Epidemiology
Epidemiology
Improve health of populations
frequencies of diseases & health states
(trends)
factors that cause
predicting occurrence & distribution
factors that prevent, prolong life, improve
health
Epidemiology
Identify / Explain causal factors
(exposures)
epidemics
epi - above/around
dem - people
Epidemiology
Distribution and determinants
disease, injury, or dysfunction
Epidemiology
Exposures risk (causal) factors
lifestyle
occupational hazards
environmental influences
interventions
Epidemiology
Descriptive
distributions / patterns
Analytic
cause and effect
make inferences
Epidemiology
Descriptive Exploratory Experimental
Describe Identify Cause and
Populations Relationships Effect
Clinical Trials
Cohort/Case-Control
Studies
Descriptive Epidemiology
Who
Where
When
Descriptive Epidemiology -
Research Designs
Case report/series
Correlational studies
Cross-sectional surveys
NO Causality
Measures of Disease
Frequency
Prevalence
# of existing cases
total population at risk
Point Prevalence
Point Prevalence
1,000 therapists in NYS during 1999
had LBP
10,000 therapists in NYS
P = 1000/10,000 = 10%
Measures of Disease
Frequency
Incidence
Cumulative Incidence
# of new cases
total population at risk
Cumulative Incidence
500 therapists in NYS developed LBP in
1999
10,000 total therapists
CI = 500/10,000 = 5%
Measures of Disease
Frequency
Incidence Rate
# of new cases
total person-time
Incident Rate
Of the 10,000 therapists in 1999 -
2,000 worked for only six months
8,000 therapists contributed 8,000
person-years
2,000 therapists contributed 1,000
person-years
IR = 500/9,000 = 5.6%
Descriptive Epidemiology
Vital Statistics
Birth rate
Mortality rate:
total mortality - all causes
crude mortality - total mortality / avg.
midyear population
Descriptive Epidemiology
Vital Statistics
Mortality rate
“cause-specific” - specific disease / avg.
midyear population (AIDS, CAD, etc.)
“case-fatality” - deaths / individuals with
disease
Age-specific rates
Analytic Epidemiology
Observational Studies
Case-Control
Cohort
Clinical Trials
Intervention Study
Descriptive Exploratory Experimental
Identify Cause and
Relationships Effect
Clinical Trials
Cohort/Case-Control
Studies
Observational Analytic
Designs
Objective:
Test hypotheses about
association/relationship of risk factors
and disease
Case-Control Studies
Case Definition
Case Selection
population-based – general population of
those w/ disorder
hospital-based – patients in medical
institution
Case-Control Studies
Analysis Issues
Selection bias
Misclassified
Observation/Interviewer bias
Extraneous variables
Cohort (follow-up) Studies
Cohort – group of individuals followed over
time
Temporal component
Limited use w/ rare disorders
Cohort Studies
Prospective
Control and monitor data collection
Subjects readily available
Retrospective
Inexpensive and faster
Incomplete/inadequate data
Cohort Studies
representative sample generalize
group identification
internal comparison
external comparison
Cohort Studies
Analysis Issues
Misclassification -
Attrition -
Clinical Trials (RCT)
Intervention Study
Causality
Rigorous - Gold standard
Prospective - intervention vs. control
Clinical Trials
Therapeutic
Effect of rx or intervention
Preventative
Agent/procedure reduce risk of
developing a disease
Clinical Trials
Subject Selection
Target/Reference
Experimental/accessible population
Clinical Trials
Validity
sample size
achievable
attrition
Clinical Trials
Analysis
randomization
blinding
bias
ethics
Clinical Trials
Analysis
tests of statistical significance
(difference)
t-tests, ANOVA, etc.
causality
inferences about the population
Measures of Association -
Observational Studies
Test Hypotheses
Relationships
Association Exposure represents a risk
factor
Measures of Association
Relative Effect
Exposed:Unexposed
Absolute Effect
Disease Rateexposed - Disease Rateunexposed
Relative Risk
Disease
Exposure
Yes No
Yes a b a+b
No c d c+d
b+d N
a+c
Relative Risk
Cumulative Incidence Estimate Exposed
(CIE)
Unexposed (CIO)
CIE a / (a + b)
RR = =
c / (c + d)
CIO
Relative Risk
Disease
Yes No
Exposure
Yes 50 33 a+b
19 259
No c+d
a+c b+d N
Relative Risk
CIE a / (a + b)
RR = =
c / (c + d)
CIO
CIE = 50/83 = 0.602
CIO = 19/278 = 0.068
RR = 0.602/0.068 = 8.9
Relative Risk
Odds Ratio – Case-control
a/c ad
OR = =
b/d bc
= (50)(259) / (33)(19)
= 20.6
Attributable Risk
Risk Difference = AR = IE - EO
a c
AR = CIE - CIO =
a+b
- c+ d
AR = 0.602 - 0.068 = 0.534
AR = 534/1,000
Attributable Proportion
AR x 100 IE -IO
AR% = = x 100
IE IE
AR% =0.534/(50/83) = 88.7%
Attributable Proportion
For case-control (Odds Ratio)
OR-1 x 100
AR% =
OR
AR% = 19.6/20.6 = 95.1%
Confounding
Extraneous (interfering) variable
associated w/ exposure
considered a risk factor - independently of
the exposure
NOT part of the causal link
Causality
Inherent to interventional research but not
observation research subject to
interpretation:
Time sequence
Strength of association
Biologic credibility
Consistency
Dose-Response
Other Research
Approaches
Historical
Evaluation
Methodological
Secondary Analysis
Historical Research
To determine:
how present conditions evolved
anticipate future events
Historical Research
Incorporates:
judgements
analyses
inferences
Establish relationships thru:
organizing
synthesizing
Historical Research
Critical Review of:
events
documents
literature
other
Sources of Historical Data
Primary
original documents
letters, videotapes, photographs, minutes
eyewitness accounts
Sources of Historical Data
Secondary
biographies
textbooks
encyclopedias
literature reviews
newspaper
summaries
Historical Research
Reliability and Validity
External Criticism
authenticity
Internal Criticism
content within context of question
Historical Research
After data is collected:
establish relationships
no cause and effect
Evaluation Research
Systematic approach to evaluating
programs
clinical
academic
Effectiveness
Evaluation Research
Establish questions/hypotheses
Choosing variables
sensitive
Methodology and design
Evaluation Research
Data Collection and analysis
Evaluations
Formative – performed as part of program
planning or during implementation
Summative – assesses outcomes after
program is implemented
Evaluation Research
Evaluation of Program Objectives
(measurable)
Quantitative
Qualitative/Behavioral
Evaluation Research
Goal-Free Evaluation
evaluating predetermined goals vs.
overall effect of program
Evaluation Research
Limitations
Bias
Complex
Long term
Usefulness
Methodological Research
Development and testing of new
instruments/measurement tools
Reliability and Validity
applications to various patient populations
sensitivity
conditions
“gold standard”
Methodological Research
Only the beginning
Secondary Analysis
Analyzing previously collected data
subsets of original data
new statistical techniques
test different hypotheses
Secondary Analysis
Advantages:
Low cost
Little wait for data
Learn from predecessors
Disadvantages:
Lack of control of data collection