Hadpop2 Flashcards
What is a census
simultaneous recording of demographic data to all persons in a defined area
What is a census useful for
- allocation of resources
- trends in populations, eg ethnicity
- projections of populations (how many people living in this area with this defined problem)
3 ways to measure Birth rates
- Crude BR: # of live births per 1000 population
- General fertility rate: # of live births per 1000 fertile women between 15-44 - more accurate but not always possible
- Total period fertility rate: average # of children born to a hypothetical woman in her lifetime
How can death rates be measured
- Crude mortality rate
2. age-specific mortality rate
Define incidence
Number of new cases per year in a set population, per 1000 people/per year
Define prevalence
Number of people with the disease in a set population
What is SMR
Standardized Mortality Rate
calculated by: observed # of deaths x 100/expected # of deaths
SMR>100 suggests excess mortality and <100 suggests less mortality with confounders accounted for
What is the p-value
indicates the percentage of the study due to chance, based around 5%
State 3 types of selection bias
- allocating
- healthy worker effect
- non-response bias
State 4 types of information bias
- instrument
- interviewer
- recall bias
- publication bias
Define a cohort study
Recruiting disease-free individuals and following up with them for an extended period of time, and classify them according to their exposure status
State 2 types of cohort studies, and define each one
- Prospective: disease-free individuals recruited and followed up
- Retrospective: choosing disease-free individuals based on a common past exposure and following up
What’s an internal comparison
Occurs when you have sub-cohorts within your original group and compare the exposed to the unexposed
What’s an external comparison
Comparing your exposed group to a reference population
What is the only study design that allows you to calculate odds ratio?
How would you do it?
Case-control studies
Case: diseased individuals, control: disease-free individuals. Want them as similar as possible
a: exposed cases
b: exposed controls (disease-free individuals)
c: unexposed cases
d: unexposed controls
a/c over b/d
A crazy big dick:)
List 4 advantages for Case-Control
- used for rare diseases
- temporal sequencing
- can look at multiple exposures at once
- Fast and inexpensive (relatively)
Define temporal sequencing
Confidently say that exposure came before the outcome
What is intention to treat
Treating everyone the same regardless if they are given the real drug or placebo in an RCT
Why is it important to keep non-compliars in a study
If non-compliars (those non-compliant) with the drug are removed, loss of randomization also occurs from the trial
What is the purpose of the Bradford Hill Criteria
To determine whether the causal-effect relationship has been established
Can you directly say this exposure caused the outcome?
List the 9 Bradford Hill Criteria
STD-C-CRABS
Strength of association: stronger the association between exposure and outcome, less likely due to chance, confounding or bias
Temporal sequencing
Dose-response: does the outcome change with increasing/decreasing dosage
Consistency of association; does same association occur in other studies
Coherence of Theory: does the observed observation confirm your scientific thinking
Reversibility: if you remove the causal factor, would you get a reduced risk of the outcome
Analogy: does a similar disease have similar outcomes
Biological Plausibility: is there a plausible biological mechanism to support the theory
Specificity of association: the outcome is associated with a specific factor
Three ways losses can be minimized by, aka three ways to minimize loss to follow up
- Follow up appointments
- No coercion or inducements
- Honesty
What is loss to follow up and when is it significant?
When a patient stops participating in the study before its completion.
A weakness of RCTs and cohort studies, since your results depend on the future results, so you don’t want your patients to drop out of the study
Three ways to maximize compliance of patients
- Simplified instructions
- Patient allowed to ask lots of questions
- Made simple and accessible for patient
How and when would you calculate RR
For RCTs:
risk of exposed with disease/total = x
risk of unexposed with disease/total = y
x/y= “Value”, Risk of the exposed getting the disease is “Value” times as likely as a nonexposed individual.
List the 4 pillars of medicine
Principle of…
- Beneficence (do good)
- Non-maleficence (do no harm)
- Autonomy (let the patient decide)
- Justice (be fair)
What is a systemative review?
Compilation of all primary studies (many RCTS), you narrow them down the most relevant and credible
List 3 characteristics that a study used in a systemative review should have
- Transparency - nothing hidden
- Reproducible - can you reproduce the exact same study again
- Explicit - clear
What’s a meta-analysis and when is it used
Within a systemative review that provides an overall value with associated CI.
Used in a forest plot
In a forest plot, what does…
a) The square represents
b) the horizontal lines
c) the size of the square
d) the diamond
e) solid vertical line
a) The odds ratio value
b) the CI
c) the amount of weight on the study
d) The pooled estimate: the average of the studies, the width of the diamond represents the overall 95% CI
e) The Null Hypothesis
List the two types of models that can be used in a meta-analysis
- Fixed effects model - assumes the studies used are homogenous, they’re so similar that any variation between data comes from within the study
- Random effects model - assumes studies are heterogeneous and variation between data comes from in the study and between separate studies
What is a funnel plot
Publication bias can be determined using a funnel plot…
Publication Bias: whereby studies are only likely to be published if results are statistically significant or have a large sample size
The more funnel-shaped, the less publication bias
List 3 advantages of RCTs
- You can introduce blinding
- eliminates confounders
- you can actually test the intervention
How is Prevalence calculated
of cases at a given time/population # at the time
How is Cumulative Incidence calculated and expressed in units, and what does the value mean?
of new cases of a disease in a period of time/# of people at risk of developing the disease at the start of the same time period
As a percentage
How many people at risk actually got the disease
How would you calculate Incidence Density and what does that value show?
of new cases of a disease/Total person-time at risk of developing the disease
Shows how many people who were at risk for x amount of time actually got the disease
List the general steps for a Cohort study
- recruit a population of disease-free individuals
- classify them based on their exposure status, e.g; How many of them have had the BCG
- Follow up with them for an extended period of time
What does Internal Validity measure
Are the finding of the study valid for the population in the study
What does external validity measure
can the findings be generalizable to other populations outside of the study?
3 advantages for a cohort study
- temporal sequencing
- can examine multiple outcomes (the final results) from one exposure
- Good for rare exposures (but bad for rare outcomes)
What three disadvantages do RCTs and cohort studies have in common
- Time
- Expensive
- risk of loss to follow up
List two disadvantages for a case-control
- Difficult to select the appropriate control groups as you want them to be as characteristically similar to the case group, (asides from lack of exposure)
e. g; similar age, gender, height, etc.
Define a confounding factor and list 5 examples
Is independently associated with the exposure and the outcome but is not on the causal pathway
age, gender, genetics, height, smoking (when it’s not the exposure being examined)
If a CI value includes the null Hypothesis it is…
Not statistically significant, therefore we accept the Null Hypothesis
If a CI value doesn’t include the null hypothesis it is…
statistically significant, therefore we accept the alternate hypothesis
What’s a case series?
The absolute fcking worst type of study design, based on reports
What does the incidence rate ratio show
It compares the incidence rates between groups with different levels of exposure
What does the relative risk show
Compares cumulative incidence (a proportion) between groups with different levels of exposure
List in order from top –> bottom the study design triangle
- Systemative review
- RCT
- Cohort studies
- Case control
- Cross sectional
- Case series
