Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
69 Cards in this Set
- Front
- Back
|
What is epidemiology? |
The basic science of pop health. The study of (1) Frequency disease occurs in diff pops, and (2) Factors determining variance btwn them |
|
|
What is a population? |
A group of people who share one or more common features |
|
|
What is a disease occurrence? |
Diseased state occurred from non-diseased state |
|
|
What is estimate of association (effect)? |
Differences in occurrence |
|
|
Define E = (N/D)/T |
Numerator: Those in the denominator to whom disease occurs Denominator: Number of people in study population
Divided by time |
|
|
What are the components of GATE frame? |
Triangle: Population Circle: Exposure and comparison Square: Outcomes (a,b top c,d bottom) Arrows: Time (V: incidence H: prevalence (T=1) |
|
|
What are EGO and CGO How do you calculate them |
EGO - exposure group outcome (a/EG) CGO - comparison group outcome (b/CG) |
|
|
Name the two data classifications types Are they interchangeable? What is an alternative if they are not? |
1. Categorical (yes/no) 2. Numerical (e.g. BP/BGL)
Numerical can be categorised (low, med, high etc.) and an alternate is to take the average |
|
|
What are the main features of a cohort study Can incidence and prevalence be measured? |
(1) Population are individuals (2) Allocated to EG/CG by measurement (3) F/up and count outcome over time Yes they are both measurable |
|
|
What is a cross sectional study? What are strengths? What are weaknesses? |
Study of prevalence of one outcome in individuals at a point in time
(S) Cheap, best for prevalence, no mx error (W) Reverse causality limits interpretation of effect, confounding common |
|
|
What does incidence measure? Strength and weaknesses
|
Measures easily seen onsets of disease over period of time. (S) Clean measure - rate determined by populations’ disease risk. Measures events, population and time (W) Can be difficult to measure as you have to observe events over time |
|
|
What does prevalence measure? What are EGO/CGO called? Is time included? |
Populations' disease status at one point in time Measures diseases with either difficult to observe onsets (e.g. HTN/DM) or which occur multiple times with varying severity (e.g. asthma attacks). EGO/CGO are prevalence measures and time isn't included |
|
|
What influences prevalence of a disease?
When might high I disease be low P and when might a low P disease be high I? |
Depends on incidence minus death/cure
A low incidence disease could be highly prevalent if no deaths/cures while a high incidence disease could have low prevalence if lots die/cured
|
|
|
Strength/weakness of measuring prevalence? |
(S) Easy to measure as ‘stop time’ and count i.e. ‘static’ measure. (W) Measures events and population only so less info than incidence. Determined by incidence and cure and death rate (a ‘dirty measure’) |
|
|
What is point prevalence? |
Disease measured at one point in time (e.g. rate of HTN today)
|
|
|
What is period prevalence and how do you measure it? |
Time point extended to longer that just one point. E.g. those who get asthma in 2014
1. Define disease e. g. “attack” is 2 attacks/yr
2. Measure if person has had "attack" in that period of time
NB are not measuring number of asthma attacks as we would for incidence!
|
|
|
When would period prevalence be better to measure than incidence? |
In diseases that do have observable onsets if the signs and symptoms come and go frequently – such as asthma attacks. |
|
|
What are strengths and weaknesses of RCT |
(S) Reduce confounding, best study for rx. or tx. questions, (W) ethics, logistics, long f/u (cost), maintenance error common, ?recruitment error (motivated volunteers) |
|
|
What is an ecological study? Strengths? Weaknesses? |
Study outcomes in populations (vs. individuals)
(S) Generally cheap & quick, useful if majority of some populations are exposed but others are not, efficient for rare outcome (W) confounding common
|
|
|
How is RD calculated Does it have units What if EGO/CGO are equal |
Calculate as EGO-CGO Has units (RD definitely does) If EGO and CGO = then RD is 0 |
|
|
How is RR calculated Does it have units What if EGO/CGO are equal |
Calculated as EGO/CGO Has no units If EGO/CGO equal then RR is 1 |
|
|
Which is better to use; RR or RD? Why? |
Better to use risk difference
Relative measures of risks & benefits are deceptive and provide less information than absolute measures and risk differences |
|
|
What is ARR and ARI When do they apply? |
If RD lower in EG it is absolute risk reduction If RD higher in EG it is absolute risk increase
|
|
|
What is RRR and RRI When do they apply? |
RR less than 1.0 is relative risk reduction RR of more than 1.0 is relative risk increase |
|
|
If EGO / CGO use means (average) then RD is called what? |
Mean difference |
|
|
If EGO / CGO use means (average) then RR is called what?
|
Relative mean |
|
|
What does RAMBOM stand for? |
Recruitment, allocation/adjustment, maintenance, blind or objective measurement |
|
|
For a selected population to be a representative sample, it must have been ...? |
Selected randomly |
|
|
You need a representative sample population to measure ...? |
Prevalence |
|
|
What are the two methods of allocation?
What is the potential source of non random error? |
1. Measurement (e.g. blood test or survey) 2. Randomised (coin toss)
Worry about allocation error if allocation not randomised e.g. in Swedish study, did they tell truth re amount of ETOH drunk? |
|
|
What is an adjustment error? How can we avoid it? |
When a confounding variable (difference) exists between EG and CG that should have been accounted for e.g. glue sniffing in heavy ETOH users.
