Ch. 4 Research Principles Flashcards
Data regarding presence or absence of exposure and disease are collected at the same time
Cross sectional study
Compares subjects with a condition to those without and examines frequency of a risk factor in each group
Case Control
study that is useful in studying infrequent events and calculated odds ratio
Case Control
study that is the gold standard in observational epidemiology and calculate relative risk
Cohort Study
longitudinal study comparing a defined group of people who share a common experience within a defined period before developing outcome of interest
Cohort study
subjects allocated to different interventions, placebo versus treatment group
RCT randomized control trial
study that can be used to test preventive interventions and determine absolute risk reduction
RCT
statistical test to compare means between two groups
student’s t-test or Wilcoxon rank sum test
statistical test to compare means between 3 or more groups
ANOVA or Kruskal-Wallis test
statistical test to compare two quantitative measurements from one source
paired t test or Wilcoxon signed rank test
statistical test to determine degree of association between to quantitative variables
Pearson coefficient of correlation or Spearman rank order coefficient
Hypothesis that there is no effect or no difference is called?
Null hypothesis
the probability that if the null hypothesis were true, we would observe a result as extreme than one observed in these study
P value
error that occurs when the null hypothesis is falsely rejected and we accept that a difference exists when it truly does not
type 1 error or alpha error
error that occurs when the null hypothesis is not rejected when a true difference exists
type 2 error or beta error
most common reason for a type 2 or beta error (not rejecting the null hypothesis)
underpowered study
ability of a test to detect a true difference in a data set or the probability of avoiding a type II error
Statistical power
statistical power is usually set to what percentage
80 to 90%
if the null hypothesis is true and you accept the null hypothesis, what is the probability?
1 - alpha
if the null hypothesis is true and you reject the null hypothesis, what is the probability?
alpha
if the null hypothesis is false and you accept the null hypothesis, what is the probability?
beta
if the null hypothesis is false and you reject the null hypothesis, what is the probability?
1-beta (equal to power)
what percentage of observations fall within 1 SD, 2 SD, 3 SD of the bell shaped curve?
68%, 95%, 99.7%
Difference between prevalence and incidence?
prevalence is percentage of patients with disease at a given time, and incidence is measure of individuals that develops disease during a specified period
Probability that a test will be positive for disease if the person has the disease
sensitivity = TP / (FN + TP)
Probability that a test will be negative if person does not have the disease
specificity = TN / (FP + TN)
a negative result for a highly sensitive tests can…?
rule out disease
a positive result for a highly specific tests can…?
rule in disease
the proportion of positive tests that are true positive
PPV = TP / (TP+FP)
the proportion of negative tests that are true negative
NPV = TN / (TN + FN)
PPV and NPV are affected by?
prevalence
How will PPV and NPV be affected by prevalence?
high prevalence increases PPV, low prevalence increases NPV
equation for absolute risk (AR)
ARC = # of events in control group / # of people in control group ART = # of events in treatment group / # of events in treatment group
equation for absolute risk reduction (ARR)
ARR = ARC - ART
equation for relative risk
RR = ART / ARC
equation for relative risk reduction
RRR = (ARC - ART) / ARC = 1 - RR
equation for number needed to treat
NNT = 1/ARR
equation for odds ratio OR
OR = (TP x TN) / (FP x FN)
three core principles in the Belmont Report
respect for persons, beneficence, justice
5 elements of informed consent
competence, disclosure, understanding, voluntariness, consent
