PM2ch33.pdf

Epidemiologic Measurement: Basic Concepts and Methods

Chapter 33

EPIDEMIOLOGIC MEASUREMENT: BASIC

CONCEPTS AND METHODS

LYNN I. LEVIN, P H D, MPH

INTRODUCTION

DEFINITION OF EPIDEMIOLOGY

BASIC MEASURES OF DISEASE FREQUENCY

DIRECT METHOD OF RATE ADJUSTMENT

TYPES OF VARIABLES

SUMMARY STATISTICS FOR CATEGORICAL AND CONTINUOUS VARI-

ABLES

METHODS FOR DISPLAYING DATA

SAMPLING

MAJOR TYPES OF STUDY DESIGNS AND MEASURES OF ASSOCIATION

USED IN EPIDEMIOLOGY

BIAS

HYPOTHESIS TESTING

STATISTICAL ASSOCIATIONS AND CAUSE-AND-EFFECT RELATIONSHIPS

SUMMARY

755

Military Preventive Medicine: Mobilization and Deployment, Volume 2

L.I. Levin ; Division of Preventive Medicine, Walter Reed Army Institute of Research, Silver Spring, MD 20910-7500

756

Epidemiologic Measurement: Basic Concepts and Methods

INTRODUCTION

This chapter will provide the public health profes-

sional with the basic tools and concepts of epidemi-

ology. The target audience is the military practitioner

of public health who may be deployed or working

in a public health setting where surveillance activi-

ties or rudimentary research studies are needed. The

epidemiologic methods covered in this chapter are

at the elementary level and are presented from the

more simple concepts to the more complex. Other

chapters in this volume specifically address out-

break investigations and surveillance activities (see

chapters 31, Disease and Nonbattle Injury Surveillance:

Outcome Measures for Force Health Protection, and

32, Outbreak Investigation). The epidemiologic

tools that are used to conduct these activities are

presented here. The focus is on data collection meth-

ods that are necessary to run programs, assess out-

breaks, allocate resources, or perform other activities

that are related to the formulation of public health

policy and practice.

DEFINITION OF EPIDEMIOLOGY

Epidemiology is the study of the distribution and

determinants of disease, injury, or health-related

factors in human populations. 1 In contrast to clinical

medicine, which focuses on the individual, epidemiol-

ogy evaluates groups of people or populations. A

basic premise of this discipline is that diseases, inju-

ries, or other medical conditions are not randomly

distributed across populations but rather vary ac-

cording to factors such as environmental exposures

(eg, deployment history) or personal characteristics

(eg, smoking status). As early as the 5th century BC,

Hippocrates hypothesized that the development of

human diseases may be associated with the environ-

ment. 2 The goal of the epidemiologist is to determine

who becomes ill and why, and this is accomplished

by comparing groups with differing characteristics.

BASIC MEASURES OF DISEASE FREQUENCY

Epidemiology is a quantitative science. One of

the fundamental measures used by epidemiologists

is the rate of disease. By computing rates, it is pos-

sible to compare two populations or groups. A rate

requires consideration of the population from which

the cases are derived:

lence depends on both the incidence of new cases

of disease and the duration of disease in those cases.

This measure typically is used to assess the need

for and costs of health services.

Number of Existing Cases of

Number of Events

Population at Risk of Event

Prevalence =

Disease at a Given Time

Total Population at a Given Time

Rate =

Typically, the denominator is the total population

at risk of disease, and the numerator is the num-

ber of people with disease. The rate is usually

expressed in some conventional base, such as

events per 1,000 individuals. There are several

types of rates used in epidemiology to describe

morbidity and mortality.

For example, a 15% prevalence of rash-associated

illness in units of an Army support battalion in

Operation Joint Endeavor, Bosnia, was determined in

this way 3 :

Prevalence = 69 Cases of Rash-associated Illness

466 Deployed Unit Members

= 0.148 = 15%

Prevalence

Incidence Rate

A commonly used measure in epidemiology is

the prevalence rate. Although it has historically

been called a rate, prevalence is a proportion rep-

resenting the fraction of the population that has a

disease at a single point in time. It is a snapshot of

the burden of disease in a defined population and

includes both new and existing cases. Thus, preva-

Incidence describes the rate of development of a

disease or a medical condition in a population over

a period of time, that is, the occurrence of new

events in a specified time period. This measure is

used to evaluate the causal or etiologic role of risk

factors and the development of disease.

757

Military Preventive Medicine: Mobilization and Deployment, Volume 2

Cumulative =

Number of New Cases

years. In calculating person-time, following 10 in-

dividuals for 3 years (30 person-years) is equivalent

to following one individual for 30 years. Incidence

density can range from zero to infinity.

For example, in a study of the incidence of HIV-

1 infection in the US Army during the period 1

November 1985 to 31 October 1993, a total of

1,061,768 active-duty soldiers were followed for a

total of 3,629,688 person-years of follow-up. Dur-

ing the period of the study, 978 soldiers with HIV-1

seroconversion were identified. 5

Incidence

During a Specified Period

Total Population at Risk of

Rate

the Disease During that Period

This measure is called the cumulative incidence

rate because it accumulates the number of cases

over time that derive from the population at risk

determined at the beginning of the time period (ie,

fixed population with the assumption that no indi-

viduals are lost to follow-up). The denominator con-

tains the count of people. Cumulative incidence is

a proportion that ranges from 0 to 1 (assuming only

one possible occurrence of the disease per person).