Adjust for it by stratification into multiple groups so those with CV are in one sub group |
|
|
What are the sources of maintenance error? |
1. Compliance 2. Contamination 3. Loss to f/up. |
|
|
What is the BOM error? How to reduce? Is blinding measurer always needed? |
Contamination of measurement of outcomes
Can be reduced by blinding (e.g. knee surgery example) or objective measures e.g. computer read blood test.
If measure is very objective then measurer doesn't need blinding |
|
|
What is random sampling error? What is the consequence? How to reduce? |
Chance error as population never perfectly representative as you cannot include everybody
Reduce with a large sample size |
|
|
What is random allocation error? How can it be reduced? |
EG and CG in RCT may differ by chance alone, particularly if small trial
Reduced by undertaking a larger study |
|
|
What is random measurement/assessment error?
How can it be reduced? |
Measurements of exposure and comparison status and of disease outcomes subject to random measurement error as ability to measure biological factors in exactly the same way every time is poor
Reduce by taking multiple measurements and average them or to use an automatic, more objective, instrument |
|
|
What does 95% CI measure?
What is this range called? |
Amount of random error in a study and describes the range of values likely to include the true result
Range is a measure of precision |
|
|
Define 95% CI |
There is about a 95% chance that the true value in a population lies within the 95% confidence interval |
|
|
What does it mean if there is no CI overlap? What will the CIs not cross? What does this mean? |
EGO and CGO are assumed truly different rather than there being an apparent difference due to random error in the measurements.
CIs will not cross no-effect line and the study results are statistically significant |
|
|
What is recruitment error? What is this called? How can it be avoided? |
Occurs when objective of study is to measure characteristics of specified eligible population but participants are not representative of those eligible. External validity error Avoid by random sampling |
|
|
Best way to overcome allocation error is (1) but it might not work if (2). To overcome this we should (3). |
(1) Randomisation (2) Participants are small in number (3) Recruit a large number of participants |
|
|
Exposure mixed with another factor is a (1). To (2) we must (3) which is called (4). |
(1) Confounding variable (2) Reduce the effects (3) Stratify (divide) those with CVs into sub-study (4) Stratified (adjusted) analysis |
|
|
After allocation and adjustment what should occur? What does this ensure? |
A baseline comparison should be done. Ensures “base” characteristics of EG and CG are similar. |
|
|
How does allocation error happen in RCT? What can be done to avoid this? Is it always necessary? |
Randomised allocation changed deliberately Can be avoided by concealment of allocation Not required if exposure/control cannot be distinguished (e.g. placebo drug is identical) |
|
|
If a participant's exposure changes this is called a (1) and the best way to overcome it is by (2) |
(1) Maintenance error (2) Blinding |
|
|
What can measurement error result in? |
Incorrect allocation to EGO or CGO |
|
|
How does measurement error occur? |
Deficiency in methods or instruments tomeasure outcome e.g. a poorly designed questionnaire, faulty scale, ornon-blinded personnel measuring, and/or using subjective measures. |
|
|
How to reduce measurement error? |
Objective measures and double blinding |
|
|
What are the sources of non-random error? |
1. Recruitment error (EVE) 2. Allocation error (CVs and in RCT, deliberate) 3. Maintenance (comp, contam, loss) 4. Measurement |
|
|
What is an RCT? |
Experimental longitudinal study. Pop are individuals allocated to EG/CG randomly then studied over time |
|
|
What is a meta-analysis? |
Combining the results of a number of studies (so as if one large study was done) to reduce random error |
|
|
What is the application of a cohort study? |
Studycause of disease incidence, or effect of interventions. |
|
|
What are the features of a cohort study? |
Individual participants Allocated EG/CG by measurement Observational (non-experimental) Tracked over time (longitudinal) Outcomes measured at f/up |
|
|
What are the strengths of a cohort study? |
Cheaper than RCTs Exposure measured before outcomeavoids reverse causality and recall bias (clear time sequence) |
|
|
What are the weaknesses of a cohort study? |
Confounding common Mx error common in longstudies |
|
|
What is the main application of a CSS? |
Measuringdis-ease prevalence (burden) in different populations |
|
|
What are the features of a CSS? |
Individuals participants Allocated EG/CG by measurement Observational (non-experimental) Outcomes measured at sametime |
|
|
What are the strengths of a CSS? |
Cheap and quick
Best for prevalence No mx error as no f/up |
|
|
What are weaknesses of CSS? |
Confounding common Possible reverse causality(uncertain time sequence) |
|
|
Application of RCT? |
Tostudy the effect of interventions (generally Rx or Tx) |
|
|
|
|
|
|
|
|
|
|
Main features of an RCT? |
Individual participants Experimental Allocated to EG/CG randoml Exposure and outcomesmeasured during f/up |
|
|
Strengths of an RCT? |
Randomisationminimises confounding |
|
|
Weaknesses of an RCT? |
Ethical limitations Logistically difficult Long-term f/up difficult Large studies expensive, so often small (randomallocation error!) Mx error common |
|
|
Features of an ecological study? |
Participants are groups of individuals Longitudinal or cross sectional Experimental or observational Randomised or non-randomised |
|
|
Strengths of an ecological study? |
Cheap & quick Useful when some populations exposed but othersare not Efficient for rare outcomes |
|
|
Weaknesses of an ecological study? |
Confoundingvery common |
|
|
Application of an ecological study? |
Investigate associationsbetween exposures & disease prevalence orincidence in differentgroups of populations |
