Step3-34 (biostat)

ismailalmokyad's version from 2018-01-27 22:28


Question Answer
selection bias examples1-nonresponse
4-berkson fallacy .
measurement (informational bias) examples1-recall bias
2-observer bias
wt is berkson fallacy biasstudy in hospital patients that doesn't represent general population
wt is referral biasstudy sample referred from practice that doesn't represent general population, for example university hospital refer more sick ppl than general hospital
wt bias raise if so many ppl lost follow upit is type of ***selection bias*** bc now the remaining sample may not reflect general population
wt is non-response biasoccur when study design allows subjects to decide whether or not to participate, so healthy or so sick ppl may not response to the study
wt is prevalence bias(neyman bias)occur when incidence of disease is estimated based on prevalence, for example comparing CAD in pt with DM and pt without DM, the result may be biased bc maybe pt with DM died bc of CAD and that is why the rate of CAD is lower in DM pts
wt is susceptibility bias occur when treatment depend on the severity of disease, for example pt with CAD who is healthy will go cath where pt who is sick may treated medically and the bias would be that pt with cath has a better out come than medicall intervention but the reallity medical intervention sample pts were sicker to begin with.
-wt is the measurement misclassified bias
-what is verification bias,
wrong labeled information for example the out come of exposed pt was labeled for non exposed pts

-verification bias is when the gold standered can't be done on all the negative pt (eg it has s/e or complications), to solve that the researcher may take a sample from the negative group and preforeme the gold test on them and generalize it to all the negative group.
wt is recall biasit is overstimation of the effect of exposulre bc pt who has bad out come will most likely report more exposure than pts who do not have the bad outcome
wt is observer bias (ascertainment)it is also called detection bias or assessment bias, it happen when the investigator knows who is exposed.
wt is stratified analysis and what is sensitive analysis-stratification is the process of removing a cofounder from a study
-sensitive analysis is when the investigator repeat primary analysis calculations after modifying certain criteria or variable ranges; the goal is to determine whether such modifications significantly affect the results initially obtained.
wt is the cofounder charactaristic must be related to both exposure and outcome.
how to limit confounding1-randomizaton
2-matching is another way(same race, neighborhood)
3-restriction which is limiting study inclusion criteria(age or sex) but this will limit generalizability (external validity)
wt is the external validitythe ability to generalize the study result on general population.
1-wt is effect modification
2- what is healthy worker effect (HWE).
modified the exposure with another factor before looking to the outcome for example looking for family hisroy of breast cancer before conculding that OCP cause breast cancer, or looking for smoking before conculding that estrogen therapy cause DVT.

(HWE) Working populations are generally healthier than the general population and often exhibit lower mortality rates(sick ppl do not work), so compare their mortality rate to another group of workers who is not exposed to the risk that is being studied
wt is the best way to deal with pt in a clinical trial who are not complaint with their medications best way by keeping them in the same group and count them on the analysis, this is called intention to treat.
wt are is intention to treat is when a pt is not complaint with the study but you still keep him in his group and count him in the result
wt is per protocol method that used in clinical trial studies is when a pt is not complaint with the study and you remove the pt out of the study and out of the result too (it affect the result )
wt is as treat method that used in clinical trial studyit is when a pt is not complain with the study and you move the pt from the treated group to the controll (he is not taking his medication), affect the out come result badly
how many lies between one SD, 2 SD, and 3 SD68%, 95% and 99.7%
wt is the Z-score and how to measure it you subtract the mean from the value then divid it the the stander deviation, this helps to know how many stander deviations this result is above the mean
wt is Two-sample t test (also called student's t test)it is used to compare means of two independent groups. the result of test gives the T statistic that will then used to calculate the P value.
wt is the paired t test it is used to compare the means for two dependent groups(unlike two-sample t which used for means of independent groups). for example comparing the mean of pt BMI before and after treatment.
what to use to analyses the result of a categorized groups study for example the ESR high or low in ppl using OCP or not using OCPif small size study use fisher's exact test, if large sample study use chi-square test.
wt to use to analyses more than two independent groups analysis of variance (ANOVA)
what is the survival analysis is the time to event data analysis during follow up studies and clinical trials.
wt is latent period (latency)the effect of the intervention need long time follow up to be statistically significant
what is type one errorconcluding that there is an association between exposure and out-come when in fact there is none
**** reject the null when it is true****
what is type two errorconcluding that there is no association between exposure and outcome when in fact there is one
*** accept the null when it is wrong***
what is the probability to committing type 1 error it is called alpha (same as P value)
what is the probability to committing type 2 error is called beta
what is 1-Bit indicate the probability of detecting an association if it exists in reality and is refferd to as the power of the study.
what is the factors that affect the power of the study1-alph level, lowering the alpha level will decrease the power of the study.
2-the magnitude of difference in outcome between the study groups(subtle difference is more difficult to detect than a big one.
3-increasing the sample size will increase the power and increase the porbability of detecting the difference.
what is Hawthorne (observer) effecttendency of study subjects to change their behavior as a result of their awareness of being studied.

Recent badges