When reporting a cumulative incidence, the time

frame must be specified. Attack rates are a measure

of cumulative incidence.

For example, in a study among US Army soldiers

in infantry basic training, 303 individuals were

followed for 12 weeks of training to determine the

incidence of training-related injuries. Of the 303 sol-

diers, 112 developed one or more lower extremity

musculoskeletal injuries. 4

Incidence Density Rate

= 978 HIV-1 Soldiers

3,629,688 Person-years

= 2.7/10,000 Person-years

Prevalence and incidence measures are interrelated.

When the incidence rate has been constant over time (a

“steady state” situation), the duration of disease has re-

mained unchanged, and the prevalence of disease over-

all is low, then the prevalence (P) equals the product of

the incidence density (I) and the average duration ( — ):

Cumulative Incidence Over a 12-week Period

= 112 Soldiers Injured

303 Soldiers in Basic Training

= 0.369 = 37%

P = I • —

If two of these measures are known, the third can

be calculated.

Incidence density rates, like other rates, can be

calculated for subpopulations, such as different age,

sex, or race-ethnic groups. The following is an age-

specific incidence rate:

Another measure of incidence is the incidence

density rate. This is a measure of the instantaneous

rate of development of a disease in a population.

Incidence Density Rate

Number of New

Cases of a Disease During a Specified Period

Age-specific Incidence Rate

Number of New Cases in a Specified Age Group

Total Person-years of Observation for the Age Group

Total Person-time of Observation During

the Given Time Period

For example, the incidence rate of HIV-1 infection

among 20- to 24-year-olds in the study 5 cited above

was determined in this way:

The numerator of the incidence density is the num-

ber of new cases occurring in the population (the

same as in the cumulative incidence). The denomi-

nator, however, is the sum of each person’s time at

risk or the sum of the time that each person was

under observation and susceptible to the disease

rather than the count of individuals at the begin-

ning of the follow-up period. Thus, this measure

can incorporate data from a dynamic population:

individuals who are followed for various periods

of time or who are lost to follow-up. An incidence

density rate must include the relevant time units in

the denominator, such as person-days or person-

Age-specific Incidence Rate for Ages 20 to 24

405 HIV-1 Soldiers

1,216,125 Person-years Among Soldiers Aged 20 to 24

= 3.3/10,000 Person-years

Mortality Rate

Incidence rates describe the incidence of disease

in a population, whereas mortality rates describe

the incidence of death in a population. The crude

mortality rate can be determined using this equation:

758

Epidemiologic Measurement: Basic Concepts and Methods

Crude Mortality Rate

Cause-specific Mortality Rate

All Deaths in a Calendar Year

Total Population at Risk of Death During Year

• 10 n

Number of Deaths from a Specific Cause in a Year

Total Population During the Year

A cause-specific mortality rate was determined for

motor vehicle mortality among deployed service-

members to the Persian Gulf region for the period

1 August 1990 to 31 July 1991 6 :

For example, the following is the crude non–battle-

related mortality rate among deployed servicemem-

bers to the Persian Gulf region, 1 August 1990 to 31

July 1991. 6

Cause-specific Mortality Rate

Crude Mortality Rate

62 Deaths Due To Motor Vehicle Accidents

264,868 Person-years in the Persian Gulf Region

225 Nonbattle-related Deaths

264,868 Person-years in the Persian Gulf Region

= 85/100,000 Person-years

100,000 Person-years

All of these measures of morbidity and mortality

are used to monitor disease trends in surveillance

activities, identify outbreaks, and plan and evalu-

ate health services.

The cause-specific mortality rate is determined in a

similar fashion:

DIRECT METHOD OF RATE ADJUSTMENT

Rate adjustment allows the public health officer to

compare rates of disease or injury in two communi-

ties that have different demographic characteristics,

such as different age distributions or race-ethnic com-

positions. Because incidence rates or mortality rates

typically vary by age and race, it is necessary to ad-

just or standardize these rates by these factors so that

the rates can be compared on equal terms. Adjusted

rates are artificial in that they are only used for com-

parison purposes and do not describe a particular

population. The process of adjusting rates requires the

use of a standard population with which both com-

munities are compared and data on the factor-specific

rates of disease in both communities. If, for example,

injury rates in the Army are to be compared with in-

jury rates in the Navy, these rates would have to be

adjusted for age and sex, because the composition of

the two services differs in these factors. Texts such as

Kahn and Sempos 7 or Selvin 8 have examples on how

to perform a direct age adjustment.

TYPES OF VARIABLES

Categorical Variables

such as race and cause of death. Some discrete vari-

ables can be ordered or ranked. Examples of ordered

variables are severity of pain (eg, mild, moderate, in-

tense) and military rank.

In addition to obtaining information on rates of

disease, epidemiologic studies can also collect in-

formation on many variables about individuals.

Variables that are divided into categories or are as-

signed codes are called categorical variables. Each

category is defined and there are a limited number

of values that can be measured. Dichotomous vari-

ables are categorical variables that assume only two

values, such as male and female or inducted into mili-

tary services and not inducted. Polychomotous vari-

ables can be divided into more than two categories,

Continuous Variables

Variables that can be measured on a continu-

ous scale, such as height or 1-mile run times, are

considered continuous variables. The values that

these variables can assume are only limited by

the level of accuracy of the scale on which they

are measured.

759

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: