HaDPop Flashcards
Use of a census
Population Health Housing Employment Transport Ethnic group
Define census
The simultaneous cording of demographic data buy the government at a particular time pertaining to all the persons who live in a particular territory
Measures of fertility
Crude birth rate
General fertility rate
Total period fertility rate
Define CBR
Crude birth rate
Number of live births per 1000 population
Define GFR
General fertility rate
Number of live births per 1000 females aged 15-44 years
Define TPFR
Total period fertility rate
The average number of children that would be born to a hypothetical woman in her life
Why use CBR?
Describes impact of births (-deaths +/- migration) on size of population
Takes men into account
Why use GFR?
Compares fertility of fertile female populations
Requires all women to be known
Why use TPFR?
Removes influence of age group structure when comparing fertility of fertile females
Removes confounding of age
Only observed in short term study, useless in long
Define fecundity
Physical ability to reproduce
Define fertility
Realisation of this potential as births
What affects fecundity?
Hysterectomy
Sterilisation
What affects fertility?
Contraception
Abortion
Good economic climate
Increased sexual activity
Measures of mortality
Crude death rate
Age specific death rate
Standardise mortality rate
Define CDR
Crude death rate
Number of deaths per 1000 population
Define ASDR
Age-specific death rate
Number of deaths per 1000 in age group
Define SMR
Standardised mortality ratio
Compared ‘observed’ number of deaths with number ‘expected’ (reference group)
Adjusted for age-sex distribution (removes confounding)
Expressed as percentage or ratio relative to 1
Reasons for collecting mortality data
Inform health services
Classify causes of death
Analyse patterns in mortality rate
Define incidence
Number of new cases of a disease
Useful for monitoring epidemics
Define prevalence
Number of people affected by disease (new and old)
Measures need for services
Measured as proportion not a rate
Does not take time into account
Incidence rate calculation
Persons x time (years)
Define incidence rate ratio
Allows us to compare incidence rate of two groups with different levels of exposure
Exposed/unexposed
Compare efficacy of treatment
Define confounding factor
Links exposure to outcome but not on causative pathway
Common confounders
Ethnicity
Age
Sex
Define confidence interval
Range within which we can be 95% certain that the true value of the underlying tendency lies
IR or SMR /x error factor
X1000 to get per 1000 person years
Define null hypothesis
Default position
No relationship between exposure and outcome
1
Define p value
Probability that the data observed is simply due to chance
<0.05 reject H0
Value lies within 95% confidence interval
Null hypothesis value of 1 lies in CI
P>0.05
Results not statistically significant and could be due to chance
Cannot reject null hypothesis
If null hypothesis value lies outside CI
Null hypothesis value of 1 does not lie in CI
P<0.05
Results are statistically significant and unlikely due to chance
Can reject null hypothesis
How to run a cohort study?
Recruit disease-free individuals
Classify into exposed and unexposed
Follow each group over time (years)
Count how many develop the disease being studied
Calculate the IR (d/p-y) for each group and then the IRR (relative risk)
Types of cohort studies
Concurrent (prospective)
Historical (retrospective)
Internal
External
Define concurrent (prospective) study
Follow up starts immediately or later
Define historical (retrospective) study
Collect follow up data from the past
Define internal study
Between two subgroups
IRR
Define external study
Follow 1 group
Compare to reference group
SMR
Prospective vs retrospective
Retro quicker but may have unknown confounding
Because collecting data from past
Problems with external cohorts
Often limited data available for reference population
Often no incidence data only mortality
Study and ref population may not be comparable (selection bias)
Problems with internal cohort
Always smaller sample size and bigger error factor than external
Problems with cohort studies
Large and resource intensive Long time High loss to follow up (survivor bias) Not good for rare disease Possible unknown confounders
Case control study
Identify group of cases (diseased)
Identify a suitable group of non-cases (controls)
Establish previous exposure to causative agent
Compare level of exposure in cases and controls(x6cases)
Odds ratio
Controls exposed x cases unexposed
Types of bias
Selection
Responder
Survivor
Information
When use IRR
Cohort
When use OR
Case controlled study
Cohort vs case control
Exposure basis Time n cost Rare disease n outcomes Range of d n o Bias n follow up Measure incidence
Exposure basis of study types
Cohort - compares outcomes based on exposure
Case-control exposure based on disease status
Cohort vs case control time n cose
Cohort large and time-consuming and so relatively expensive
Case-control quicker and relatively cheap
Cohort vs case control rare
Cohort not good for rare outcomes or diseases
Case-control not good for rare exposures
Cohort vs case control range disease
Cohort - study range of outcomes or diseases for each exposure
Case-control - study range of exposures for each outcome/disease
Cohort vs case control bias
Cohort prone to losses to follow up
Case-control prone to selection and information bias
Types of information bias
Recall
Publication
Types of selection bias
Allocation
Healthy worker effect