2. How we view the worldâĻ..How we view the worldâĻ..
īŽ Pessimist:Pessimist: The glass isThe glass is
half empty.half empty.
īŽ OptimistOptimist: The glass is: The glass is
half full.half full.
īŽ EpidemiologistEpidemiologist: As: As
compared to what?compared to what?
6. Epidemiology is...
īŽ "The science of long division....
I'=[(480)(log2)(10E6)]/[(9.1)(0.955po)
+0.45n]"
īŽ Statistician
7. Definition of Epidemiology*
"The STUDY of the DISTRIBUTION and
DETERMINANTS of HEALTH-
RELATED STATES in specified
POPULATIONS, and the application of this
study to CONTROL of health problems."
*Last, J.M. 1988. A Dictionary of Epidemiology, 2nd ed.
10. Definition of Epidemiology
īŽ A quantitative basic science, built on a working
knowledge of probability, statistics and sound
research methods.
īŽ A method of causal reasoning, based on
developing and testing biologically plausible
hypothesis pertaining to occurrence and
prevention of morbidity and mortality.
īŽ A tool for public health action to promote and
protect the public's health based on science, causal
reasoning, and a dose of practical common sense.
11. Epidemiology is a QuantitativeEpidemiology is a Quantitative
DisciplineDiscipline
īŽ Measures of frequencyMeasures of frequency
īŽ Counts and ratesCounts and rates
īŽ Measures of associationMeasures of association
īŽ Relative riskRelative risk
īŽ Odds ratioOdds ratio
īŽ Statistical inferenceStatistical inference
īŽ P-valueP-value
īŽ Confidence limitsConfidence limits
13. EpidemiologyEpidemiology
īŽ DescribesDescribes
ī health eventshealth events
ī cause and risk factors of diseasecause and risk factors of disease
ī clinical pattern of diseaseclinical pattern of disease
īļ Identify syndromesIdentify syndromes
īŧ Identify control and/or preventive measuresIdentify control and/or preventive measures
14. So, EpidemiologySo, Epidemiology
īŽ Is theIs the basic sciencebasic science of public healthof public health
īŽ Provides insight regarding theProvides insight regarding the naturenature,, causescauses,,
andand extentextent of health and diseaseof health and disease
īŽ Provides information needed toProvides information needed to planplan andand targettarget
resourcesresources appropriatelyappropriately
15. Kinds of EpidemiologyKinds of Epidemiology
īŽ DescriptiveDescriptive
īŽ AnalyticAnalytic
īŽ ExperimentalExperimental
Further studies to determine the
validity of a hypothesis concerning
the occurrence of disease.
Deliberate manipulation of the
cause is predictably followed
by an alteration in the effect
not due to chance
Study of the occurrence and
distribution of disease
16. Overview of epidemiologic designOverview of epidemiologic design
strategiesstrategies
īŽ DescriptiveDescriptive
īļ Populations{Correlational studies}Populations{Correlational studies}
īļ IndividualIndividual
ī Case reportCase report
ī Case seriesCase series
ī Cross sectional studiesCross sectional studies
īŽ Analytic studiesAnalytic studies
īļ ObservationalObservational
ī Case controlCase control
ī CohortCohort
ī§ RetrospectiveRetrospective
ī§ ProspectiveProspective
īļ Interventional/ExperimentalInterventional/Experimental
ī Randomized controlled trialRandomized controlled trial
ī Field trialField trial
ī Clinical trialClinical trial
17. Descriptive vs. Analytic EpidemiologyDescriptive vs. Analytic Epidemiology
DescriptiveDescriptive
īŽ Used when little isUsed when little is
known about theknown about the
diseasedisease
īŽ Rely on preexistingRely on preexisting
datadata
īŽ Who, where, whenWho, where, when
īŽ Illustrates potentialIllustrates potential
associationsassociations
AnalyticAnalytic
īŽ Used when insight aboutUsed when insight about
various aspects of disease isvarious aspects of disease is
availableavailable
īŽ Rely on development of newRely on development of new
datadata
īŽ WhyWhy
īŽ Evaluates the causality ofEvaluates the causality of
associationsassociations
Both are
18. Descriptive StudiesDescriptive Studies
īŽ Relatively inexpensive and less time-consumingRelatively inexpensive and less time-consuming
than analytic studies, they describe,than analytic studies, they describe,
īŽ Patterns of disease occurrence, in terms of,Patterns of disease occurrence, in terms of,
