E and PH Flashcards
Ecological fallacy
Where the study looks at population (often geographical) averages, and not the individual level. So for example you may have a study showing that in county with higher, there is higher pulmonary fibrosis. But we don’t know that the people
ROC curve overview
Which study can detect on 1 outcome and which can detect only 1 exposure?
Case control and cohort resp
Good study for rare disease or disease with long latent period
Case control
Out of exposure and outcome, which is the dependent and independent variable
Exposure is the independent variable
What stats technique would we use to assess quantative exposure and a dichotomous outcome
Logistics regression
Larger sample sizes do what to T1 and T2 errors
Likely increase and decrease respectively
Intention to Tx vs per protocol vs as treated
ITT - it’s the shit, analyse based on inital allocation
Per protocol - particular about it, exclude anyone who deviates
As treated - ass, analyse based on what people end up doing
Discuss the importance of p values in ANOVA
statistically significant ANOVA
(ie, p-value is below the significance level, usually 0.05) indicates only that at least 1 group mean is different from the rest. Therefore, further evaluation is required to determine which group means differ. It is possible to use confidence intervals (CIs) to determine whether means of >2 groups are different.
Which way is proven to have highest impact on medication reconciliation
Interventions that target pharmacy personnel and high-risk patients in particular appear to have the most impact on the quality of patient care.
In a mean difference, what would statistical and clinical significance mean
The difference in means; would determine clinically significant, given that it’s relevant to what is in dicussion. The difference is statistically significant if (p = <0.05).
4 measurement bias
Observer bias. Attention bias (Hawthorn effect) Expectation bias. Verification or workup
bias.
RRR equation
RRR = ARR/risk in control