Public Health Sciences Flashcards
1
Q
Observational studies
- cross sectional: what is it, question asked, measures
- case control: what is it, question, example
- cohort: what is it, goal, question for prospective vs retrospective
- crossover: what is it, conditions, measures
- twin- what is it, measures
- adoption- what is it, measures
A
- frequency of dx and risks are assessed in present, “what is happening?”; dxx prevelance, can show risk factors but not cause
- compares group w/ dx to group w/o, looks to see if odds of prior exposure or risk factor differ by disease state; “what hapened”, pts with COPD had higher odds of smoking history than those w/o COPD
- compares group with given exposure or risk to group w/o, looks to see if that group later develops a dx, prospective “ who will develop dx” or retrospective “who developed the dx”
- compares the effect of a series of 2 or more tx on a participant, order in which pts receive tx is randomized w/ washout period between each tx; allows participants to serve as their own controls
- compares frequency with which monoxygotic vs dizygotic twins develop a dx; measures hereitability and influence of env factors
- compares siblings raised by biological vs adoptive parents; measures hereitability and influence of env factors
2
Q
Clinical trials
- what is it
- improved w/
- phases
A
- study involving humans comparing the benefits of 2 or more tx or of tx and placebo
- study that is randomized, controlled, and double blinded
- Is it safe (safety, tox, pharmacokinetics), does it work ( tx efficacy, dosing, adverse effects), is it as good or better ( compares new tx to old txs), can it stay (long term effects)
3
Q
Eval of diagnostic tests
- sensitivity
- specificity
- positive predictive value
- negative predictive value
A
- probability that when the dx is present the test is positive; closer to 100 will indicate low false negative rate; TP/ TP+FN
- probability that when the dx is absent the test is negative, closer to 100 = low false positive rate; TN/ TN+FP
- probability that person who has positive test results actually has dx; TP/ TP+FP
- probability that person w/ negative test actually does not have dx; TN/ TN+FN
4
Q
Likelihood ratio
- what is it
- useful test/ equations
A
- likelihood that given test result would be expected in a pt w/ target disorder compared to likelihood that same result would be expected in a pt w/o targeted disorder
- LR+ (TP/FP) > 10 or LR - (FN/ TN) <0.1 indicates useful diagnostic test
5
Q
Quantifying risk
- odds ratio
- relative risk
- relative risk reduction
- attributable risk
- absolute risk reduction
- number needed to treat
- number needed to harm
- case fatality rate
A
- odds of certain outcome given exposure vs odds w/o exposure; case-control studies
- risk of developing disease in exposed group divided by risk in unexposed group; cohort studies; =1: no association btwn exposure and dx, >1 exposure associated w/ increased disease, <1 exposure associated w/ decrease disease
- proportion of risk reduction attributable to intervention as compared w/ control
- the difference in risk between exposed and unexposed groups
- difference in risk attributable to intervention as compared to control
- number of pts who need to be treated for 1 pt to benefit
- number of pts who need to be exposed to risk factor for 1 pt to be harmed
- percentage of deaths that occur over dx course; if 4 pts die after 10 cases of meningitis then it is 40 %
6
Q
Incidence vs prevalence
- what is incidence
- what is prevalence
- increase in survival time
- increase mortality
- therapy initiation
- faster recovery time
- extensive vaccine admin
- decrease in risk factors
A
- # new cases/ # ppl at risk
- # existing cases/ total # people in population
- increases prevalence
- decrease prevalence
- decreases prevalence
- decrease prevalence
- decrease prevalence
- decrease prevalence
7
Q
Precision vs Accuracy
- what is precision; what effects it; increases cause
- what is accuracy
A
- consistency and reproducibility of a test; random error will decrease it ; decrease in standard deviation and increase in statistical power
- trueness of test measurements
8
Q
Bias: what is it, example
- selection
- recall
- measurement
- procedure
- observer-expectancy
- confounding
- lead time
- length time
A
- non random sampling or treatment allocation so that study population is not representative of target population; berkson - cases and controls selected from hospitals are less healthy than gen pop, attritionparticipants lost to follow up have different prognosis than those that complete the study
- awareness of disorder alters recall by subjectives; pts w/ recall disease recall exposure after learning about similar cases
- information gathered in systematically distorted manner; participants change behavior once aware they are bein watched or using faulty equipment to gather information
- subjects in diff groups not treated same; pts in tx group spend more time in highly specialized hospital units
- researcher belief in efficacy of tx changes the outcome of the ts; observer expecting tx group show signs of recovery more likely to develop positive outcome
- factor related to both exposure and outcome distort effect of exposure on outcome; uncontrolled study shows association between drinking coffee and lung CA but coffee drinkers smoke more accounting for association
- early detection confused w/ increased survival
- screening test detects dx w/ long latency period while those with short latency period become symptomatic earlier; slow progressing CA easier to detect w/ screen test rather than rapidly progressive
9
Q
Statistical distribution
- measures of central tendency: includes
- measure of dispersion: standard deviation vs standard error
- normal distribution: shape
A
- mean= avg, median, mode
- standard deviation: how much variability exists in set of values around the mean of these values; standar error: estimate of how much variability exists in setof sample means around true population mean
- bell shaped (gaussian)
10
Q
Non-normal distribution
- bimodal: suggests
- positive skew: mean, median, mode; looks like
- negative skew: mean, median, mode; looks like
A
- suggests 2 diff population
- mean> median> mode, asymmetry w/ longer tail on right
- mean < median < mode, asymmetry w/ longer tail on left
11
Q
Stat hypotheses
- null
- alternative
A
- hypothesis of no difference or relationship -> no association between dx and risk factors
- hypothesis of some diff or relationship; some association between dx and risk factors
12
Q
Outcome of statistical hypothesis testing
- correct
- type I error, alpha
- type II error, beta
A
- stating there is difference when one occurs or stating there is no difference when there is not one
- stating there is an effect when none exists
- stating that there is not an effect when one exists
13
Q
Confidence interval
- what is it
- often used
A
- range of values within which true mean of population is expected to fall w/ specified probability
- 95% w/ alpha = 0.05
14
Q
meta analysis
- what is it
- estimates
- improves
A
- method of statistical analysis that pools summary data from mult studies for more precise estimate of the size of an effect
- estimates heterogenocity of effect sizes between studies
- improves strength of evidence and generalizability if study findings
15
Q
common stats tests
- t-test
- ANOVA
- chi-square
A
- checks diff between means of 2 groups
- checks diff between means of 3 or more groups
- checks diff between 2 or more percentages or proportions of categorical outcomes not mean