īŽ Who gets sick and/or who does notWho gets sick and/or who does not
īŽ Where rates are highest and lowestWhere rates are highest and lowest
īŽ Temporal patterns of diseaseTemporal patterns of disease
īŽ Data provided are useful for,Data provided are useful for,
īŽ Public health administrators (for allocation of resources)Public health administrators (for allocation of resources)
īŽ Epidemiologists (first step in risk factor determination)Epidemiologists (first step in risk factor determination)
20. Correlational Studies (Ecological Studies)Correlational Studies (Ecological Studies)
īŽ Uses measures that represent characteristics ofUses measures that represent characteristics of
entire populationsentire populations
īŽ It describes outcomes in relation to age, time,It describes outcomes in relation to age, time,
utilization of services, or exposuresutilization of services, or exposures
īŽ ADVANTAGESADVANTAGES
īŽ We can generate hypotheses for case-control studies andWe can generate hypotheses for case-control studies and
environmental studiesenvironmental studies
īŽ We can target high-risk populations, time-periods, orWe can target high-risk populations, time-periods, or
geographic regions for future studiesgeographic regions for future studies
21. Correlational StudiesCorrelational Studies
īŽ LIMITATIONSLIMITATIONS
īŽ Because data are for groups, we cannot link disease andBecause data are for groups, we cannot link disease and
exposure in individualexposure in individual
īŽ We cannot control for potential confoundersWe cannot control for potential confounders
īŽ Data represent average exposures rather than individualData represent average exposures rather than individual
exposures, so we cannot determine a dose-responseexposures, so we cannot determine a dose-response
relationshiprelationship
īŽ Caution must be taken to avoid drawing inappropriateCaution must be taken to avoid drawing inappropriate
conclusions, orconclusions, or ecological fallacyecological fallacy
22. Patterns of disease Occurrence :Patterns of disease Occurrence :
CorrelationCorrelation ofof PopulationPopulation statisticsstatistics
īŽ EcologicEcologic (( correlationcorrelation ) studies) studies
ī Used as first step in determining associationUsed as first step in determining association
plotplot :: disease (population) burden [ Y axis ]disease (population) burden [ Y axis ]
vs.vs. prevalence of ârisk factorâ [ X axis ]prevalence of ârisk factorâ [ X axis ]
e.g. smoking vs. lung cancere.g. smoking vs. lung cancer
īŽ -- correlation coefficient : r ; + 1 to -1-- correlation coefficient : r ; + 1 to -1
īŽ Quantifies linear relationship between exposure & diseaseQuantifies linear relationship between exposure & disease
23. Case Reports (case series)Case Reports (case series)
īŽ Report of a single individual or a group ofReport of a single individual or a group of
individuals with the same diagnosisindividuals with the same diagnosis
īŽ AdvantagesAdvantages
īļ We can aggregate cases from disparate sources to generateWe can aggregate cases from disparate sources to generate
hypotheses and describe new syndromeshypotheses and describe new syndromes
Example: hepatitis, AIDSExample: hepatitis, AIDS
īŽ LimitationsLimitations
īļ We cannot test for statistical association because there is noWe cannot test for statistical association because there is no
relevant comparison grouprelevant comparison group
īļ Based on individual exposure {may simply be coincidental}Based on individual exposure {may simply be coincidental}
24. Case report/Case series(contd.)Case report/Case series(contd.)
īŽ ImportantImportant interfaceinterface between clinical medicine &between clinical medicine &
epidemiologyepidemiology
īŽ Most common type of studies published inMost common type of studies published in
medical journals{1/3medical journals{1/3rdrd
of all}of all}
īŽ e.g. Frisbee finger , break dancing necke.g. Frisbee finger , break dancing neck
īŽ AIDS ~ b/w oct1980-may81, 5 cases of P.cariniiAIDS ~ b/w oct1980-may81, 5 cases of P.carinii
pneumonia were diagnosed among previously healthypneumonia were diagnosed among previously healthy
young homosexual males in L.A.young homosexual males in L.A.
