epidemiology. improve health of populations zfrequencies of diseases & health states (trends)...

Post on 19-Dec-2015

215 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Epidemiology

Epidemiology

Improve health of populations

frequencies of diseases & health states (trends)

factors that causepredicting occurrence & distributionfactors that prevent, prolong life,

improve health

Epidemiology

Identify / Explain causal factors (exposures)

epidemics

epi - above/arounddem - people

Epidemiology

Distribution and determinants

disease, injury, or dysfunction

Epidemiology

Exposures risk (causal) factors

lifestyleoccupational hazardsenvironmental influencesinterventions

Epidemiology

Descriptivedistributions / patterns

Analyticcause and effectmake inferences

Epidemiology

DescriptiveDescriptive ExploratoryExploratory ExperimentalExperimental

DescribePopulations

IdentifyRelationships

Cause and Effect

Clinical TrialsClinical Trials

Cohort/Case-ControlCohort/Case-ControlStudiesStudies

Descriptive Epidemiology

Who

Where

When

Descriptive Epidemiology - Research Designs

Case report/series

Correlational studies

Cross-sectional surveys

NO Causality

Measures of Disease Frequency

Prevalence

Point Prevalence

# of existing cases

total population at risk

Point Prevalence

1,000 therapists in NYS during 1999 had LBP10,000 therapists in NYS

P = 1000/10,000 = 10%

Measures of Disease Frequency

IncidenceCumulative Incidence

# of new cases

total population at risk

Cumulative Incidence

500 therapists in NYS developed LBP in 199910,000 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 months8,000 therapists contributed 8,000 person-years2,000 therapists contributed 1,000 person-yearsIR = 500/9,000 = 5.6%

Descriptive EpidemiologyVital Statistics

Birth rateMortality rate:

total mortality - all causes

crude mortality - total mortality / avg. midyear population

Descriptive EpidemiologyVital Statistics

Mortality rate“cause-specific” - specific disease /

avg. midyear population (AIDS, CAD, etc.)

“case-fatality” - deaths / individuals with disease

Age-specific rates

Analytic EpidemiologyObservational Studies Case-Control CohortClinical Trials Intervention Study

DescriptiveDescriptive ExploratoryExploratory ExperimentalExperimental

IdentifyRelationships

Cause and Effect

Clinical TrialsClinical Trials

Cohort/Case-ControlCohort/Case-ControlStudiesStudies

Observational Analytic Designs

Objective:Test hypotheses about

association/relationship of risk factors and disease

Case-Control Studies

Case DefinitionCase Selectionpopulation-based – general

population of those w/ disorder

hospital-based – patients in medical institution

Case-Control Studies

Analysis IssuesSelection 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

ProspectiveControl and monitor data collectionSubjects readily availableRetrospectiveInexpensive and fasterIncomplete/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

TherapeuticEffect of rx or intervention

PreventativeAgent/procedure reduce risk of

developing a disease

Clinical Trials

Subject SelectionTarget/Reference

Experimental/accessible population

Clinical Trials

Validitysample size

achievable

attrition

Clinical Trials

Analysisrandomizationblindingbiasethics

Clinical Trials

Analysistests of statistical significance

(difference)t-tests, ANOVA, etc.causalityinferences about the population

Measures of Association - Observational Studies

Test Hypotheses

Relationships

Association Exposure represents a risk factor

Measures of Association

Relative EffectExposed:Unexposed

Absolute EffectDisease Rateexposed - Disease

Rateunexposed

Relative Risk

a bc d

DiseaseYes No

Exp

osur

e

Yes

No

a + b

c + d

a + c b + d N

Relative Risk

Cumulative Incidence Estimate Exposed (CIE)

Unexposed (CIO)

RR = CIE

CIO

= a / (a + b)

c / (c + d)

Relative Risk

50 3319 259

Disease

Yes No

Exp

osur

e Yes

No

a + bc + d

a + c b + d N

Relative Risk

CIE = 50/83 = 0.602

CIO = 19/278 = 0.068

RR = 0.602/0.068 = 8.9

RR = CIE

CIO

= a / (a + b)

c / (c + d)

Relative Risk

Odds Ratio – Case-control

OR = a / cb / d

=adbc

= (50)(259) / (33)(19)

= 20.6

Attributable Risk

Risk Difference = AR = IE - EO

AR = a

a + b - cc+ d

CIE - CIO =

AR = 0.602 - 0.068 = 0.534

AR = 534/1,000

Attributable Proportion

AR% = ARIE

x 100IE -IO

IE

x 100=

AR% =0.534/(50/83) = 88.7%

Attributable Proportion

AR% = OR-1OR

x 100

AR% = 19.6/20.6 = 95.1%

For case-control (Odds Ratio)

Confounding

Extraneous (interfering) variable

associated w/ exposureconsidered a risk factor -

independently of the exposureNOT part of the causal link

Causality

Inherent to interventional research but not observation research subject to interpretation:

Time sequenceStrength of associationBiologic credibilityConsistencyDose-Response

Other Research Approaches

Historical

Evaluation

Methodological

Secondary Analysis

Historical Research

To determine:

how present conditions evolved

anticipate future events

Historical Research

Incorporates:judgementsanalysesinferencesEstablish relationships thru:organizingsynthesizing

Historical Research

Critical Review of:eventsdocumentsliteratureother

Sources of Historical Data

Primaryoriginal documents

letters, videotapes, photographs, minutes

eyewitness accounts

Sources of Historical Data

Secondarybiographiestextbooksencyclopediasliterature reviewsnewspapersummaries

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

clinicalacademic

Effectiveness

Evaluation Research

Establish questions/hypotheses

Choosing variables sensitive

Methodology and design

Evaluation Research

Data Collection and analysis

EvaluationsFormative – 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

LimitationsBiasComplexLong termUsefulness

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 datanew statistical techniquestest different hypotheses

Secondary Analysis

Advantages:Low costLittle wait for dataLearn from predecessorsDisadvantages:Lack of control of data collection

top related