Study Designs
What is an ecological study? What fallacy can occur?
- populations/groups studied instead of individuals
- group level data can be obtained
- ecologic fallacy: association observed between variables isn't representative of the association at an individual level causing error (uncontrolled factors)
What is mass vs. selective screening?
mass: everyone in pop. regardless of risk (newborns)
selective: high risk groups screened (STIs, genetic)
What are the five infectious agents?
bacteria, virus, fungi, protist, prion
What's the difference between a one-sided and two-sided t test?
one-sided: better
two-sided: better or worse
What is selection vs. information bias? What are some other biases we've discussed?
selection: how cases and controls are chosen is an apparent association (exclusion, compensation) -> systematic error
information: obtaining information is flawed so exposure/disease info is wrong (misclassification)
hawthorne, recall, reporting, surveillance
What is an experimental study? What can occur with experimental studies?
- intervention
- intentional change
- randomized control trial or quasi-experiment (individuals vs. an area being exposed)
- clinical trial using randomization
- single, double, no blinding -> preventing observer bias
What are the types of prevention?
primary: pre pathogenesis (vaccine, diet)
secondary: early pathogenesis (screening)
territory: late pathogenesis (insulin, PT)
What is active vs. passive immunity?
active: natural infection or vaccine -> long
passive: antibodies from another person, transplacental antibodies -> short
What information do we need to determine a necessary sample size?
- difference in response rates
- estimate of response rate in one group
- what alpha do we want
- what power do we want
- do we want a one or two sided test
What is syngerism vs. antagonism?
synergism:pos. interaction where effect is greater than expected
antagonism: neg. interaction where effect is less than expected
What is the difference between case reports and cross sectional studies?
case report: single or a few patients with a rare disease
cross-sectional: study examining data from a larger population to observe prevalence
What are the screening test categories?
newborn: state level at hospital after birth for hearing, heart, blood screening
children/adolescents: developmental early screening, vision, dental, behavioral risk, chronic risk
adults: cancer, CVD, diabetes
genetic: cancers, PD, recessive diseases
others: HIV, mental health
How can infectious agents transmit?
Direct: person-to-person through direct contact or droplets
Indirect: vehicle borne where infection comes from contaminated non-moving objects, airborne where particles are present in the air creating dust, vector borne where insect transmits to host
You have a portal of entry and exit with these!
What is our alpha, beta, and power?
alpha: probability of making a type I error
beta: probability of making a type II error
power: 1 - beta i.e. probability to correctly conclude if that treatments differ
How do we determine if there's an interaction?
Has an association been observed?
Is it due to confounding?
Is the association equally strong in the strata?
What is a case-control study? What do you use to measure it?
- start with cases and controls
- look backward to compare past exposures
- group vs. individual matching
- odds-ratio = (cases w exposure/cases w.o. exposure)/(controls w exposure/controls w.o. exposure)
- (A/C)/(B/D) = AD/CB
- positive association, 2x more, protective factor, no association
What is reliability vs. validity?
reliability: are results consistent
validity: are results accurate
screening tests should be both or can be invalid and reliable!
Define these terms:
Herd immunity, incubation period, subclinical infection, carrier, index case, zoonosis, bioterrorism, emerging infection
bioterrorism attack: deliberate release of agents to cause illness in organisms
emerging infection: newly appeared in a population or known for some time but incidence is at a rapid incline
zoonosis: vertebrate animal to human
index case: patient zero
carrier: unknown source of infection
subclinical infection: doesn't show obvious signs or symptoms
incubation period: time from invasion of agent to first signs of disease
herd immunity: large proportion are immune/vaccinated to protect whole community
What are recruitment and ethical issues involved with these errors?
- not having enough people to get results
- informed consent and voluntary participation
- benefits, risks, vulnerable groups, coercion
What is the additive model vs. the multiplicative model? What is used for each?
multiplicative: one exposure multiplies to the effect of the other (relative risk)
What is a cohort study and what are the different types? What do you use to measure it?
- look at exposed vs. non-exposed to see who gets the disease and who doesn't
- prospective cohort: present to future- retrospective cohort: past to present
- historical prospective: past to future
- relative risk = (exposed and gets disease/total exposed)/(not exposed and gets disease/total not exposed)
- (A/(A+B))/(C/(C+D))
- same risk, 2x as high, protective effect
What are the measures of validity? What are the four results you can get? Be specific.
sensitivity: ability of test to identify all correctly screened individuals who have disease (a/(a+c))
specificity: ability of test to identify all correctly screened individuals who don't have disease (d/(b+d))
pos. predictive value: proportion of individuals screened positive who have disease (a/(a+b))
neg. predictive value: proportion of individuals screened negative who do not have disease (d/(c+d))
false pos, false neg, true pos, true neg
What are Hill's criteria? Briefly describe each.
strength
consistency
specificity
temporality
biological gradient
plausibility
coherence
experiment
analogy
How would you set up a table for type I and type II errors?
correct decision: we conclude treatments aren't different and treatments aren't different
correct decision: we conclude treatments aren't different and treatments aren't different
type I error: we conclude treatments are different when treatments are actually not different
type II error: we conclude treatments are not different when treatments are actually different
What is preventative vs. therapeutic study?
preventative: prevention of health problems (primary, secondary, tertiary)
therapeutic: existing health issues (meds, surgery)