midterm Flashcards
Descriptive epidemiology
Classify according to person, place and time variables
analytical epidemiology
explores causal hypotheses between exposures and health conditions
Simple random sampling (SRS)
sample is selected using random processes
stratified random sampling
to ensure adequate representation of minority groups that might be underrepresented in SRS
convenience sampling
selecting individuals from easily accessible and available groups such as those who visit a specific clinic
Systematic sampling
choosing via a systematic, predetermined approach from the sampling frame
Cluster sampling
selecting groups of individuals
Central tendency
in normal distribution, mean, median and mode are equal and situated at center of distribution with values symmetrically distributed around central point
skewed distribution
mean, median and mode are often different- median is most accurate because it is resistant to outliers
symmetric distribution
mean and median are equivalent- mean is preferred bc it is more stable/precise across different sample sizes
Multimodal distribution
several peaks in frequency of data marking distinct groups/clusters within dataset (ex. Age-related changes)
Epidemic curve
represents distribution of cases over specified time frame to identity patterns in disease outbreaks
Dose-response curve
correlation between exposure and effect
Incidence
number of new cases occurring in specific population in specific time period
Prevalence
total number of existing cases in population at specific time, expressed as proportion, New cases and treatment advances that extend survival can boost prevalence
person variables
Age, sex, race, socioeconomic status, marital status
place variables
Morbidity and mortality influenced by geographic factors (ie life expectancy), urban vs rural, localized patterns of disease
time variables
Secular trends (long-term changes in frequency of diseases), cyclic (seasonal) trends, point epidemics, clustering
Criteria for causal relationship
Strength of association
Consistency
Specificity
Temporality- establishing that exposure occurs before health outcome
Biological gradient- depicting how increased exposure correlates with increased disease incidence
Plausibility
Coherence
Experimental evidence
Analogy in causality- comparison against established causal relationships
Risk factor
Factors correlated with an increased likelihood of disease but are not sufficient on their own to cause it
observational studies
ecological, case-control, cohort, cross-sectional
experimental studies
clinical trials, field trials, community trials
Advantages: less prone to biases, control for confounding variables, causal relationships
Disadvantages: complex and costly, large sample size, ethical considerations, patient non-compliance
bias
Systematic error that leads to deviation from truth
Hawthorne effect
altered behavior bc know they’re being studied
Recall bias
remembering exposure history
Selection bias
selecting participants leads to misrepresentation of association between exposure and outcome
ecological studies
estimate exposure levels for groups based on data from environment, applying same exposure level to all group members despite variations within group
Advantages: broader context, conducted quickly with less resources because data is more readily available, useful when cant get individual level data
Disadvantages: ecologic fallacy- correlations at group do not necessarily apply to individuals, overlook individual/personal behaviors and genetic predispositions
case-control studies
cases have outcome/disease, retrospective approach
Advantages: useful for studying low-prevalence conditions or rare diseases, fewer resources and quicker than traditional follow up studies
Disadvantages: innacurate exposure measurements, representativeness of cases and controls may be uncertain, indirect estimates of risk, temporal relationships may not be clear
cohort studies
tracks incidence overtime following cohort that shares common characteristic
Fixed- individuals classified as exposed or unexposed at specific baseline time and followed
Dynamic- continuous recruitment of new eligible candidates over time- inds may shift from unexposed to exposed group
Advantages- evaluating incidence rates, direct observation of risk, relationship between exposure and outcomes over time, uncommon exposures or diseases with clear temporal relationships between exposures and outcomes
Disadvantages- costly and time consuming, not useful in rare conditions, planning and tracking large groups, non-participation may affect demographics
cross-sectional studies
examines defined population or random sample at specific point in time- assessing both presence of disease and potential determinants, mainly descriptive (not causal relationships)
Advantages: snapshot, inexpensive and easy bc do not follow over time
Disadvantages: establishing causal, controlling for confounding variables, accurate reporting of exposure by patients, selection bias, length bias
clinical trials- randomized controlled trials (RCTs)
phase 1 tests safety, phase 2 immune responses and phase 3 randomization
randomly allocated into test and control groups
field trials and community trials
involve healthy individuals at risk, data collection in natural settings, evaluate intervention designed to prevent diseases
involve groups, can be harder
environmental epidemiology
etiology (causes/origins), natural history, health status of pop, effectiveness of interventions, geographic focus
Natural history of disease
Pathological onset, presymptomatic stage, clinical stage
Prognosis- predicting likely course of disease
Efficacy- benefits of treatment under ideal conditions
Effectiveness- benefits in real-world clinical settings (RCTs)
dose-effect/dose-response relationship
Does-effect identify which adverse effects should be prioritized for prevention
Dose-response to determine what constitutes acceptable level for exposure
directed acyclic graphs (DAGs)
relationships/causal links (arrows) between variables (nodes), no sequence of arrows forms a closed loop, if 2 or more variables share same cause (“parent” cause should be included), unmeasured confounding/cause (arrow pointing both directions between 2 nodes)