25. Cross-Sectional Studies (prevalence studies)Cross-Sectional Studies (prevalence studies)
īŽ Measures disease and exposure simultaneously in aMeasures disease and exposure simultaneously in a
well-defined populationwell-defined population
īŽ AdvantagesAdvantages
īŽ They cut across the general population, not simply thoseThey cut across the general population, not simply those
seeking medical careseeking medical care
īŽ Good for identifying prevalence of common outcomes, suchGood for identifying prevalence of common outcomes, such
as arthritis, blood pressure or allergiesas arthritis, blood pressure or allergies
īŽ LimitationsLimitations
īŽ Cannot determine whether exposure preceded diseaseCannot determine whether exposure preceded disease
īŽ It considers prevalent rather than incident cases, resultsIt considers prevalent rather than incident cases, results
will be influenced by survival factorswill be influenced by survival factors
īŽ Remember: P = I x DRemember: P = I x D
26. Cross-Sectional StudiesCross-Sectional Studies
ī Can be used as a type of analytic study for testingCan be used as a type of analytic study for testing
hypothesis, when;hypothesis, when;
īļ Current values of exposure variables are unalterable overCurrent values of exposure variables are unalterable over
timetime
īļ Represents value present at initiation of diseaseRepresents value present at initiation of disease
īļ E.g. eye colour or blood groupE.g. eye colour or blood group
īļ If risk factor is subject to alterations by disease, onlyIf risk factor is subject to alterations by disease, only
hypothesis formulation can be donehypothesis formulation can be done
27. The epidemiologic approach:The epidemiologic approach:
Steps to public health actionSteps to public health action
MEASURES
īą
Counts
īą
Times
īą
Rates
īą
Risks/Odds
īą
Prevalence
METHODS
īą
Design
īą
Conduct
īą
Analysis
īą
Interpretation
ALTERNATIVE
EXPLANATION
S
īļ
Chance
īļ
Bias
īļ
Confounding
INFERENCES
īļ
Epidemiologic
īļ
Causal
ACTION
ī
Behavioural
ī
Clinical
ī
Community
ī
Environmental
DESCRIPTIVE
īŽ
What (case
definition)
īŽ
Who (person)
īŽ
Where (place)
īŽ
When (time)
īŽ
How many
(measures)
ANALYTIC
īŽ
Why (Causes)
īŽ
How (Causes)
30. What are the three categories ofWhat are the three categories of
descriptive epidemiologic clues?descriptive epidemiologic clues?
īŽ âĄâĄ Person:Person: WhoWho is getting sick?is getting sick?
īŽ âĄâĄ Place:Place: WhereWhere is the sickness occurring?is the sickness occurring?
īŽ âĄâĄ Time:Time: WhenWhen is the sickness occurring?is the sickness occurring?
īŽ PPT = person, place, timePPT = person, place, time
33. Tetanus â by year, USA, 1955-2000Tetanus â by year, USA, 1955-2000
During 2000, a total of 35 cases of tetanus were reported. The percentage of cases among persons aged 25-59 years
Has increased in the last decade. Note: A tetanus vaccine was first available in 1933.
0
100
200
300
400
500
600
700
800
900
1955 1960 1965 1970 1975 1980 1985 1990 1995 2000
Year
34. Possible Reasons for Changes inPossible Reasons for Changes in
TrendsTrends
īŽ ArtifactualArtifactual
īŽ Errors in numerator due toErrors in numerator due to
īŽChanges in the recognition of diseaseChanges in the recognition of disease
īŽChanges in the rules and procedures forChanges in the rules and procedures for
classification of causes of deathclassification of causes of death
īŽChanges in the classification code of causes ofChanges in the classification code of causes of
deathdeath
īŽChanges in accuracy of reporting age at deathChanges in accuracy of reporting age at death
īŽErrors in the denominator due to error in theErrors in the denominator due to error in the
enumeration of the populationenumeration of the population
35. Possible Reasons for Changes inPossible Reasons for Changes in
Trends (cont.)Trends (cont.)
īŽ RealReal
īŽ Changes in age distribution of the populationChanges in age distribution of the population
īŽ Changes in survivorshipChanges in survivorship
īŽ Changes in incidence of disease resultingChanges in incidence of disease resulting
fromfrom
īŽ Genetic factorsGenetic factors
īŽ Environmental factorsEnvironmental factors
36. Other phrasesOther phrases
īŧ Cyclic trends ~ recurrent alterations in
occurrence , interval or frequency of disease
īļSecular cyclicity
īLevels of immunizations
īBuild up of susceptibles
īŧ e.g. Hep A-7 yr cycle,Measles-2yr cycle
īļShort term cyclicity
īŧ Chickenpox,salmonella(yearly basis)
39. SeasonalSeasonal
īŽ A cyclic variation in disease frequencyA cyclic variation in disease frequency
by time of year & seasonby time of year & season..
