Basic intro Flashcards
Public Health Framework
- Define the problem
- Identify risk and protective factors
- Develop and test prevention strategies
- Assure widespread adoption
Epidemiology:
The study of measuring dis-ease in different populations
Risk of dis-ease equation:
Number of cases of dis-ease
_________________________
Number of people in the population
Health:
The absence of death, disease or disability
Acronym for GATE Frame:
PECOT
PECOT:
Participants Exposure group (EG) One half of pop Comparison group (CG) Other section of pop Outcomes Time
Cohort study:
Also called follow up study, when a group is followed over time and relevant dis-ease events are counted. Description of study usually says “observed”. Results in a death rate e.g. deaths out of 1000 people per year
EGO and CGO:
Exposure group outcome
Comparison group outcome
EGO=
A/Exposure group (deaths in EG/population of EG)
CGO=
B/Comparison group (deaths in CG/population of CG)
Arrow to the right =
Incidence
EGO and CGO are called an incidence measure of occurrence when:
Dis-ease events are counted as they occur over a period of time. e.g. km per hour - must include time in the equation
EGO and CGO are called a prevalence measure of occurrence when:
People with dis-ease are counted at one point in time. eg: if a car had travelled 200,000km when we checked today
If you want to know how fast a car can go you measure:
Incidence (km per hour)
If you want to measure how worn out a car is you measure:
Prevalence (mileage)
Arrow straight down =
Prevalence
Which measure of occurrence gives more data and why?
Incidence because it is only determined by the dis-ease risk in a population and because it measures events, population and time
If it difficult to measure when a dis-ease event occurs or it occurs frequently, it is measured IF (yes/no) it has occurred over a period of time, this is a measure of:
Period Prevalence
Obesity, asthma attacks and diabetes are measured using _____
Period Prevalence
Cross sectional study:
Measuring everything at the same time
EG measuring alcohol consumption and drug consumption at the same time and studying the relationship between the 2
Relative Risk=
EGO/CGO
Relative risk doesn’t have:
Units
In what studies is a cohort study often inaccurate:
Diet/disease/drug alcohol use
How is death always measured and why:
Death is always a measure of incidence as it happens over time
Ecological study:
Comparing averages of whole countries rather than individuals
(e.g. NZ vs AUS rather than NZ men vs NZ women aged 30-50)
RCT stands for:
Randomised controlled trial
What is a randomised controlled trial?
Similar to a cohort study except participants are randomly allocated to EG and CG
Can you measure prevalence within a cohort study?
Yes, you can stop EG after a year of following EG and CG and taking a measurement
What are the benefits of RCT?
You have an equal chance of being in EG and CG so the outcome can only be due to the outcome of the exposure
Double blind:
Neither the participants nor the investigators know which intervention was given to which participant
Singe blind:
Only the participants are unaware of which intervention was given
In what situation is RCT the best type of study?
When answering questions about the effect of a treatment if it is both ethical and practical
Risk Difference:
Risk of one group - the risk of the other group (EGO-CGO or CGO-EGO)
If the risk of 1 group is 30/100 and the risk of the treatment group is 20/100 what is the effect of treatment?
10 per 100, reduction in the amount of disease, or the benefit of giving 100 people the treatment is 10 more people not dying
If EGO=CGO what is the risk difference?
0 - those given the treatment had the same outcome of those without the treatment, there is no difference in the risk of death whether treated or not treated
Risk ratio: (relative risk)
EGO/CGO or CGO/EGO
If the relative risk is 2/3, what is the risk reduced by?
1/3
For example if RR of disease without treatment is 0.66, the risk of t disease is 33% lower in that have been treated
If EGO=CGO what is the risk ratio (or relative risk)?
1
If you are told you can reduce the risk by 50% what do you need to ask yourself?
Are they subtracting one risk from the other or dividing them
How do you all the difference between relative risk and risk difference?
Risk difference has units (deaths/population/time) while relative risk has no units (due to cancelling out)
Other name for risk difference?
absolute risk difference
Other name for relative risk?
RELATIVE risk difference
If the risk of dis-ease is 0 what is the benefit of treatment?
There is no benefit because 0/30 is still 0
When an error occurs by chance what is it called?
Random error
What is it called if an error occurs because of poor study design, poor study processes?
Non random error (bias)
What causes a study to be considered a valid study?
A small amount of random or non random errors
If people are allocated by measurement to EC and CG what study type is it?
Cross sectional
When the exposure measured in a group (e.g. high solvent use) is also associated with another factor (e.g. high alcohol consumption) what is this called and what is the result?
Confounding, which causes a higher death rate and therefore inaccurate results
What is the solution to the mixing of factors within an EG or CG (confounding)?
Adjustment through the dividing of the study into sub studies = sub EG and CG to compare the relationship between the test subject and external factors
_____ is the best way to reduce confounding?
Randomised Allocation (RCT) as EG and CG will be very similar
Why does a cross sectional study require no maintenance and therefore is not effected by confounding?
Exposure and outcome are measured at the same time=no follow up to be maintained
Is asking the severity of an asthma attack an objective measurement?
