Final Flashcards
What is descriptive epi and what questions does it ask
defines frequency and distribution of diseases and other health related events
how many, who, where, when
What is analytic epi and what questions does it ask
analyzes determinants of health problems
how, why
What are the different types of descriptive epi studies
case reports
case series
ecological
cross-sectional
What is a case report study
a single occurence, description of symptoms, signs and diagnosis
What is a case series study
several occurrences with common features
what is an ecological study
study of aggregate data on populations
what is a cross-sectional study
snapshot of a study pop, data at indiv level
What is a big issue in ecological studies
temporality
What are some strengths in cross-sectional studies
- data at indiv level
- Subgroups based on E+ or O+
- generalizeable
- helps identify resource needs for hlth intervention
- generate new hypotheses on relationship between exposure and disease
What are some limitations of ecological
- No objective measure of the temporal ordering of exposure and disease
- Cannot examine causality
- Identify high proportion of prevalent cases of long duration
- May under-represent diseases with short duration
What studies are analytic
Observational:
case control
cohort
Experimental: RCT
What are you analyzing in analytic epi
- Hypothesis
- Exposure causes outcome/ exposure precedes the outcome
- E-> O
- If the exposure casus the disease, then exposed indiv will have more disease than the unexposed indiv
What are the 4 causal criteria you need to know?
Temporality:does the exposure precede the outcome
Consistency
Strength of association: stronger association more credible the results will be
Biological gradient (dose response): the more of the exposure the more disease you can have. Is the correct type of medicine being used and dosing at the right amount
What is the difference between observational and experiemental
O: investigators do not intervene on or control study participants’ exposure status
E: investigators assign an exposure to study participants
Describe a case control study
- groups selected in terms of whether they do or do not have outcome of interest
- groups compared retrospectively with respect to exposure to potential risk factors
Describe cohort study
- indiv selected based on their exposure
2. incidence of outcome over time is assessed
what is a case control study design
- Start with outcome those who have the disease (cases)
- Select controls those who do not have the disease (controls)
- Then examine if each group was exposed or not exposed
Advantages of case-control studies
- Fast and inexpensive
- rare diseases
- diseases with long delay between E+ and O+
- Population is dynamic
- Look at multiple E+
6.
Disadvantages of Case-control studies
- 1 hlth O+ at a time
- inefficient for rare diseases
- Temporal relationship between E+ and disease is unclear
- Participant recall
- Cannot measure disease frequency since indiv with disease are pre-selected by investigator
Are case control studies
Retrospective or prospective
How are indiv selected?
What measure do you use
- Retrospective
- Selected on their health outcomes (cases) - Disease - No disease
- Measure prior exposure between disease and no disease
- Odds Ratio
What is an odds ratio
OR = ad/bc
What are the interpretations of OR
=1
>1
<1
OR = 1 – exposure has no effect
OR > 1 – exposure is a risk factor
OR < 1 – exposure is a protective factor
What is the classic cohort study design
- there is a defined population that is not randomly assigned
- Then they are put into exposed or not exposed
- within each group there is disease or no disease.
Why choose prospective designs?
More control over cohort selection, exposure measurement, follow-up procedures, and outcome measurement
Greater ability to account for other variables (i.e. confounders)
Who choose retrospective designs?
Faster and less expensive
Existing records available to reconstruct the cohort, assess exposure history, and assess outcomes
Strengths of cohort studies
Can directly measure incidence outcomes
- Temporality of E+ and O+ is clear
- Good for common outcomes that occur within a relatively short time frame
- Useful for rare exposures: can study multiple
Limitations of cohort studies
- Need large samples
- Inefficient for rare or very delayed outcomes
- Potential for confounding when E+ is not randomly assigned
- Can be expensive and labor-intensive
- LTFU
- continuity of staff and funding is uncertain
What are the three possible sources of bias when evaluating a screening program that may result in a false picture of its efficacy:
- Volunteer Bias (People who choose to participate in the screening program may be healthier or at higher risk of developing the disease than those that don’t participate)
- Lead-time bias (Lead-time is the amount of time by which the diagnosis was advanced due to screening. Lead time bias means that survival may erroneously appear to be increased among screen-detected cases simply because the diagnosis was made earlier in the course of the disease.)
