Measures of Disease Frequency 0903_gaohongcai(1)

8/14/2019 Measures of Disease Frequency 0903_gaohongcai(1)

1/89

Me asu res o fDise ase sFrequency(1)Gao Hongcai

Email: [email protected]

School of Public Health, Jining MedicalCollege


2/89

03/20/09 2

Introduction Epidemiologists use a variety of

methods to summarize data. One

fundamental method is the frequencydistribution. The frequency distributionis a table which displays how manypeople fall into each category of a

variable such as age, income level, ordisease status.

In later lessons you will learn about

frequency distribution (time placepeople).


3/89

03/20/09 3

Introduction cases:

people who have a disease or conditionbeing studiedare generally referred to as

cases.

People without the disease are called non-

cases.

non-cases :


4/89

03/20/09 4

IntroductionThe most common form of

organizing epidemiologic informationis through collecting information

about cases and non-cases.


5/89

03/20/09 5

Introduction counting cases of disease in a

population is the unique domain of

epidemiologyit is the corecomponent of disease surveillanceand a critical step in investigating an

outbreak.


6/89

03/20/09 6

Introduction Case counts must be placed in proper

perspective, however, by using rates to

characterize the risk of disease for apopulation. Calculating rates for differentsubgroups of age, sex, exposure historyand other characteristics may identify high-

risk groups and causal factors. Such information is vital to the

development and targeting of effectivecontrol and prevention measures.


7/89

03/20/09 7

Introduction The numbers are usually translated into

rates.

Rates indicate the risk of disease for apopulation.

This helps to identify high-risk groups and

causal factors. Data elements collected from cases and

non-cases are called variables (age, sex,race, weight, height, temperature, bloodpressure, disease, death etc)


8/89

03/20/09 8

Why measure disease?To quantify or measure the

occurrence of disease is fundamental

in understanding: Description of health events in the

population

Impact of disease in the population Possible associations between factors

and disease

Possible mechanisms of spread


9/89

03/20/09 9

Rates, Ratios and

Proportions Variables or data elements are summarized

to provide number of cases, proportions,rates and ratios = frequency

In epidemiology many variables have onlytwo possible categories these are called dichotomous variables(e.g. disease or no disease; alive or dead).

The frequency measures we use withdichotomous variables are ratios,proportions and rates.


10/89

03/20/09 10

Rates, Ratios and

Proportions Before you learn about specific measures,

it is important to understand the

relationship between the three types ofmeasures and how they differ from eachother.

All three measures are based on the sameformula:

Ratio, proportion, rate = 10n

y

x


11/89

03/20/09 11

Rates, Ratios and

Proportions

10n(10 to the nth power),is a constant that we

use to transform the result of the divisioninto a uniform quantity. The size of 10n

may equal 1, 10, 100,1000 and so on

depending upon the value ofn.


12/89

03/20/09 12

Rates, Ratios and

Proportions For example,

100 = 1

101 = 10 102 = 10 10 = 100 103 = 10 10 10 = 1000

You will learn what value of 10n

touse when you learn about specificratios, proportions, and rates.


13/89

03/20/09 13

= 5 /2 = 2.5 /1

Ratio

The quotient of 2 numbers Numerator NOT necessarily INCLUDED in the

denominator

Allows to compare quantities of different nature

Rates, Ratios and

Proportions


14/89

03/20/09 14

For example, the sex of children attendingan immunization clinic could be compared

in either of the following ways:

In the first option,x(female) is completelyindependent ofy(male).

In the second,x(female) is included iny(all).Both examples are ratios.


15/89

03/20/09 15

2--- = 0.5 100 = 50%4

Proportion

The quotient of 2 numbers Numerator NECESSARELY INCLUDED

in the denominator

Quantities have to be of same nature

Proportion always ranges between 0 and 1

Percentage = proportion 100


16/89

03/20/09 16

Of the two ratios shown above,

the first is not a proportion, becausexis not a part ofy.

The second is a proportion, because

xis part ofy.


17/89

03/20/09 17

Rate

2----- = 0.02 /year100

Observed in 1998

Numerator- number EVENTS observed for a given time

Denominator- population in which the events occur

(population at risk)- includes time

The quotient of 2 numbers

Speed of occurrence of an event over time


18/89

03/20/09 18

rate, is often a proportion, with an addeddimension: it measures the occurrence of

an event in a population over time. The basic formula for a rate is as follows:


19/89

03/20/09 19

Notice three important aspects of this

formula.