īSeasonal fluctuations in,Seasonal fluctuations in,
īļEnvironmental factorsEnvironmental factors
īļOccupational activitiesOccupational activities
īļRecreational activitiesRecreational activities
42. Endemic, Epidemic and Pandemic
īŽ Endemic - The habitual presence (or usual occurrence) of a
disease within a given geographic area
īŽ Epidemic - The occurrence of an infectious disease clearly in
excess of normal expectancy, and generated
from a common or propagated source
īŽ Pandemic - A worldwide epidemic affecting an exceptionally
high proportion of the global population
Number
of Cases
of
Disease
Time
43. Time clusteringTime clustering
Time Place Cluster/disease clusterTime Place Cluster/disease cluster
īŽ A group of cases occur close togetherA group of cases occur close together
& have a well aligned distribution& have a well aligned distribution
patternpattern {{in terms ofin terms of time and placetime and place}}
īŽ Cluster analysis-used for rare or special diseaseCluster analysis-used for rare or special disease
events.events.
44. Time/Place clustering analysis using theTime/Place clustering analysis using the
Poisson modelPoisson model
{Poisson spatial/nearest neighbor distribution}{Poisson spatial/nearest neighbor distribution}
īŽ Poisson probability distribution is an inferential statistics probabilityPoisson probability distribution is an inferential statistics probability
measure.measure.
īŽ Describes objects/events as they are distributed geographically.Describes objects/events as they are distributed geographically.
īŽ Geographical area divided into a series of equal square areas.Geographical area divided into a series of equal square areas.
īŽ Randomization i.e. each case has equal probability of falling into eachRandomization i.e. each case has equal probability of falling into each
square.square.
īŽ If clustering occurs, probability of cause-effect relationship goes up &If clustering occurs, probability of cause-effect relationship goes up &
vice versa.vice versa.
45. PlacePlace
īŽ Diagnosis is Made
īŽ Contact occurred
between agent and
host
īŽ Source became
infected
Geographic Area Example Action Level
Home â Patient ill
Restaurant â Food
Eaten
Farm â Eggs Infected
Investigation
Control
Prevention
46.
47. PersonPerson
Age Hobbies
Sex Pets
Occupation Travel
Immunization status Personal Habits
Underlying disease Stress
Medication Family unit
Nutritional status School
Socioeconomic factors Genetics
Crowding Religion
48. Descriptive epidemiologyDescriptive epidemiology ::
Patterns of Disease OccurrencePatterns of Disease Occurrence
īŽ distributiondistribution of disease in populationsof disease in populations
numerator ( âeventâ count ) / denominator ( group âatnumerator ( âeventâ count ) / denominator ( group âat
riskâ )riskâ )
īŽ by âby âpersonpersonâ : age , race / ethnicity , gender ,â : age , race / ethnicity , gender ,
occupation , education , marital status , geneticoccupation , education , marital status , genetic
marker , sexual preferencemarker , sexual preference
īŽ by âby âplaceplaceâ : residence (urban vs. rural) , worksite ,â : residence (urban vs. rural) , worksite ,
social eventsocial event
īŽ by âby âtimetimeâ : week , month , year ; sporadic , seasonalâ : week , month , year ; sporadic , seasonal
, trends, trends
--- incubation period ; latency--- incubation period ; latency
49. Sources of informationSources of information
īŽ Census dataCensus data
īŽ Vital statistical recordsVital statistical records
īŽ Employment health examinationsEmployment health examinations
īŽ Clinical records from hospitalsClinical records from hospitals
īŽ National figures on food consumption ,National figures on food consumption ,
medications, health events etcmedications, health events etc
50. Epidemiologic (Epidemiologic ( scientificscientific ) Approach) Approach
īŽ 1. Identify a PROBLEM1. Identify a PROBLEM ::
clinical suspicion ; case series ; review of medical literatureclinical suspicion ; case series ; review of medical literature
īŽ 2. Formulate a HYPOTHESIS2. Formulate a HYPOTHESIS ( asking the right question ) ;( asking the right question ) ;
good hypotheses are: Specific, Measurable, and Plausiblegood hypotheses are: Specific, Measurable, and Plausible
īŽ 3. TEST that HYPOTHESIS3. TEST that HYPOTHESIS ( assumptions vs. type of data )( assumptions vs. type of data )
īŽ 4. always Question the VALIDITY of the result(s)4. always Question the VALIDITY of the result(s) ::
Chance ; Bias ; and CausalityChance ; Bias ; and Causality
51. Epidemiologic Study: threats to ValidityEpidemiologic Study: threats to Validity
īļ ChanceChance : role of: role of randomrandom error in outcome measure(s)error in outcome measure(s)
( p - value ; power of the study and the confidence interval )( p - value ; power of the study and the confidence interval )
--- largely determined by sample size--- largely determined by sample size
īļ BiasBias : role of: role of systematicsystematic error in outcome measure(s)error in outcome measure(s)
īŽ SelectionSelection bias - subjects not representativebias - subjects not representative
īŽ InformationInformation bias - error(s) in subject data / classificationbias - error(s) in subject data / classification
īŽ ConfoundingConfounding - 3rd variable (causal) assoc. w/ both X and Y- 3rd variable (causal) assoc. w/ both X and Y
