epidemiology and social determinants of health Flashcards
how is incidence reported
as a rate (denominator includes time)
are cohort studies prospective or retrospective?
prospective - we recruit people according to risk factor status (exposed or not-exposed) outcome happens after recruitment
what is the income where once you go past - life expectancy doesnt change much
$10000 per annum
what kind of studies can you get incidence from?
only from longitudinal studies
what are the 4 stages of health transition
age of infection and famine age of receding pandemics age of degenerative and man-made diseases age of delayed degenerative diseases
which study designs are observational?
case series, ecological, cross-sectional, case-control, cohort
what did the whitehall study find
that the lower in the work hierachy you are the greater the risk of death
what is the bonus of using 95% confidence intervals over p vales?
CI also provides an indication of the precision of the result
is case control studies prospective or retrospective
retrospective - we come in after disease has already occurred
how do you minimize confounding in the analysis of a study
restriction stratification - analysis of sub group multivariate analyses
what is intention to treat analysis
analyse results assuming that subjects remained in randomized group, regardless of cross over
what is number needed to harm
the number of people needed to undergo the intervention in order to harm one person (when interventions increase the risk/rate of outcome)
selection bias
systematic difference in characteristics of people selected for the study and those not selected
what are the 10 social determinants of health
AEFSSSSTUV the social gradient stress early start social exclusion work unemployment social support addiction food transport
a ROC indicates how well a test discriminates….
people with disease from people without disease
proximate risk factors for TB
- exposure - decreased IS - malnutrition - indoor air pollution - alcohol abuse - other disease - depression and stress
pros and cons of cross sectional studies
pros: relatively cheap and easy cons: need for representative sample, no temporal relationships (association only), weak evidence of causality
how does the social gradient determine health
life expectancy is shorter and disease is more common down the social ladder in each society
what are PPV and NPV dependent on
- sensitivity and specificity - underlying prevalence of the disease (lower prevalence = lower PPV)
hazard
special type of rate that is continuously updated as a longitudinal study progresses
what is confounding
the relationship between exposure and outcome may be influenced by exposure to a confounding factor related to the exposure
how do you try to reduce confounding in clinical trials
try and get treatment groups identical in all aspects other than the intervention
which study designs are interventional?
clinical trials
Risk
probability of disease occurring in a disease free population during a specified time period
how are controls matched for cases in case control studiest
by confounders - eg. age, sex
what is a retrospective cohort study
researchers come in at a time when the cohort has already been assembled and use the data available for examination
how is risk calculated
risk = new cases in a defined period / population at risk
what is internal validity dependent on?
appropriate study design data collection data analyses - need to have reduced confounding, information bias and selection bias
what is health
health is a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity
What are the 2 broad types of research questions?
descriptive - describing a condition or the risk factors for a condition analytical - case and effect
what is the rationale for randomisation of participants in a trial?
to try and reduce confounding and reduce selection bias
cohort studies
comparison of outcomes between subgroups - longitudinal with follow up of subjects
rate
probability of disease occurring in a disease free population during the SUM OF INDIVIDUAL FOLLOW UP PERIODS
key output of case control studies
odds ratio - (approximation of relative risk)
bias leads to what
systematic difference between groups and therefore under/over estimation of true results
how does stress determine health
stressful circumstances –> worry and anxiety –> damaging to health and may lead to premature death
how does having a job determine health
people who have more control over their work have better health
p value measures…
the probability that the observed result arose from change
definition of sensitivity and formula
% of people with disease that test positive TP/TP+FN
outputs of cross sectional studies are mostly what?
descriptive - especially prevalence
which study designs are longitudinal?
cohort, clinical trials
When looking at the clinical setting of a trial, what do you have to look for?
PICOT Population Intervention Comparator/control Outcome Timing
how does food determine health
a good diet and adequate food supply are central for promoting health and well-being
what are clinical trials
longitudinal studies designed to assess if an intervention changes the incidence of an outcome
what does a hazard ratio tell you
at any given point in time within the period of follow-up, the probability of outcome in intervention group is….. compared to that of the control group
relative risk
indicates the relative magnitude of change in risk/rate of outcome, associated with exposure
definition and forumla for positive predictive value
% of positive tests that is truly positive TP/TP+FP
how is prevalence reported
as a proportion/percentage
how does unemployment determine health
job security increases health, well-being and job satisfaction. Higher rates of unemployment cause more illness and premature death
definition and formula for negative predictive value
% of negative tests that is truly negative TN/TN+FN
what is the advantage of randomisation
we can even out confounders that we are not even aware of - just be change
screening limitation
- selection - healthy more likely to be screened - lead-time - early detection, not necessarily prolonged survival - length-time - detection of non-aggressive disease
which study designs are analytical?
case-control, cohort, clinical trials
main outputs of cohort studies
incidence relative risk
which study designs are non-longitudinal?
case series, ecological, cross-sectional, case-control
what factors push towards higher life expectancy and higher income
new technology emerging economy aid trade peace industrialisation
what is the focus for improving global health?
improving health and achieving equity in health in all people
population attributable risk
indicates the additional or excess risk/rate of the outcome in the population, due to the exposure
What is the defining factor between longitudinal and non-longitudinal studies?
longitudinal studies involve follow up of study subjects (information progressively collected over multiple encounters) while non-longitudinal does not (information collected in only one encounter).
