HADPOP Guide Flashcards
What is prevalence?
What is it affected by?
What is point prevalence?
Number of current cases at a given moment.
Increased by prolonging lives of sufferers.
Decreased by cures and deaths.
No of sufferers/No at risk
What is the crude death rate?
What is the age-specific death rate?
No of deaths per 1000 population
No of deaths per 1000 in an age group
What are the four disadvantages of case controlled studies?
- Only estimates relative risk
- No absolute incidence rates
- Often prone to recall and selection bias
- Can’t be sure exposure preceeds outcome
What is the incidence?
How do you calculate the incidence rate?
No. of new cases in a given time
Rate=number of new cases/person-years
Person-years=Number of people observed x Number of years observed
What three things can apparent associations results from?
-Chance
-Bias:
Selction, Information
-Confounders
What are Confidence Intervals?
How do you calculate them?
Values you can be 95% sure the true value lies between.
- Upper = Observed x error factor
- Lower = Observed/error factor
What are the 6 uses of mortality data?
- Identifies cause of death to analyse patterns
- Identifies health problems to inform services
- Uses as incidence date for rapid diseases
- Errors from omissions
- Incorrect diagnoses
- ‘Fashions’ in diagnosis
State and define the fertility determinants
What is fertility increased and decreased by?
Fecundity - Physical ability to reproduces
Fertility - Realisation of fecundity
Conceptions - Liver births + miscarriages + abortions
Increased by sexual activity + economic climate
Decreased by contraception and abortions
What is publication bias?
Studies with significant/favourable results are more likely to be published
What are the disadvantages of using historical controls? (3)
- May have been treated differently
- Less information about controls makes confounder adjusment difficult
- Comparison with patients can overestimate the benefits of a new treatment
What are population estimates?
Applying what’s known abouts births, deaths and migrations to the present population
What are the characteristics of the ideal outcome? (12, but divide into 6 groups of two)
Appropriate: to patient, clinician, society
Valid: reasonably linked to treatments compared
Specific
Reliable: different people in different settings with same results
Simple and sustainable: repeatable
Cheap and timely: not expensive or long
What are the four advantages of case-controlled?
- Study rare diseases
- Study range of exposures
- Cheap
- Quick
What is the paradox of the commons, you sexy beast?
The principal that the optimum stratagey from an individual is not the optimum strategy for the community?
Why are there losses to follow up in studies?
How can they be minimised? (four)
Clinical condition may necessitate removal or may choose to leave
- Make follow up practical and convenient
- Make clear commitment involved at start of study
- Avoid coercion
- Maintain contact
What are the three types of data?
Binary exposure
Several categories
Continuous
What is a case controlled study?
Classify on the basis of disease and the compare exposure status
What three factors contribute to population size?
Births, deaths and migration
What are the three types of blinding?
What are its benefits?
What are its difficulties?
Single, double and triple
Avoids bias
Surgery, lifestyle interventions
What are observed quantities?
What is random variation?
What is affected by random variation to produce observed quantities?
Values that depart from their true value via random variation e.g. incidence, prevalence
Fluctuations in disease patterns that can’t be explained by systematic causes
True underlying tendencies
What is systematic variation?
How can it be used?
What is the problem with it?
Risk varies systematically throughout population
Can give clues as to cause of disesae
Can confound relationships between 2 variables
What are the seven disadvantages to cohort studies?
Really Tired Elephants Share Ears with Randy Creeps
Resource intensive Time consuming Expensive Risk of high number of loss --> Survivor bias Ethical dilemmas Can't study rare diseases Difficulty with confounding variable
What is the problem with incidence and prevalence?
They assume that all members of a population run the same rish, which they don’t
Why use random allocation when designing studies?
Minimise selection biase and confounders
What is a confounder?
How do you minimise them?
Affects both exposure and outcome, but does not lie on causal pathway
Matching using important confounders e.g. Age, Sex and ethnicity
What is the primary outcome?
What is the secondary outcome?
Sample size
Side effects
What is selection bias?
How do you minimise it?
Specifically selecting cases and controls so more likely to achieve desired result
Cases representative of cases
Controls representative of population
What are the ethical issues behind placebos? (3)
- Deception
- Should only be used when no standard treatment available for comparison
- Patients should be fully informed that they may receive a placebo
What is the placebo effect?
Compariosn of treatment to non-treatment
-Measure of the attitude change to the illness itself by the pure thought it being treated
What are population projections?
Applying future estimates of births, deaths and migrations to future population estimations.
What is causality?
Epidemiological addumption that no disease occurs at random and has preventable and causal factors
Who writes the death certificate?
What is in the death certificate?
When must it be written?
Attending doctor
Cause of death
Within five days of death
What are the 6 advantages of cohort studies?
Rigorously defined disease + exposure
Study ran of outcomes
Study rare exposures
Establish exposure preceeds outcome
Good for conditions that fluctuate with age
Prospective - detailed assement of exposure outcomes, personal characteristics and confounders.
What is a hypothesis?
What is the null hypothesis?
A statement that an underlying tendency of scientific interest takes a particular value
States that there is no statistically significant different between 2 populations
What is the meaning of :
- Necessary
- Sufficient
Necessary - Exposure must always preceed disease
Sufficient - Exposures capable of causing disease on its own
What is the total period fertility rate?
How is it calculated?
Average no of babies born to hypothetical worman between 15-44
Sum of age specific fertility rates
What are the five advantages of clinical trials?
- Provide reliable evidence of treatment efficacu and safety
- Fair
- Controlled
- Reproducible
- Intervention studies
What is the standardised mortality ratio?