52. What is a hypothesis?What is a hypothesis?
īŽ An educated guessAn educated guess
īŽ an unproven ideaan unproven idea
īŽ based on observation or reasoning, that can bebased on observation or reasoning, that can be
proven or disproven through investigation.proven or disproven through investigation.
53. What goes into a hypothesis?What goes into a hypothesis?
īŽ Characteristics of the diseaseCharacteristics of the disease
īŽ The illnessThe illness
īŽ Established modes of transmissionEstablished modes of transmission
īŽ DistributionDistribution
īŽ In timeIn time
īŽ By placeBy place
īŽ By personBy person
54. Hypothesis formulationHypothesis formulation
īŽ 4 methods {derived from4 methods {derived from 5 canons of inductive5 canons of inductive
reasoningreasoning by John Stuart Mill}by John Stuart Mill}
īļ Method of differenceMethod of difference
īļ Method of agreementMethod of agreement
īļ Method of concomitant variationMethod of concomitant variation
īļ Method of analogyMethod of analogy
55. MeasuresMeasures
īŽ Morbidity: Refers to the presence of disease in aMorbidity: Refers to the presence of disease in a
populationpopulation
īŽ Mortality: Refers to the occurrence of death in aMortality: Refers to the occurrence of death in a
populationpopulation
56. Methods for MeasuringMethods for Measuring
How do we determine disease frequency for aHow do we determine disease frequency for a
population?population?
īŽ Rate = Frequency of defined events in specifiedRate = Frequency of defined events in specified
population for given time periodpopulation for given time period
īŽ Rates allow comparisons between two or moreRates allow comparisons between two or more
populations of different sizes or of a populationpopulations of different sizes or of a population
over timeover time
57. Compute Disease RateCompute Disease Rate
Number of persons at risk = 5,595,211Number of persons at risk = 5,595,211
Number of persons with disease = 17,382Number of persons with disease = 17,382
Rate =Rate = 17,382 persons with heart disease17,382 persons with heart disease
5,595,211 persons5,595,211 persons
== ..003107 heart disease / resident / year003107 heart disease / resident / year
58. RatesRates
Rates are usually expressed as integers andRates are usually expressed as integers and
decimals for populations at risk during specifieddecimals for populations at risk during specified
periods to make comparisons easier.periods to make comparisons easier.
.003107 heart disease / resident / year.003107 heart disease / resident / year x 100,000x 100,000
== 310.7310.7 heart disease /heart disease / 100,000100,000 residents / yearresidents / year
59. PrevalencePrevalence vsvs. Incidence. Incidence
īŽ Prevalence is the number ofPrevalence is the number of existingexisting cases ofcases of
disease in the population during a defineddisease in the population during a defined
period.period.
īŽ Incidence is the number ofIncidence is the number of newnew cases ofcases of
disease that develop in the population during adisease that develop in the population during a
defined period.defined period.
60. IncidenceIncidence
īŽ Incidence rate is a measure of theIncidence rate is a measure of the
probability of the event among persons atprobability of the event among persons at
risk.risk.
