risk
the probability of an event in a given time period
relative risk
used to evaluate results of cohort (prospective) studies; compares the incidence of a disease in a group exposed to a particular risk factor (RF) w/ the incidence in those not exposed to the RF; an RR 1 means that the event is more likely in that group
odds ratio
used in case-control (retrospective) studies; compares rate of exposure among those w/ & w/o disease; considered less accurate than RR
Absolute risk reduction (ARR) or attributable risk
measures risk accounted for by exposure to a given factor, taking into account the background of the disease; useful in RCTs; numerically, ARR=the absolute risk (rate of adverse events) in placebo group minus absolute risk in treated pts
relative risk reduction (RRR)
used in RCTs; the ratio between 2 risks; numerically, RRR=[the event rate in control pts minus event rate in experimental pts] divided by the event rate in control pts
***can be deceptive & is clinically far less important than ARR
number needed to treat (NNT)
the number of pts who would need to be treated to prevent 1 event; NNT=1/ARR
statistical significance/p-value
expresses likelihood that an observed outcome was due to random chance; a p-value <0.05 is gen. accepted as indicating that an outcome is statistically significant
confidence interval (CI)
like the p-value, expresses certainty that observation is real or is a product of random chance; used with ORs & RR, the 95% CI says that the observed risk or odds have a 95% chance of being w/in the interval
***if a 95% CI includes 1.0, the results are NOT significant
cohort study
a population is observed over time, grouped on the basis of exposure to a particular factor, & watched for a specific outcome
- not good studies for rare conditions
- can be prospective or retrospective
- use RR to interpret results
case-control study
a retrospective study involving a group of people w/ a given disease & an otherwise similar group of people w/o disease who are compared for exposure to RFs; good for rare diseases; use OR to interpret results
randomized controlled trial (RCT)
a prospective study that randomly assigns participants to a treatment group or to a placebo group; the placebo group & treatment group are then compared to determine if the treatment made a difference
*the double-blind RCT is the gold standard of experimental design
Standardized mortality ratio (SMR)
Represents an adjusted measure of overall mortality and is calculated by dividing the observed number if deaths in the population of interest by he expected number derived from the reference population; mortality is typically adjusted for age and less commonly for gender, race, etc.
Bell curves and standard deviations
In a normal (bell-shaped) distribution curve, 68% of observations lie w/in 1 SD of the mean, 95% w/in 2 SDs, and 99.7% w/in 3 SDs