case control studies
compares previous exposure to a risk factor to cases and conrol
what is a receiver operator characteristic curve
a plot showing 1-specificity vs sensitivity for various thresholds for a test - graphical representation of trade-off between sensitivity and specificity in tests
instantaneous rate
rate applies to an exact point in time
attributable risk percent
proportion of incident disease among exposed people that is due to exposure
what is the rationale for blinding?
reduce information bias so the prejudice about the intervention does not influence the outcome
how is rate calculated
rate = new cases in a defined period total person time / total person time of follow up
effects of chronic stress
impaired memory, increased risk of depression, deteriorated immune response, hypertension, high hormone levels, higher risk of infertility and miscarriage
how are outcomes ascertained?
according to strict, standardised, objective criteria. Determined by centralized committees (blinded as well)
what is blinding
non-awareness of intervention allocation (single-blind or double-blind)
absolute risk/rate
isolated measurement of risk/rate (no indication of association with exposure)
what is the number needed to treat
the number of people needed to undergo the intervention in order to prevent outcome in one person
what are the risk factors for CVD?
smoking, hypertension, diabetes, male, older, genetics, obesity, hypercholesterolaemia, inactivity, chronic stress, alcohol
what was the main reason for the decline in TB incidence over the 20th century
due to the economic growth, social reform, poverty reduction, improved living conditions and advances in medicine and public health
what is a meta-analysis
an analysis of combined data from multiple studies
purpose of a diagnostic versus a screening test
diagnostic test - provides confirmation of a disease (results definitive) screening test - provides identification of patients who may have disease (results are preliminary)
how is population attributable risk calculated?
PAR = risk/rate in whole population (exposed and unexposed) - risk/rate among unexposed
which study designs are descriptive?
case series, ecological, cross sectional
what are the effects of acute stress?
increased alertness, less perception of pain, immune system readied, HR increases and vasoconstriction to bring more oxygen to the muscles, mobilize energy supplies, reproductive functions temporarily suppressed
cons of cohort studies
difficult to study rare outcomes not cheap or easy
how does addition determine health
individuals turn to alcohol, drugs and tobacco and suffer from their use
how do you minimize confounding in the design and execution of a study
matching by confounder (match age and sex between groups) restriction (only recruit certain sex or age)
pros of case control studies
can gain temporal relationship between exposure and outcomes useful for studying rare outcomes
attributable risk
indicates the absolute magnitude of change in risk/rate of outcome, associated with exposure
what is survival analysis
measure the time to event
key statistical information gained from meta-anaysis
- outcome - weighted average effect size - weighting of individual studies - heterogeneity
population attributable risk percent
proportion of incident disease among whole population that is due to exposure
what is the advantage and disadvantage of intention to treat analysis
advantage - reduces selection bias disadvantage - always under-estimates any treatment effect
best kind of trial is what
a systematic review of randomized controlled trials
Cross sectional studies
sample of population. Information obtained one one point/period in time (participant only contributes data once)
4 purposes for a meta-analysis
- increase power - resolve uncertainty - improve estimates of effect size - answers other questions
what was major the reason for the results found in the whitehall study
stress! lower down the hierachy - more chronic stress - less control in the workplace
definition of specificity and formula
% of people without disease that test negative TN/TN+FP
how is attributable risk calculated?
AR = risk/rate among exposed - risk/rate among unexposed
how is number needed to treat calculated
1 / (absolute risk/rate reduction)
what are the bradford hill criteria for causality?
temporal relationship strength dose-response relationship consistency plausibility exclude alternatives experimental evidence specificity coherence
how is survival analysis shown
on a Kaplan-Meier curve
what is internal validity
the extent to which the results of a study are valid for the sample of patients being studied
What does the WHO think the major determinant of health is
the factors in the social environment - determine access to health services and influence lifestyle choices
universal confounders
age and sex
prevalence
the number of EXISTING cases of an outcome of interest in a defined population, AT ONE POINT IN TIME
what was the main reason for the increased TB incidence in the 17th-18th centuries
- rapid industrialisation and urbanisation - urban poor (increased population density, crowded living conditions, poor nutritional status)
key outcomes of clinical trials
relative risks hazard ratios absolute risk/rate reduction number needed to treat survival analysis
2 main types of bias
selection bias information (measurement bias)
what is the 95% confidence interval
the interval, within which, there is 95% confidence that the true” value lies”
how does transport determine health
healthy transport means less driving and more walking and cycling backed by better public transport
when is a 95% confidence interval statistically significant?
when the null hypothesis is not included in the interval
what factors push towards poor life-expectancy and poor income
epidemic disease global and intra-country inequality poverty colonization world and civil war
incidence
number of NEW CASES of an outcome of interest arising from a defined population DURING A TIME INTERVAL
What is external validity?
the extent to which the results of this trial are applicable to the population that I want to apply these results to
pros of cohort study
explicit knowledge on temporal relationships between exposure and outcome can include multiple exposures and outcomes research hypotheses can be addressed post hoc
associations make inferences about what
cause and effect correlation
definition of likelihood ratios
likelihood that a given test result would be expected in a patient with the disease compared to a patient without the disease
information bias
systematic difference in the way information is collected between/among groups being compared
how is relative risk calculated?
RR = Risk/rate among exposed / risk/rate among unexposed
In cross sectional studies, how is data collected?
mainly questionnaires