- Length-biased sampling (•Slowly progressing forms of a disorder tend to be more easily detected and have a better prognosis than rapidly progressing forms•Screening tests appear to be more effective than they are when comparing screened to unscreened populations in term of survival rates)
Definition of incidence
Number of new cases that occur during a specified period of time within a population at risk
What is the RR formula
CI among exposed/ CI among unexposed
A/ (A+B)/ C/ (C+D)
What is the AR formula and interpretation
CI among exposed- CI among Unexposed
Excess amount of disease due to the exposure
what is the AR % formula and interpretation
(CI among E+ - CI among Unexposed)/ CI among exposed
The percentage of disease in the exposed that can be attributable to the exposure
Recall bias
Cases may be more likely to accurately recall an exposure than controls
E.g., people diagnosed with a serious illness are likely to spend a lot of time thinking about what might have caused it
Controls have not spent this time reflecting on their past, so may less accurately remember exposures
Cases may be incorrectly remembering their exposure history
Reporting Bias
A.k.a. social desirability bias, where participants are reluctant to report exposures because of social norms
E.g., studies of risky sexual behaviors; studies of illegal drug use
Observer bias
Observers may have preconceived expectations of what they should find in an examination
In abstracting electronic records, differential recording of exposures in the records of cases vs. controls
Hawthorne effect
behavior of participants changes because they are aware of being oberved
which study is especially prone to selection bias
case-control
What is selection bias
A systematic distortion of findings that results from the processes used to recruit study participants
Why is selection bias important
If the process by which participants are recruited favors, or overlooks, certain types of participants, then the study population will not be representative of the population for which we are attempting to obtain estimates
What is confounding?
- Mixing or blurring of effects
- Occurs when the effect of a third factor is measured instead of the association between exposure and outcome
- Meets 3 criteria
- Associated with exposure
- Affects the outcome
- But is not intermediate in causal chain between exposure and outcome
What are ways to adjust for confounding?
Design phase: -randomization -restriction -matching Analysis Phase: -Stratification -Mathematical modeling
What is effect modiciation
Occurs when the magnitude of the effect of an exposure on an outcome differs depending on the level of a third variable
Third factor is known as the effect modifier
Also called interaction
effect modification is a biological phenomenon
- If R1 and R2 are the same but are different from RR crude
If R1 and R2 and RR crude are all moving in the same direction you have
If RR1 and RR2 not moving in the same direction you have
If all three are equal you have
- confounding
- EM and Confounding
- EM but no confounding
- nothing
what is the RCT design
Start with study population
then they are randomly assigned to current treatment or new treatment
Then are studied to see if they improve or did not improve
what does randomization protect against
confounders
selection bias
-Participants’ assignment to a treatment or intervention group is determined after enrolling in the study
information bias
-Exposure status is assigned, not dependent on memory or records
-Participants and investigators are usually “blinded” or “masked” to participant assignment
what are the types of masked or blinded studies
Single-blind study- participants are unaware of whether they are in the experimental or control group
Double-blind study- both participants and observers are unaware of the participants’ group assignments
Triple-blind study: participants, observers, and data analyst are all unaware of group assignments
what are strengths of RCT’s
Protects against confounding
Observed factors
Unobserved factors
Protects against bias
Provides the best evidence for causality
What are the limitations of RCT’s
Cannot always randomly assign a treatment (exposure)
Very expensive
Not suitable for rare outcomes
Only as strong as the protocol
Masking/blinding
Adherence
Screening in Public Health and Medicine
Definition:Presumptive identification of an unrecognized disease or defect by the application of tests, examinations, or other procedures
Purpose:
Delay onset of symptomatic or clinical disease
write out the 2x2
Test result on left
Screen positive
SCreen negative
Disease status
With disease
Without disease
A= true positive b= fast positive c= false negatie d= true negative
what does true positive, false positive, false negative, and true negative mean
True positive= heave disease and test was positive
B= no disease but test was positive
c= have disease but test was negative
d= no disease and test was negative
Sensitivity vs specificity formula
Sensitivity= with D+ A/(A+C) = TP/ (TP + FN)
Specificity= W/o D+ D/(D+B) = Tn/(Tn+FP)
Positive Predictive Value:
TP/(TP + FP)
Given the test is positive, what proportion have the disease (top half of table)
Negative Predictive Value:
TN/(FN + TN)
Given the test is negative, what proportion are without disease (bottom half of table)
Sensitivity in words
Sensitivity: What proportion of people who actually have the disease test positive?
