Epidemiology Final Exam Questions and Answers

what is bias? - error that is built into the study causes the results to be different from the truth (overestimation or underestimation of the association between exposure and outcome) where is the best place to prevent bias during a study? - the design stage (before you even look for results) name the four type of bias common in cross-sectional studies - selection bias interpretation bias recall bias non-response bias name the common types of bias in case-control studies? - selection bias observation bias what are the two types of bias that observation bias encompasses? - recall bias interviewer bias what is selection bias? - bias in how study subjects are selected or excluded from the study (ex. assessing college students then only using hofstra students) what is recall bias? - when the patient recalls information differently when it actually happened common with studies that use surveys or questionnaires what are the two types of selection bias? - control selection bias self-selection bias what is control selection bias? - control patients are selected from a different population that the case patients the case patients should be taken from the same population as the control patients what is self selection bias? - different rates of participating between patients who are cases and patient who are controls participant blind studies help remove this bias bc the patient then doesn't know if they are a control or a case what is diagnostic test bias? - patients get hospitalized differently based on their specific exposure women who take OCPs with a VTE are more likely to be hospitalized what is berkson's bias? - hospital patients are NOT representative of the general population because theyre older and more sick than an average person what is prevalence-incidence bias (aka survivor bias)? - estimation of a disease is taken at a specific POINT IN TIME vs information taken over a TIME PERIOD common in cross sectional studies since they are at one point in time what is information bias? - the measure of the data is incorrect and thus you are getting data that is not true ex. faulty BP cuff common selection bias' in cohort studies? - healthy worker effect loss to follow up what is the healthy worker effect? - workers are usually younger and healthier than the general population - a problem if the cohort is made up of workers what is the cohort effect (aka generation effect) - the study population is effected by events at the current time and cannot always be extrapolated to a population at a different time what is confounding? - a variable that is associated with both the outcome and exposure separately ex. smoking is associated with both coffee and heart disease is a large sample size good? - yes it is good what do odds ratio and relative risk measure? - they measure how related an exposure and health event are what does and RR or OR of 1 mean? - random chance, there is no extra risk or less risk if the RR or OR is over 1, what does that mean? - there is a positive association between the exposure and the outcome if the RR or OR is less than 1, what does that mean? - there is a negative association between the exposure and the outcome when is risk ratio (RR) used? - incidence based prospective studies ex. cohorts and RTCs when is odds ratio (OR) used? - for retrospective studies (case-control studies) what do you use to calculate the OR and RR? - 2X2 contigency table formula for OR? - (a x d) / (b x c) RR formula? - risk in population A divided by risk in population B a/(a + b) / c/(c + d) what is the most common p value? - less than 0.05 what is the most common confidence interval? - 95% what is the effect of sample size on confidence interval? - larger sample size -- narrow confidence interval what is attributable risk? - amount of disease that is due to the specific exposure that is being discussed ex. what percent of lung cancer is caused by smoking what is the population attributable risk? - incidence of disease in the population that is due to the exposure what is another name for an experimental study? - intervention study what is between group design? - outcomes are measured between two or more groups receiving different levels of the intervention what is a non randomized study? (aka convenience sample study) - you make the comparison and experimental group with a NON-RANDOM process what is a single blind study? - subjects are blind what is a double blind study? - subjects and investigators are blind what is a triple blind study? - subjects, investigators, and data analysts are blind what is the benefit of blinding patients? - placebo effect is constant with all participants because they all think they are getting the intervention Can RCTs demonstrate a cause-effect relationship? - yes they can thats why they are the gold standard of epidemiological studies what is the purpose of phase 1 drug trials? - determines safety in humans (no blinding, no control study with a few amount of patients) uses patients with advanced disease that have little options what is the purpose of phase 2 drug trials? - blinded trial with a small test group that tests for safe dosages and side effects what is the purpose of phase 3 drug trials? - large trials with thousands of patients what are the four ethical principles in research on human subjects - - maximize good - avoid doing harm - respect for all persons - fairness to aill what is causality? - a cause must precede the effect what is hills criteria for causal association? - 9 criteria that determine if a cause/effect relationship exists what are the 9 criteria in hills criteria? - 1. strength of the association 2. consistency 3. specificity 4. temporality 5. biological gradient 6. plausability 7. coherence 8. experiment 9. analogy what is temporality? - the cause PRECEDES the effect what is a continuous or ratio variable type? - numbers on a scale that can take on ANY VALUEEEEEE (1.1 can exist and 100 can exist and 100.34582 can exist) what is count or interval variable type - numbers that can only be integer values no DECIMALS NONEEEEEEE what are the three types of categorical data? - binomial nominal ordinal what is categorical binomial data? - variables with only two responses or options (yes/no type situation) what is categorial nominal data? - variable with more than 2 responses but no order to them NO ORDERRRR ex. blood type, marital status