61. Incidence RatesIncidence Rates
īŽ Population denominator:Population denominator:
IR =IR = # new cases during time period X K# new cases during time period X K
specified populationspecified population at riskat risk
62. Example (Incidence Rate)Example (Incidence Rate)
During a six-month time period, a total of 53 nosocomialDuring a six-month time period, a total of 53 nosocomial
infections were recorded by an infection control nurseinfections were recorded by an infection control nurse
at a community hospital. During this time, there wereat a community hospital. During this time, there were
832 patients with a total of 1,290 patient days. What is832 patients with a total of 1,290 patient days. What is
the rate of nosocomial infections per 100 patient days?the rate of nosocomial infections per 100 patient days?
I R =
53 X 100
1,290 pt. days
=
4.1 infections per
100 pt. days
63. Mortality RatesMortality Rates
īŽ A special type of incidence rateA special type of incidence rate
īŽ Number of deaths occurring in a specifiedNumber of deaths occurring in a specified
population in a given time periodpopulation in a given time period
64. Use of Mortality ratesUse of Mortality rates
īŽ Mortality rates are used to estimate diseaseMortality rates are used to estimate disease
frequency whenâĻfrequency whenâĻ
īŽ incidence data are not available,
īŽ case-fatality rates are high,
īŽ goal is to reduce mortality among screened or
targeted populations
65. Mortality Rates: ExamplesMortality Rates: Examples
īŽ Crude mortalityCrude mortality: death rate in an entire: death rate in an entire
populationpopulation
īŽ Rates can also be calculated for sub-groups withinRates can also be calculated for sub-groups within
the populationthe population
īŽ Cause-specific mortalityCause-specific mortality: rate at which deaths: rate at which deaths
occur for a specific causeoccur for a specific cause
66. Mortality Rates: ExamplesMortality Rates: Examples
īŽ Case-fatalityCase-fatality: Rate at which deaths occur from a: Rate at which deaths occur from a
disease among those with the diseasedisease among those with the disease
īŽ Maternal mortalityMaternal mortality: Ratio of death from: Ratio of death from
childbearing for a given time period per numberchildbearing for a given time period per number
of live births during same time periodof live births during same time period
67. Mortality Rates: ExamplesMortality Rates: Examples
īŽ Infant mortalityInfant mortality: Rate of death for children less: Rate of death for children less
than 1 year per number of live birthsthan 1 year per number of live births
īŽ Neonatal mortalityNeonatal mortality: Rate of death for children: Rate of death for children
less than 28 days of age per number of liveless than 28 days of age per number of live
birthsbirths
68. PrevalencePrevalence
īŽ Prevalence: Existing cases in a specifiedPrevalence: Existing cases in a specified
population during a specified time period (bothpopulation during a specified time period (both
new and ongoing cases)new and ongoing cases)
īŽ Prevalence is a measure of burden of disease orPrevalence is a measure of burden of disease or
health problem in a populationhealth problem in a population
69. PrevalencePrevalence
Prevalence: The number of existing cases in thePrevalence: The number of existing cases in the
population during a given time period.population during a given time period.
PRPR == # existing cases during time period# existing cases during time period
population at same point in timepopulation at same point in time
Prevalence rates are often expressed as a percentage.Prevalence rates are often expressed as a percentage.
70. Factors Influencing PrevalenceFactors Influencing Prevalence
Increased by:
īŽ Longer duration of the
disease
īŽ Prolongation of life of
patients without cure
īŽ Increase in new cases
īŽ (increase in incidence)
īŽ In-migration of cases
īŽ Out-migration of
healthy people
īŽ In-migration of
susceptible people
īŽ Improved diagnostic
facilities
īŽ (better reporting)
Decreased by:
ī§Shorter duration of
disease
ī§High case-fatality
rate from disease
ī§Decrease in new
cases (decrease in
incidence)
ī§In-migration of
healthy people
ī§Out-migration of
cases
ī§ Improved cure rate
of cases
71. Basic Measures of AssociationBasic Measures of Association
īŽ Relative risk& odds ratioRelative risk& odds ratio
īŽ We often need to know the relationship betweenWe often need to know the relationship between
an outcome and certain factors (e.g., age, sex,an outcome and certain factors (e.g., age, sex,
race, smoking status, etc.)race, smoking status, etc.)
īŽ Used to guide planning and interventionUsed to guide planning and intervention
strategiesstrategies
72. 2 x 2 contingency table for Calculation of2 x 2 contingency table for Calculation of
Measures of AssociationMeasures of Association
  Outcome OutcomeÂ
ExposureExposure PresentPresent AbsentAbsent TOTALTOTAL
PresentPresent aa bb a+ba+b
AbsentAbsent cc dd c+dc+d
TOTALTOTAL a+ca+c b+db+d a+b+c+da+b+c+d
Note: âExposureâ is a broad term that represents any
factor that may be related to an outcome.