The persons in the denominator must reflect thepopulation from which the cases in the numerator

arose.

The counts in the numerator and denominatorshould cover the same time period.

In theory, the persons in the denominator must

be at risk for the event, that is, it should have

been possible for them to experience the event.


20/89

03/20/09 20

Rates and Proportions Rates: tell us how fast the disease isoccurring in a population

Proportions: tell us what fraction of thepopulation is affected

Rates and Proportions can be used to estimate

the extent and significance of disease burden ina community or population


21/89

03/20/09 21

As you can see from the above discussion,

ratios, proportions, and rates are NOT three

distinctly different kinds of frequency

measures. They are all ratios:

proportions are a particular type ratio, and

some rates are a particular type of

proportion.


22/89

03/20/09 22

In epidemiology, however, we often

shorten the terms for these

measures in a way that makes it

sound as though they are completely

different. When we call a measure aratio, we usually mean a

nonproportionalratio;


23/89

03/20/09 23

when we call a measure a

proportion, we usually mean a

proportional ratio that doesnt

measure an event over time, and

when we use the term rate, wefrequently refer to a proportional ratio

that does measure an event in a

population over time.


24/89

03/20/09 24

We also use ratios, proportions, and,

most important rates to describethree aspects of the human

condition:

morbidity (disease),

mortality (death),

and natality (birth).


25/89

03/20/09 25

Morbidity FrequencyMeasures


26/89

03/20/09 26

To describe the presence ofdisease in a population, or the

probability (risk) of its occurrence,

we use one of the morbidity

frequency measures.

In public healthterms, disease

includes illness,injury, or disability.


27/89

03/20/09 27

Table 1 shows several morbiditymeasures.


28/89

03/20/09 28

All of these can be further elaborated intospecific measures for age, race, sex,

or some other characteristic of aparticular population being described.


29/89

03/20/09 29

Table 2. shows a summary of the formulas forfrequently used morbidity measures.


30/89

03/20/09 30

I will describe how you calculateeach of the morbidity measures

and when you would use it later.


31/89

03/20/09 31

Incidence Rate

Definition:

Incidence refers to the occurrence of

new cases in a population during agiven period of time.


32/89

03/20/09 32

Incidence Rate

Incidence rate is the most commonway of measuring and comparing the

frequency of disease in populations. We use incidence rates instead of raw

numbers for comparing disease

occurrence in different populationsbecause rates adjust for differences inpopulation sizes.


33/89

03/20/09 33

Incidence Rate

Rate of development of disease during agiven period of time.

The incidence rate expresses theprobability or risk of illness in a populationover a period of time.

It is used to estimate the risk of developinga disease in a specified population during aspecified period of time.


34/89

03/20/09 34

Incidence Rate

Since incidence is a measure of risk,when one population has a higher

incidence of disease than another, wesay that the first population is at ahigher risk of developing disease thanthe second, all other factors being

equal. We can also express this by saying that

the first population is a high-riskgrouprelative to the second population.


35/89

03/20/09 35

Incidence Rate

An incidence rate (sometimesreferred to simply as incidence) is a

measure of the frequency with whichan event, such as a new case ofillness, occurs in a population over a

period of time.


36/89

03/20/09 36

Incidence Rate

The formula for calculating anincidence rate as follows:

k=periodtimesametheduringriskatpopulation

periodgiven timeaduringocurringcasesnewrateIncidence

k=10n,

n=2,3,4,5,6


37/89

03/20/09 37

Incidence Rate

Three key elements: Only new cases included in numerator

Total population at risk in thedenominator

Time element period over which new

cases developed


38/89

03/20/09 38

-

numerator

The numerator of an incidence rate

should reflect newcases of disease

which occurred or were diagnosed

during the specified period.

The numerator should notinclude

cases which occurred or were

diagnosed earlier.


39/89

03/20/09 39

-

denominator

Notice that the denominatoris thepopulation at risk. This means that persons

who are included in the denominator shouldbe able to develop the disease that is beingdescribed during the time period covered.