Specificity in words
Specificity: What proportion who do not have the disease test negative?
Power is the probability of correctly concluding that the two treatments (e.g., intervention vs. control groups) differ.
True
For a dose-response relationship, the threshold refers to:
The lowest dose at which a particular response occurs
In a stratified randomization RCT, subjects are first randomized then stratified by potential confounding variables.
False
Screening identifies an illness that would not have shown clinical signs before death from other cases
Over diagnosis bias- •This situation happens when the screening identifies an illness that would not have shown clinical signs before death from other causes
Non-differential misclassification is the probability that misclassification does not vary for the different study groups.
True
An ideal screening program occurs before the on-set of clinical symptoms of the disease.
True
Background risk is the amount of low-level risk that only unexposed individuals have
False
H. pylori is clearly linked to peptic ulcers. Over a 10-year period of time (prospective cohort study), about 11% of people with the bacteria went on to develop peptic ulcers.
Temporality
In a case-control study comparing people with STIs to those without STIs, the exposure of interest is number of sexual partners. Cases are less likely to under-report the number of sexual partners they have had, when compared to controls
Differential
A _run-in design__ is particularly useful for increasing the proportion of study participants who adhere to the intervention and follow-up procedures.
True
what are the interpretation of OR
OR (disease) = the odds of being a case if you were exposed are [OR] times the odds of being a case if you were unexposed
what are main concerns with designing studies
Validity-
Precision
What is non-differential misclassificiation
Subjects are erroneously categorized with respect to either exposure or disease status
Proportions of subjects erroneously classified are approximately equal
Increased similarity between the exposure (or disease) groups
Biases association between exposure and disease towards the null
What is differential misclassification
Proportions of subjects misclassified differ between the study groups
Observed estimate or effect can be biased in the direction of producing either an overestimate or underestimate of true association (toward or away from the null)
What are the subcategories of selection bias
Non-response
Berksonian-
Neyman-
Healthy Worker effect-
What is non-response Bias
- People who respond to a study often differ systematically from people who do not respond
- Non-response can occur at recruitment or follow-up
- Recruitment procedures miss eligible participants
- Eligible participants not interested or difficult to reach
- Study participants were lost to follow-up in non-random manner
Why is non-response bias important
- Generally the reasons for non-participation are different between cases and controls
- Often information on reasons for nonparticipation is not available
- We can’t know whether non participation is based on exposure status
- Low participation potentially introduces bias
- Differences in recruitment of cases and controls may result in incorrect estimates of causal associations and wrong study conclusions
What is berksonian bias
- A form of selection bias that applies to hospital-based epidemiologic studies
- People in hospital are likely to suffer from multiple diseases or engage in unhealthy behaviors (e.g., smoking, drinking, food-related issues, etc.)
- As a result, hospitalized study participants are not typical of the communitypopulation
Why is bersonian bias important
may result in larger odds ratios than actually exist
What is Neyman bias
- A.k.a. prevalence-incidence bias
- Refers to selective survival among prevalent cases
- If persons with more severe disease die quickly, those available for study are not as sick
why is neyman bias important
May lower ORs and mask strength of causal association
What is the healthy worker effect
•As a result, workers may not be typical of the larger population.
Ill and disabled people are less likely to be employed
•A form of selection bias that affects studies in occupational epidemiology
•Employed workers are healthier than other segments of the population
Why is healthy worker effect important
- May reduce the validity of the exposure data
* Estimates of association may be lower than actually exist
What are the two experimental units and what is the difference
Individual- individuals are randomized to intervention or control conditions
Group- groups of people are randomized to an intervention or control condition
PPV in words
PPV: What proportion of people who test positive actually have the disease?
NPV in words
NPV: What proportion of people who test negative actually do not have the disease?