73. Relative RiskRelative Risk
īŽ Ratio of the incidence rates between two groupsRatio of the incidence rates between two groups
īŽ Can only be calculated from prospective studiesCan only be calculated from prospective studies
(cohort studies)(cohort studies)
īŽ InterpretationInterpretation
īŽ RR > 1: Increased risk of outcome among âexposedâRR > 1: Increased risk of outcome among âexposedâ
groupgroup
īŽ RR < 1: Decreased risk, or protective effects, amongRR < 1: Decreased risk, or protective effects, among
âexposedâ groupâexposedâ group
īŽ RR = 1: No association between exposure andRR = 1: No association between exposure and
outcomeoutcome
74. Calculation of Relative RiskCalculation of Relative Risk
incidence rate among exposedincidence rate among exposed
RR =RR =
incidence rate among non-exposedincidence rate among non-exposed
75. Calculation of Relative RiskCalculation of Relative Risk
  OutcomeOutcome Â
ExposureExposure PresentPresent AbsentAbsent TOTALTOTAL
PresentPresent aa bb a+ba+b
AbsentAbsent cc dd c+dc+d
TOTALTOTAL a+ca+c b+db+d a+b+c+da+b+c+d
Relative Risk =
a
a b
c
c d
+
īŖĢ
īŖ
īŖŦ
īŖļ
īŖ¸
īŖˇ
+
īŖĢ
īŖ
īŖŦ
īŖļ
īŖ¸
īŖˇ
76. Relative Risk Case StudyRelative Risk Case Study
  Birth WeightBirth Weight
Smoking statusSmoking status <2500 g<2500 g >>2500 g2500 g TOTALTOTAL
SmokerSmoker 120120 240240 360360
Non-smokerNon-smoker 6060 580580 640640
TOTALTOTAL 180180 820820 10001000
Smoking and low birth weight
77. Answers to Relative Risk Case StudyAnswers to Relative Risk Case Study
īŽ 1. Incidence of LBW among1. Incidence of LBW among
smokerssmokers
īŽ 2. Incidence of LBW among2. Incidence of LBW among
non-smokersnon-smokers
īŽ 3.3. Relative riskRelative risk for having afor having a
LBW baby among smokersLBW baby among smokers
versus non-smokersversus non-smokers
= =
1 2 0
3 6 0
1 0 0 0 3 3 3 3x , .
= =
6 0
6 4 0
1 0 0 0 9 3 8x , .
= â
3 3 3 3
9 3 8
3 6
.
.
.
78. Understanding Probability and OddsUnderstanding Probability and Odds
īŽ Probability:Probability: Chance or risk of an event occurring (aChance or risk of an event occurring (a
proportion)proportion)
īŽ Probability=Probability= no. of times an event occursno. of times an event occurs
no. of times an event can occurno. of times an event can occur
īŽ Odds:Odds: ratio of the probability of an event occurring toratio of the probability of an event occurring to
the probability of an event not occurringthe probability of an event not occurring
īŽ Odds = P/(1-P)Odds = P/(1-P)
79. Calculation of Odds RatioCalculation of Odds Ratio
  OutcomeOutcome Â
ExposureExposure PresentPresent AbsentAbsent TOTALTOTAL
PresentPresent aa bb a+ba+b
AbsentAbsent cc dd c+dc+d
TOTALTOTAL a+ca+c b+db+d a+b+c+da+b+c+d
Odds Ratio =
a d
b c
80. Odds RatioOdds Ratio
īŽ The odds ratio (OR) is a ratio of two odds.The odds ratio (OR) is a ratio of two odds.
īŽ The OR can be calculated for all three studyThe OR can be calculated for all three study
designsdesigns
ī Cross-sectionalCross-sectional
ī Case-controlCase-control
ī Cohort.Cohort.