Unfortunately, unless we conduct a special

study, we usually cannot identify andeliminate persons who are not susceptibleto the disease from available populationdata.


40/89

03/20/09 40

-

denominator

In practice, we usually use Census population

counts or estimates for the midpoint of the time

period under consideration.

If the population being studied is small and very

specific, howeversuch as a nursing home

populationwe can and should use exact

denominator data.


41/89

03/20/09 41

-

denominator

The denominator should representthe population from which the

cases in the numerator arose. E.g. : If the numerator was limited

to civilian cases, it was necessary

for us to restrict the denominatorto civilians as well.


42/89

03/20/09 42

-

denominator

Depending on the circumstances, the mostappropriate denominator will be one of the

following: average size of the population over the time

period.

size of the population (either total or at risk)

at the middle of the time period. size of the population at the start of the

time period.


43/89

03/20/09 43

Incidence rate -10n For 10n, any value ofn can be used.

For most nationally notifiable

diseases, a value of 100,000 or 105 isused for 10n.

Otherwise, we usually select a value

for 10n so that the smallest ratecalculated in a series yields a smallwhole number (for example, 4.2/100,

not 0.42/1,000; 9.6/100,000, not


44/89

03/20/09 44

Incidence rate -10n Since any value ofn is possible, the

investigator should clearly indicate

which value is being used. If weselected a value of 100,000, ourincidence rate is reported as 297.4

per 100,000.


45/89

03/20/09 45

Incidence rate -time Rates imply a change over time. For disease incidence rates, the

change is from a healthy state todisease. The period of time mustbe specified.

For surveillance purposes, the periodof time most commonly used is thecalendar year, but any interval maybe used as long as the limits of the

interval are identified.


46/89

03/20/09 46

Incidence rate -types When the denominator is the size of the

population at the start of the timeperiod, the measure is sometimes calledcumulative incidence.

This measure is a proportion, becauseall persons in the numerator are also inthe denominator. It is a measure of theprobability or riskof disease, i.e.,what proportion of the population willdevelop illness during the specified timeperiod.


47/89

03/20/09 47

-types

In contrast, the incidence rate islike velocity or speed measured in

miles per hour. It indicates howquicklypeople become illmeasured in people per year.


48/89

03/20/09 48

Cumulative Incidence

There are two types of Incidence rates :

1. Cumulative Incidence (CI): is the

proportion of people who become diseasedduring a specified period of time.

k

riskatpopulationTotal

periodspecificaduringdiseaseaofcasesnewofNumber

k=10n,

n=2,3,4,5,6


49/89

03/20/09 49


CI provides an estimate of theprobability, or risk that an individual will

develop a disease during a specifiedperiod of time.

The time period must be clearly

specified when reporting the CI.


50/89

03/20/09 50


The CI assumesThat the entire population is at risk at

the beginning of the study periodThat the populations has been followed

for the specified time interval


51/89

03/20/09 51


In 2000: 733,151 new cases of gonorrhea were

reported among the United States civilianpopulation.

The mid-year civilian population wasestimated to be 246,552,000.

CI=[733,151/246,552,000] x 1,000=

2.97 per 1000 population.


52/89

03/20/09 52

Incidence Density

2. Incidence density (ID) = measure of riskin a changing population where people are

free of disease at start and observed foronset of disease for different time periods

Shows how rapidly cases develop

In presenting an incidence density, it is

essential to specify the relevant time units.-i.e. number of cases per person-day,person-month, person-year, etc.


53/89

03/20/09 53

Incidence Density

Person-time is the number of personsat risk for disease multiplied by the

length of time they are observed foronset of disease

Person years = # of person # of

years followed up


54/89

03/20/09 54

Incidence Density Example: In a cohort study, 600 contraceptive

users were followed over the course of three yearsas follows:

100 women for 1year 200 women for 2 years

300 women for 3 years

The number of person-years of observation in this

study is: 100 women x 1 = 100 person-years

200 women x 2 = 400 person-years

300 women x 3 = 900 person-years

Total = 1400 person-years


55/89

03/20/09 55

Incidence Density

Of these 600 people 10 developed thromboembolism

The incidence rate ofthromboembolism is calculated as: (10/1400)*k = 7.1 per 1000 person

years of follow up.


