Descriptive Epidem iology & Descriptive Epidem iology & - - PowerPoint PPT Presentation

Descriptive Epidem iology & Descriptive Epidem iology & Study design Study design

Potjam an Potjam an Siriarayapon Siriarayapon

Bureau of Epidem iology Bureau of Epidem iology

2

Outline

General concept of descriptive epidemiology Study design in epidemiology

Descriptive Analytic

3

KEY CONCEPTS

฀ ฀ “

“Descriptive epidem iology Descriptive epidem iology” ” is the first step in epidemiological study. Careful observation of available information alone have led many successful preventive measures in the past. Describing health or disease occurrence according to place, person and time can lead to very meaningful hypothesis formulations.

4

TIME

Disease rates change over time. Some of

these change occur regularly and can be predicted.

By examining events that precede a disease

rate increase or decrease, we may identify causes and appropriate actions to control or prevent further occurrence of the disease.

5

1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 6 0 0 1 9 6 1 9 6 5 1 9 7 1 9 7 5 1 9 8 1 9 8 5 1 9 9

Year Reported cases (per 1,000)

Vaccine Licensed

Measles by year of report, United States, 1960-1989

Secular ( long Secular ( long-

• term ) trends:

term ) trends: to predict or evaluate control program

6

Seasonality: Seasonality: to see seasonal pattern to see seasonal pattern

• ver years
• ver years

฀

2 4 6 8 1 0

F e b A p r J u n A u g O c t D e c

9 7 9 8 9 9

Cases of eosinophilic meningitis, Nan province, Thailand, 1997-1999

Cases Onset by Month

7

Day of w eek and tim e of day: Day of w eek and tim e of day: especially especially important for condition that related to important for condition that related to

• ccupational or environmental condition
• ccupational or environmental condition

5 1 0 1 5 2 0 2 5 3 0 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8

Deaths Time of Day

Fatalities associated with farm tractor injuries by time of day, Georgia, 1971-1981

8

Epidem ic period: Epidem ic period: to show the time course to show the time course

• f disease outbreak
• f disease outbreak

฀

10 20 30 40 19 21 23 25 29 1 3 5 7 9 11 13 15 17 19 21

Flood

Cases Date of onset

Num ber of leptospirosis patients by date of onset, Nov 2 5 -Dec 2 1 , 2 0 0 0

9

PLACE

We describe a health event by place

to gain insight into the geographical extent of the problem.

We may use place of residence,

birthplace, place of employment, school district, hospital unit, etc., depending on which may be related to the occurrence of the health event.

Myanmar

Umphang district

Nov.-Dec. Jan-Feb. Mar.

village Refugee camp Meningococcal cases, Meningococcal cases, Tak Tak province, province, 2 0 0 2 2 0 0 2 -

• 2 0 0 3

2 0 0 3

11

Thai-Myanmar border area in Umphang district

12

I n the refugee cam p

13

House of Karen people

• utside the cam p

14

PERSON

There are several person categories

available:

inherent characteristics: age, sex, race acquired characteristics: marital status,

immune

activities: occupation, use of

medication/ tobacco/ drugs

condition under which their live:

socioeconomic status, access to medical care

15

Sex specific attack rates of reported Sex specific attack rates of reported leptospirosis leptospirosis cases cases, , Songkla Songkla, , Nov Nov 25 25-

• Dec

Dec 15, 2000 15, 2000

5 9 .5 2 0

2 0 4 0 6 0 8 0 M ale F em a le S ex C a ses /100,0 00

16

Age Age specific specific attack rates of attack rates of reported reported leptospirosis leptospirosis cases cases, , Songkla Songkla, , Nov Nov 25 25-

• Dec

Dec 15, 15, 2000 2000

21.6 33.9 37.3 51.1 40.7 58.3 16.7 20 40 60 80 0-9 10-19 20-29 30-39 40-49 50-59 60+ Age group Cases /100,000

17

18

DEFINITION OF EPIDEMIOLOGY DEFINITION OF EPIDEMIOLOGY

"Epidemiolgy is the study of the distribution and determinants of health- related states or events in specified populations and the application of this study to the control of health problems" ( John M. Last, 1 9 8 8 )

19

Study design Study design

20

Why Why “ “ type of study type of study” ” ? ?

• Design more appropriate studies to get

right answers to the specific questions

• Getting more optional study design to

the questions

• Shape up proposals for getting more

valid results

TYPES OF OBSERVATIONAL STUDIES

Descriptive Studies

Little is known about the

• ccurrence, or

determinants of the disease

• Identify cases, estimate disease

frequency, examine time trend

• Justify additional (analytic)

studies to test hypotheses

Analytic Studies

Enough is known about the d disease, specific hypotheses can be tested

• Test specific etiologic hypotheses
• Justify additional studies
• Suggest potential for disease

prevention

22

STUDY DESIGN OF DESCRIPTIVE STUDY

23

STUDY DESIGN OF DESCRIPTIVE STUDY STUDY DESIGN OF DESCRIPTIVE STUDY

Individual level Case reports or case series Cross-sectional surveys of individual Population level Ecological study (Correlational study)

24

CASE REPORTS AND CASE SERIES CASE REPORTS AND CASE SERIES

• Case reports

Case reports : : describe experience of a single patients. Case reports document unusual medical occurrence and can represent the first clues in the identification

• f new diseases.
• Case series

Case series : : collections of individual case

• reports. Investigation of the activities of the

individual in case reports can lead to formulation of a hypothesis.