81. Various approaches to Odds ratioVarious approaches to Odds ratio
īŽ Cross product/odds ratioCross product/odds ratio
ī 2 x 2 contingency table (ad/bc)2 x 2 contingency table (ad/bc)
īŽ Prevalence odds ratioPrevalence odds ratio
ī cross sectional studiescross sectional studies
īŽ Exposure odds ratioExposure odds ratio(( odds of exposure in diseased vs. nondiseased)odds of exposure in diseased vs. nondiseased)
ī In rare cases or exotic diseasesIn rare cases or exotic diseases
īŽ Disease odds/Rate odds ratioDisease odds/Rate odds ratio(o(odds of getting a disease if exposeddds of getting a disease if exposed
or unexposed)or unexposed)
ī Cohort & cross sectionalCohort & cross sectional
īŽ Risk odds ratioRisk odds ratio
ī Cross sectional ,cohort & case controlCross sectional ,cohort & case control
82. Odds RatioOdds Ratio
īŽ For cohort & cross sectional studiesFor cohort & cross sectional studies: OR is a: OR is a
ratio of the odds of the outcome in exposedratio of the odds of the outcome in exposed
persons to the odds of the outcome in non-persons to the odds of the outcome in non-
exposed persons.exposed persons.
īŽ For case-control studiesFor case-control studies: OR is a ratio of the: OR is a ratio of the
odds of exposure in cases to the odds ofodds of exposure in cases to the odds of
exposure in controls.exposure in controls.
īŽ Provides an estimate of the relative risk whenProvides an estimate of the relative risk when
the outcome is rarethe outcome is rare
83. Interpretation of Odds RatioInterpretation of Odds Ratio
īŽ OR > 1: Increased odds of exposure among thoseOR > 1: Increased odds of exposure among those
with outcomewith outcome
īŽ OR < 1: Decreased odds, or protective effects,OR < 1: Decreased odds, or protective effects,
among those with outcomeamong those with outcome
īŽ OR = 1: No association between exposure andOR = 1: No association between exposure and
outcomeoutcome
84. Keeping the Terms StraightKeeping the Terms Straight
īŽ ââRisk ratioâ = ârelative riskâRisk ratioâ = ârelative riskâ
īŽ ââRelative oddsâ = âodds ratioâRelative oddsâ = âodds ratioâ
īŽ Remember â the key is recognizing the termsRemember â the key is recognizing the terms
âriskâ and âoddsââriskâ and âoddsâ
85. Appropriateness of MeasuresAppropriateness of Measures
īŽ Remember that the relative risk can only beRemember that the relative risk can only be
calculated in prospective studiescalculated in prospective studies
īŽ Odds ratio can be calculated for any designOdds ratio can be calculated for any design
īŽ Cohort / prospectiveCohort / prospective
īŽ Case-controlCase-control
īŽ Cross-sectionalCross-sectional
86. InferenceInference
īŽ The relative risk and odds ratio provide theThe relative risk and odds ratio provide the
magnitude of difference between some factormagnitude of difference between some factor
and an outcomeand an outcome
īŽ How do we know if the magnitude isHow do we know if the magnitude is statisticallystatistically
significant?significant?
87. Confidence IntervalsConfidence Intervals
īŽ A confidence interval is a range of values that isA confidence interval is a range of values that is
likely (e.g., 95%) to contain the true value in thelikely (e.g., 95%) to contain the true value in the
underlying populationunderlying population
88. The 10 Steps of Outbreak InvestigationThe 10 Steps of Outbreak Investigation
īŽ Prepare for field workPrepare for field work
īŽ Establish the existence of an outbreakEstablish the existence of an outbreak
īŽ Verify the diagnosisVerify the diagnosis
īŽ Define & identify casesDefine & identify cases
īŽ PerformPerform descriptive epidemiologydescriptive epidemiology
īŽ Develop hypothesesDevelop hypotheses
īŽ PerformPerform analytic epidemiologyanalytic epidemiology
īŽ Refine hypotheses & conduct additional studiesRefine hypotheses & conduct additional studies
īŽ Implement control & prevention measuresImplement control & prevention measures
īŽ Communicate findingsCommunicate findings
89. Objectives of Descriptive EpidemiologyObjectives of Descriptive Epidemiology
īŽ To evaluate trends in health and disease and allowTo evaluate trends in health and disease and allow
comparisons among countries and subgroups withincomparisons among countries and subgroups within
countriescountries
īŽ To provide a basis for planning, provision andTo provide a basis for planning, provision and
evaluation of servicesevaluation of services
īŽ To identify problems to be studied by analytic methodsTo identify problems to be studied by analytic methods
and to test hypotheses related to those problemsand to test hypotheses related to those problems