56/89

03/20/09 56

Prevalence Rate

Prevalence, sometimes referred to as

prevalence rate, Is the proportion ofpersons in a population who have a

particular disease or attribute at a specified

point in time or over a specified period of

time.


57/89

03/20/09 57

Prevalence Rate

The formula for presence of disease is:

k=

periodtimesametheduringpopulation

periodgiven timeaduringcasesexisting-preandnewallPrevalence

k=10n

The formula for prevalence of anattribute is:

k=periodtimesametheduringpopulation

periodgiven timeaduringattributeexisting-preandnewallPrevalence


58/89

03/20/09 58

Prevalence Rate

Gives an indication of the burden ofdisease


59/89

03/20/09 59

Prevalence Rate

The value of 10n is usually 1 or 100for common attributes.

The value of 10n may be1,000,100,000, or even 1,000,000for rare traits and for most

diseases.


60/89

03/20/09 60

Prevalence Rate

There are two types of prevalence:

1. Point prevalence: is the presence

of a disease at a single point in time(snap shot).

kriskatpopulationpoint-midTotal

timeofpointspecificaatcasesexistingofNumber

k=10n,

n=2,3,4,5,6


61/89

03/20/09 61

Prevalence Rate

The amount of disease present in a

population is constantly changing.

Sometimes, we want to know how much ofa particular disease is present in a

population at a single point in timeto get

a kind of stop action or snapshot lookat the population with regard to that

disease.

We use point prevalence for that


62/89

03/20/09 62

Prevalence Rate

Point prevalence is not an incidence

rate, because the numeratorincludes pre-existing cases;

it is a proportion, because the

persons in the numerator are also in

the denominator.


63/89

03/20/09 63

Prevalence Rate

At other times we want to know howmuch of a particular disease is present

in a population over a longer period.Then, we use period prevalence.


64/89

03/20/09 64

Prevalence Rate

2. Period Prevalence: the presence of adisease at a particular time interval .

kriskatpopulationperiod-midTotal

timeofperiodspecificaatcasesexistingofNumber

Prevalence usually refers to pointrevalence

k=10n,

n=2,3,4,5,6


65/89

03/20/09 65

Prevalence Rate

The numerator in period prevalence isthe number of persons who had a

particular disease or attribute at anytime during a particular interval.The interval can be a week, month, year,

decade, or any other specified timeperiod.

l i hi


66/89

03/20/09 66

Relationship BetweenPrevalence and Incidence

The prevalence and incidence of disease are

frequently confused. They are similar, but

differ in what cases are included in thenumerator.

Numerator ofIncidence = new cases

occurring during a given time period

Numerator ofPrevalence = all cases

(existing and new)


67/89

03/20/09 67

As you can see, the numerator of anincidence rate consists only of persons

whose illness began during a specifiedinterval.

The numerator for prevalence includesallpersons ill from a specified cause

during a specified interval (or at aspecified point in time) regardless ofwhen the illness began. It includesnot only new cases, but also old cases

during the specified interval.


68/89

03/20/09 68


Prevalence and Incidence are related

by average duration (D) of disease ina stable population


69/89

03/20/09 69

High prevalence of a disease within apopulation may reflect high risk, or itmay reflect prolonged survival withoutcure.

Conversely, low prevalence mayindicate low incidence, a rapidly fatalprocess, or rapid recovery.


70/89

03/20/09 70

We often use prevalence ratherthan incidence to measure the

occurrence of chronic diseasessuch as osteoarthritis which havelong duration and dates of onset

which are difficult to pinpoint.

R l ti hi B t


71/89

03/20/09 71


Deaths,

Cured,

Lost...

DurationPrevalence

Incidence

P=ID

R l ti hi B t


72/89

03/20/09 72


If a disease lasts a long time, there will behigher prevalence in the population

If incidence is low but those affected havethe condition for long = the prevalence willbe high relative to the incidence

If disease is short-lived, there will be lowprevalence in the population

If incidence is high but the disease durationis short then the prevalence will be lowrelative to the incidence

R l ti hi B t


73/89

03/20/09 73


Example: a follow up study of theFramingham data showed males and

females had same IR for heart diseasebut females had a higher prevalenceof disease

WHY?