How do you calculate it?
Compares the observed deaths in a population with the expected deaths of that age group, assuming the age-sex distributions are identical
No of observed deaths/No of expected deaths
What are clinical trials?
Planned experiments involving patients which elucidates the most appropriate method of treatment for future patients with a given medical condition
What is association?
Statistical dependance between 2 or more events, characteristics or other variables
How many controls in case controlled study and why?
Increase in cases means increase in controls
Required to reduce error factor
Beyond 5 controls to cases, the decrease in error factor is disproportionate to the increase in costs of controls?
What is internal comparison?
What observed value is calculated?
Between 2 groups within the cohort
Incidence rate ratio
What is the point of clinical appraisal?
Distingus between good and bad evidence to provide optimal care for patiens
What is the criteria for inferring causality? (9)
Strength of association
Specificity of association
Consitency of association
Coherence of theory
Analogy
Biological plausability
Reversibility
Temporal sequence
Dose response
What is the incidence rate ratio?
How is it calculated?
What do its results indicate?
Comparison of incidence rates between 2 groups, varying in exposure
Rate A/Rate B
> 1 suggests increase risk in group A
<1 suggests increase in risk in group B
What is the crude birth rate?
What is the general fertility rate?
No of babies born per 1000 population (live births)
Number of live births per 1000 women aged 15-44
Why do we need a population perspective of medicine?
Look at large groups of people to discover causes of disease and evaluate preventative/curative measures
What is external comparison?
What is the observed value calculated?
What are its limitations? (2)
Compare incidence within cohort with reference population
SMR
- Limited data for reference population
- Reference and study pops may not be comparable due to selection bias and healthy worker bias
How do you calculate odds ratio?
—————-Cases—-Controls
Exposed—–A————-B—-
Unexposed–C————D
AxD/BxC = Odds of being exposed (cases)/odds of being exposed (controls)
What are the problems with hospital acquired infections? (3)
How can HAIs be prevented? (3)
Increase costs
Increase hospital stay
Cause deaths
Hand washing
Restrict anti-biotic useage
Cohort colonised and non colonised staff and patients together
How do you interpret a 95% C.I?
Does the value of the null hypothesis lie between the CIs?
- Yes
- Pt accept hypothesis
Why is there variable study quality? (4)
How do you deal with this? (3)
Poor study design
Design protocol
Protocol implementation
Prone to bias + confounders
Define basic quality standard
Score each study
Incorporate into weighting
What is a cohort study?
What are the types?
Classify patients based on exposure and then follow up to look for outcome
- Prospective - Exposure status and follow up
- Retrospective - Exposure status and follow up based on historical records
What are the 5 contentious issues associated with health care information use?
Completeness/duplication Accuracy Confidentiality Numerator/Denominator mismatch Varying diagnosis of disease
What is age sex standardisation?
Why use it?
A procedure that adjusts for age-sex differences in population structure, to provide a single summary measure
Age and sex are both factors that can confound disease risk
What is health information used for? (3)
Identify health and healthcare needs
Monitor trends in disease
Monitor peformance in healthcare
What is a cencus?
What information about the population does it produce?
Simultaneous recording of data by the government at a particular time, pertaining to all persons living in a particular territory
Population size and structure
(S9 +)
What are meta analyses?
What observed value is calculated?
Quantitive synthesis of the results of 2 or more primary studies, that addressed the same hypothsis in the smae way
Pooled estimate odds ratio
(S9 +)
What are forest plots?
Squares = Individual ORs
= Size x Weight
Lines = 95% CIs
Diamond = Pooled estimate
Centre = Pooled OR
Width = Pooled CI
Solid line = Null hypothesis
RCTs with OR>1 = greater survival odds(S9 +)
Why are systematic reviews so credible? (5)
Unbiased Objective Explicit Transparent Reproducible
(S9 +)
What is intention to treat analysis?
Everyone is included
More representative of population
(S9 +)
What are systematic reviews?
Overview of primary studies, which the synthesis can include meta-analyses
Why are outcomes predefined? (S9 +)
Prevents data dredging Prevents repeat analyses Sets data collection protocol Provides agreed criteria for measurement Assesment of outcomes
What is as-treated analysis? (S9 +)
What is its problem?
Non-compliers excluded
- 2 groups no longer random
- Lost immunity to confounders
What 2 methods are used for calculating pooled estimate OR? (S9 +)
Fixed-effect model:
Studies estimate the same effect size
Random effect model:
Studies estimate similar effect size
What is a nested case controlled study? (S9 +)
Collection of data from an evolving exposure and outcome database of a propective cohort study
What is a funnel plot? (S9 +)
Measure study size against measure of effect
Balanced funnel means no bias
What are the steps in conduction of Randomised Control Trial? (S9 +)
Identify Recruit Consent Maintain Allocate Assess
What are the advantages of nested case-controlled over cohort and case-controlled studies? (S9 +)
Cohort:
Collect more detailed inmfromation for a minority of participants
Case-controlled:
- Incidence rates can be defined
- Population for sampling of controls is already defined
Define: (S9 +) Collective ethic - 1 Individual ethic - 2 Clinical equipoise -3 Valid consent -4
1- All individuals have the right to safe and efficient treatment
2-RCTs go against beneficence, malfeansance, autonomy and justice
3-Genuine ingnorance of no better treatment
4-From a knowledgeable informant with appropriate information
What is information bias?
Bias as a result of misclassification of exposure of outcomes. Can be differential or non-differential.
Information bias definition
Error due to systematic differences in the measurement or classification of subjects in the groups being studied
