Apply Lecture 2/5 Flashcards
What is critical appraisal?
Systemic review to determine:
- quality
- trustworthiness
- relevance
What are the 5 parts of a research article?
Abstract Background/Rationale/Intro/Hypothesis Methods Results Discussion References, acknowledgements, conflicts of interest
What does an abstract include?
context purpose procedures main findings conclusions new aspects
What is included in the Methods?
State IRB approval state informed consent answer info about PICO components describe study design sample size estimate (determined ahead of study) describe statistical analysis used randomization method allocation concealed masking or blinding
What do the results include?
How precise is the estimate of the treatment effect?
Validity
Generalizability of results
how do we determine how precise is the estimate of the treatment effect?
confidence intervals
do CI’s include values that indicate no difference in the treatment effect
What is validity?/why do we care about it?
likely influence of chance
bias or confounding variable
what is included in the results
tables and charts
actual data values
means and medians
percentages
what is usually in Table 1 of a study?
key characteristics of participants to compare groups
what are confidence interval and p value used for?
applying to measures of associations observed
how do we obtain P value?
through significance or hypothesis testing
what is in the discussion?
context of the study
implications of results on health or condition and intervention involved
study limitations
honest and frank discussion
what are the 8 Bradford Hill’s Criteria of Causality?
- Strength of Association:
- Risk difference of absolute risk reduction
- Relative risk
- Odds ratio - Consistency of an Observed Association
- Repeated observation in different populations in different circumstances - Specificity of Association (cause leads to a single effect)
- Temporal Sequence (can be long time)
- Dose-response relationship (ex. “pack years”)
- Biological Plausibility (ex. HIV increased risk of AIDS)
- Experimental evidence (remove factor to see if decreased)
- Analogy (ex. If one virus can cause disease, 2nd may cause similar disease)
what is internal validity?
was the research done “right”?
what is external validity?
does the same thing happen in other settings…. other labs…. real life?
what is CONSORT stand for?
consolidated standards for reporting trials
what does CONSORT do?
shows how participats move along in a study, and if there are drop outs
what are some measures of variability?
standard deviation
variance
standard error of the mean
what is effect size?
quantifies the difference between groups
how do you calculate effect size?
mean difference/pooled standard deviation
is .20 a large, medium or small effect size?
small
is .50 a small medium or large effect size?
medium
is .8 a small medium or large effect size?
large
if you are confused about effect size…. watch this video
https://www.youtube.com/watch?v=6uYNVCy-8NA
what is a correlation?
measures the strength and direction of a linear relationship between two variables
is correlation an indicator of cause and effect?
no
what does a correlation coefficient of 0 mean?
no relationship
what does a correlation coefficient of +1 mean
positive linear relationship. as you increase one, the other increase
what does a correlation coefficient of -1 mean?
negative linear relationship. as you increase one, the other decreases.
what is the strength of the correlation?
r value
what is a good r value?
as close to 1 as possible.
what is considered a good correlation (r value?)
.7 and closer to 1
what is regression
given a moderately strong correlation, Y can be predicted given X.
A regression equation is found using the slope and y intercept.
AKA… y=mx+b
what does r^2 mean or represent?
r^2 represents the amount of variation in the data that is explained by the regression
what is the null hypothesis?
the hypothesis that the groups are equal and there is no difference
what is ALPHA
agreed upon level for probability of chance in explaining the results of the study. usually 5% (.05)
p value
an expression of the probability that the difference has been identified in the study is due to chance. p=.02
which value do we set? p value or alpha?
alpha
what does it mean if the p is lower then the alpha?
then we reject the null hypothesis!!! bc there is a difference
