ambirnba's version from 2015-06-04 20:51

Section 1

Question Answer
Parametric tests can be applied...Can only be applied to numerical data w/ normal distribution (can calculate mean & SD)
Non-parametric testsNon-normal distribution May not be numerical data, i.e. categorical/nominal data, e.g. ordinal data (exist along a linear scale)
non-parametric data, comparing 2 groupsmann-whitney U test
non-parametric data, comparing 2+ groupskruskal-wallis test
non-parametric data, comparing rank dataWilcoxon rak-sum test
Single blindedpt doesn’t know if they’re receiving treatment or placebo (to account for placebo effect)
Double blindedresearcher & pt don’t know if pt is receiving treatment or placebo (to account for researcher’s biases)
Intention to treat analysisincludes all pts who entered the study, including those who dropped out or weren’t compliant
per protocol (PP) analysisincludes only those pts who completed the study as originally intended
which is better: intention to treat or per protocol analysis? why?PP analysis tends to inflate the benefits & does not account for biases that led pts to be noncompliant, so ITT analysis is better

Section 2

Question Answer
regression is a mathematical model to describe relationship b/w explanatory (exposure/predictor) variable & dependent (outcome) variable
linear regression is a form of regression w/ continuous outcome variable
correlation coefficient symbolr
correlation coefficient ranges between -1 and 1
when does correlation coefficient |r|=1 ?If exposure variable accounts for all the variation in the outcome variable
when is r negative?if outcome decreases w/ increasing exposure
what is important to keep in mind about correlation?Correlation ≠ causation
wilcoxon rank sum test is a ______ test to calculate ___ for ____ samplesNon-parametric test to calculate p for 2 samples (i.e. non-normally distributed data)
how do you calculate for a wilcoxon rank sum test?Assigns ranks to data from 2 samples, put data from both samples in order, give them a rank, sum up the ranks of each sample, & then do actual math to calculate p
cofounder is a factor associated w/ both the exposure & the outcome
is cofounding an issue in observational studies? in randomized control trials?Confounding is especially an issue for observational studies; randomized controlled trials (RCTs) control for possible confounders (but only if they’re correctly identified!)
you can turn a confounder into a covariate
how do deal with confounders? (3 steps)Must identify confounders, measure their effects on the outcome, & remove that effect in statistical analysis

Section 3

Question Answer
ARR = RR in placebo group - RR in intervention group
NNH (# needed to harm) = 1/ARI (abs risk increase)
ARI (absolute risk increase) = RI in placebo group - RI in intervention group
hazard: the rate over time that an event happens (e.g. developing diabetes)
multi-linear regression modelfor continuous outcome, but for 2+ explanatory variables
logistic regressionfor categorical outcome variable (e.g. yes/no for DM dx; dead/alive for Kaplan-Meier curve)
Proportional hazards models tell you the impact of a risk factor on the time until an event occurs
adjusts for people dropping out(Cox) proportional hazard model
(Cox) proportional hazards regression assesses a rateit models the # of new cases (or deaths) per at-risk population per unit time
(Cox) proportional hazards regression allows you to analyze the effect of ____ on _____Allows you to analyze the effect of several risk factors on survival
________ assesses the cumulative incidence (a measure of disease frequency during a period of time)logistical regression
Fixed effect in regression analysis not relevant outside of this study; something being held constant; those factors aren’t generalizable outside the study
random (as opposed to fixed effects)may be generalized outside the study

Section 4

Question Answer
4 principal activities of preventionimmunization, screening, behavioral counseling, & chemoprevention
Primary prevention: keeping dz from occurring by removing its causes
Secondary prevention: detecting early, asymptomatic dz & treating it to stop it from progressing
Tertiary prevention: preventing deterioration/reducing complications after a dz is already present; basically the same as treatment
Sensitivity: true positives/total # sick pts; A/(A+C)
Specificity: true negatives/total # healthy pts; D/(D+B)
Positive Predictive Value (PPV): true positives/test positives; A/(A+B)
Negative Predictive Value (NPV): true negatives/test negatives; D/(D+C)
As prevalence increases, PPV _____ & NPV ___PPV increases & NPV decreases
a method of using mathematical procedures to make intelligent guesses about a population based on a sample & to generate new knowledgestatistical inference
Statistical significance is defined asthe likelihood that the outcome wasn’t chance (i.e. p<0.05)
Clinically, “significance” is importance/relevance
_______is used for ranked data; correlation coefficient is used for ______spearman rank correlation; continuous outcome

Section 5

Question Answer
purpose of Federal Trade Commission (FTC) meant to protect consumers & prevent anti-competitive or deceptive business practices (e.g. scam cancer tx, dietary supplements, illegal online pharmacies, erectile dysfunction tx, direct-to-consumer genetic tests)
another name for clinical hypothesis testingco-selection
Clinical hypothesis testing = simultaneously gathering clinical data & searching for illness scripts to generate & test hypotheses
Clinical reasoningprocess by which physicians analyze & organize clinical data to arrive at ddx & write assessment
Precontemplationunaware of a problem or no intention to change in the foreseeable future
Contemplationambivalent about changing with no commitment to change in the near future
Determination/preparationintends to take action within the next month or has made some recent changes in behavior
Actionactively attempting to change; generally 1st 6 mo
Maintenancehas changed and needs to prevent slips and relapse; generally after 6 mo
adolescents are in formal operational stagedevelop capacity for hypothetical thinking
Propositional thoughtability to reason w/ symbols w/o using concrete representations
Higher order thinking, incl. metacognitionawareness of mental strategies employed to solve a problem

Recent badges