R l ti hi B t


74/89

03/20/09 74


PR = What proportion of thepopulation has this condition?

IR = At what rate do new casesarise over time in the population?

R l ti hi B t


75/89

03/20/09 75


Example: in Richmond City, 2006, there

were a total of 5,000 hepatitis B cases of

which 1,000 were newly diagnosed. The total population of Richmond City is

200,000.

Prevalence in 2006:(5,000/200,000)*1000=25 per 1,000

population

Incidence in 2006: (1,000/200,000)*

U f P l d


76/89

03/20/09 76

Uses of Prevalence andIncidence Measures

Prevalence is mostly used for

planning, evaluating

Incidence is used to identify causal

relationships

Factors which increase or decrease prevalence rate


77/89

03/20/09 77

Factors which increase or decrease prevalence rate

Rates Are Increased by Rates Are Decreased by

Immigration of ill cases Immigrationofhealthypersons

Emigration of healthy persons Emigration of ill cases

Immigration of susceptible cases orthose with potential of becomingcases

Improved cure rate of cases

Increaseddeath rates from the diseases

Prolongation of life of caseswithout cure (increase of durationof disease)

Decrease in occurrence of new cases

Shorter duration of disease

Increase in occurrence of new cases(increase in incidence)

Death


78/89

03/20/09 78

Attack Rate

Attack Rate: Is a variant of anincidence rate, applied to a narrowly

defined population observed for alimited time, such as during anepidemic or outbreak of a disease.

It is usually expressed as a percent, so10n equals 100.


79/89

03/20/09 79

Attack Rate

For a definedpopulation (the population at risk),during a limited time period,

100periodtheofbeginningat theriskatPopulation

periodtheduringpopulationtheamongcasesnewofNo.rateAttack =


80/89

03/20/09 80

Attack Rate

Example: Of 75 persons who attended achurch picnic, 46 subsequently developed

gastroenteritis. To calculate the attack rate ofgastroenteritis we first define the numeratorand denominator:

x= Cases of gastroenteritis occurringwithin the incubation period forgastroenteritis among persons who

attended the picnic = 46


81/89

03/20/09 81

Attack Rate

The attack rate of gastroenteritis is:

The risk of developing gastroenteritis in this

population was 61%.

%611007546 =


82/89

03/20/09 82

Attack Rate Notice that the attack rate is a proportion

the persons in the numerator are also inthe denominator.

This proportion is a measure of theprobabilityor riskof becoming a case.

In the example above, we could say that,

among persons who attended the picnic,the probability of developinggastroenteritis was 61%, or the risk ofdeveloping gastroenteritis was 61%.


83/89

03/20/09 83

Secondary Attack Rate

Secondary attack rate:is a measure ofthe frequency of new cases of a diseaseamong the contacts of known cases.

The formula is as follows:

k=contactsofnumberTotal

periodtheduringcasesprimaryofcontactsamongcaseofNo.rateattackSecondary

k=10n,

n=2,3,4,5,6


84/89

03/20/09 84

To calculate the total number of householdcontacts, we usually subtract the number ofprimary cases from the total number ofpeople residing in those households.


85/89

03/20/09 85

Secondary Attack Rate

Example:

Seven cases of hepatitis A occurred

among 70 children attending a child carecenter.

Each infected child came from a differentfamily.

The total number of persons in the 7affected families were 32

One incubation period later, 5 family

members of the 7 infected children also


86/89

03/20/09 86


87/89

03/20/09 87

Calculate the attack rate in thechild care center and the

secondary attack rate amongfamily contacts of those cases.


88/89

03/20/09 88

1. Attack rate in child care center:

x= cases of hepatitis A among

children in child care center = 7

y= number of children enrolled in

the child care center = 70

Attack rate= 100 = 100 =

10%y

x

70

7


89/89

2. Secondary attack rate:

x= cases of hepatitis A among family contacts of

children with hepatitis A = 5 y= number of persons at risk in the families (total

number of family memberschildren alreadyinfected) = 32 7 = 25

Secondary attack rate

= 100 = 100 = 20%y

x

25

5

Measures of Disease Frequency 0903_gaohongcai(1)

Documents

Transcript of Measures of Disease Frequency 0903_gaohongcai(1)