Objective as the severity depends on what the person considers as severe
Is asking if someone has has an asthma attack an objective measurement?
No - you either have an asthma attack or not, it is not considering severity
What should you expect when you repeat a study in which you got extreme results? And what is this concept called?
Less extreme results - regression to the mean
Differences of results from similar studies are called ____ _____
Random differences ( because they happen randomly or by chance)
Only studying a sample of the population results in:
Random errors that cannot be avoided as you cannot study the whole population
Random error is measured though the:
95% confidence interval
The smaller the sample, the _____ the chance of random sampling errors
Greater
If you did 100 identical studies using samples from the same population, how many of the 95% confidence intervals will include the true value for the population that the studies were sampled from?
95/100 of the 95% confidence intervals
The range of results likely to include the true result in a whole population is determined through the:
95% confidence interval
Wider confidence intervals = _____ uncertainty
Greater
If the confidence intervals of EGO and CGO _____ they are considered statistically different from each other
Don’t overlap
If the risk difference (EGO-CGO) is statistically different from 0 OR
The relative risk difference (EGO/CGO) is statistically different from 1….
EGO will be statistically different from CGO
If the 95% confidence intervals for EGO and CGO do overlap…
There is an apparent difference that could be due to random error in the measurements
Random sampling errors can be reduced by:
Doing a bigger study
Small confidence interval/larger study = ____ uncertainty
Less
Large confidence interval/smaller study = ____ uncertainty
More
Combining studies = ____ relative risk/confidence interval
Less relative risk/smaller confidence interval
You would be unable to determine if EGO is different from CGO if:
There is a large overlap of the 95% confidence intervals of EGO and CGO
The risk difference and relative risk difference will cross the ‘no effect’ line if:
There is a large overlap of the 95% confidence intervals of EGO and CGO
‘There is too much random error to determine if there is a real difference between EGO and CGO’ when:
The risk difference and relative risk difference cross the ‘no effect’ line
The results of study will be statistically different when:
The confidence intervals for risk difference and relative risk difference do not cross the ‘no effect’ line
The upper 95% confidence interval touches the ‘no effect’ line, meaning that the result of the study is:
Borderline statistically significant
If a clinician would make the same decision whether the true result was at one end of the confidence interval or the other the results are both statistically and ______ significant
Clinically
A study in which participants are allocated to EG and CG by the investigators is an ______ study
Experimental
Most experimental studies are:
Randomised controlled trials
When participants are allocated to EG and CG by measurement it is an ______ study
Observational
Participants can be allocated to EG and CG by measurement or the investigators in a ______ study
Cohort
In a ______ study, participants can only be allocated into EG and CG and by measurement, outcomes can be measured ay the same time and prevalence can be measured.
Cross sectional
Incidence can not be measure in a ______ study because there is no opportunity to measure events over a period of time.
Cross sectional
The main measure of disease in a cohort study is _____
Incidence
The most common cause of bias in _____ are poor allocation processes and poor maintenance
Randomised Controlled Trials
All _____ studies are observational
Cross sectional
The main goal of a survey is the measure the ______ of something within a defined population
Prevalence
In a study of the prevalence of asthma among smoking and non smoking UoA students at the time of a POPLHLTH lecture, the outcome would be ______
Asthma
The prevalence or incidence of dis-ease in the same population over a number of years is an _______ study
Ecological
______ studies are very prone to confounding because there are usually many other difference between populations that may influence the study outcome
Ecological
A study that involves searching for and then recruiting a group of studies addressing the same question is a _____
Systematic review
If recruited studies have a low level of bias and have reasonably similar findings they can be combined using _____
Meta-Analysis
Excess risk is measured using _____
Risk difference (EGO-CGO)
The excess risk of dis-ease is ‘x’ over group over period, is the layout for the measuring of ______
Risk difference (EGO-CGO)
_____ as a measure of risk is expressed as a fraction of the group over time, not a percentage
Risk difference (EGO-CGO)
_____ is always measured as a percentage
Relative risk ratio (EGO/CGO)
______ explains the strength of association between dis-ease and outcome
Relative risk ratio (EGO/CGO)
40/48 = 83% is a measure of _____ because there is no time period
Prevalence
Measurement of migraines is self reported and therefore not objective (true/false) because you cannot be randomly allocated into groups when measuring dis-ease
True
Vertical arrow =
Incidence
The proportion of people that have a condition at some time within a given period, and includes those who have the condition when it starts. People are not followed for a year but are asked about the last year. EG severe (2-3) asthma attacks in a 1 year period
Period prevalence
Measuring what happens in an instant, measuring who has dis-ease at a specific point in time
What happens in an instant (photograph)
Point prevalence
What type of recruitment error is it when the participants recruited are not representative of the eligible population that refuse to take part?
External validity error
What is a key requirement of an epidemiological study?
The dis-ease outcomes (Numerator) must come from a defined population (Denominator)
When participants are divided into separate strata and are analysed as separate studies to reduce confounding is called _______
Stratified analysis