25

Streptococcus suis

26

CASE REPORTS AND CASE SERIES

• Advantage

Advantage

discover new diseases bring background information to form

hypothesis about risk factors

• Disadvantage

Disadvantage

might base on only one case do not have a comparison group

27

CROSS CROSS-

• SECTIONAL

SECTIONAL STUDY STUDY Total population Cases

28

Character of cross-sectional study

Collect data for all population; case &

healthy people

Information obtain is the number of total

cases among total population at that study period: prevalence

Can be either descriptive or analytic,

depend on design

29

CROSS-SECTIONAL STUDY

Descriptive Descriptive

Collected num ber of

cases and num ber

• f total population

Can assess only

prevalence of disease or other health events, also called prevalence study

Analytic Analytic

Exposure and

disease status are assessed sim ultaneously

Can determ ine

association betw een exposure and disease

30

Vaccine coverage among Thai children

Proportion of children received vaccine Education level of mother and vaccine status

in children

31

Cross Cross-

• sectional study

sectional study

Defined Population

Exposed Exposed

Have disease

Not exposed:

Have disease

Not Exposed:

Do not have disease

Exposed Exposed

Do not have disease

Gather Data on Exposure and Disease

32

Hypothetical illustration of the interrelationship between an occupational exposure and prevalence of disease

Job A Job A ( (hazardous hazardous) )

100 100 Workers Workers 80 well 20 ill

Job B (non-hazardous)

100 Workers 95 well 5 ill 80 well 10 10 ill ill 10 ill 95 well 15 15 ill ill

10 change jobs due to illness

Point X

Prevalence of job A = 20 % ( 20/100 ) Prevalence of job B = 5 % ( 5/100 )

Prevalence ratio Prevalence ratio = = 4 4

33

Hypothetical illustration of the interrelationship between an occupational exposure and prevalence of disease

Job A Job A ( ( hazardous hazardous) ) 1 0 0 1 0 0 W orkers W orkers 8 0 w ell 2 0 ill Job B ( non-hazardous) 1 0 0 W orkers 9 5 w ell 5 ill 8 0 w ell 1 0 1 0 ill ill 1 0 ill 9 5 w ell 1 5 1 5 ill ill

10 change jobs due to illness

Point Y Prevalence of job A = 1 1 % ( 1 0 / 9 0 ) Prevalence of job B = 1 4 % ( 1 5 / 1 1 0 ) Prevalence ratio Prevalence ratio = = 0 .8 0 .8

Advantage

Easy and less time consumed Estimate magnitude of problem especially for

rare disease

First step for exploring and forming

hypothesis

Able to use data for further case-control or

cohort during analytic study

Disadvantage

Difficult to confirm that risk factor come before

disease (temporal relationship)

Length biased sampling: Disease that has long

duration will over-represent the magnitude of illness while short duration will under- represent illness.

35

STUDY DESIGN OF DESCRIPTIVE STUDY STUDY DESIGN OF DESCRIPTIVE STUDY

Individual level Case reports or case series Cross-sectional surveys of individual Population level Ecological study (Correlational study)

36

ECOLOGICAL STUDY ECOLOGICAL STUDY

These studies are an extension of the use of

routine data.

Essentially the average exposure of the

population is plotted against the rate of the

• utcome for that population

This is done for several populations and the

data are then examined for evidence of an association between exposure and outcome.

Group of Proportion of Suicide rate Prussian Province Protestant ( per 1 0 0 ,0 0 0 pop.) A 9 .5 6 0 .3 0 B 1 6 .3 6 0 .4 5 C 2 2 .0 0 0 .7 9 D 2 6 .4 6 0 .9 5

Study to find relationship between suicide rate and proportion of Protestant

38

0.2 0.4 0.6 0.8 1 10 20 30 A B C D

• Prov. Proportion of Suicide

Protestant

rate A 9 .5 6 0 .3 0 B 1 6 .3 6 0 .4 5 C 2 2 .0 0 0 .7 9 D 2 6 .4 6 0 .9 5 Suicide rate (per 10 5/year) Proportion of Protestant (%)

Correlation betw een proportion of protestant and suicidal rate in 4 groups of Prussian provinces

Incidence Average temperature

15.3 24.3

.33

14.34

Correlation of JE incidence and average temperature in Dali prefecture, Yunnan province, China, 2001

40

Rationale for ecological study Rationale for ecological study

Low cost and convenience: make use of

existing data

Measurement limitations of individual-level

studies: environmental study

Interest in ecological effect: social norm,

population intervention

41

ECOLOGICAL STUDY ECOLOGICAL STUDY

• Advantage

Advantage : : cheap, quick and convenient since it usually come from existing data

• Disadvantage

Disadvantage : :

inability to link exposure with disease in

individual (ecological fallacy)

limit to control effect of other factors ฀

42

In summary

• Ecological study is a convenient, cheap and

simple study.

• Unit of study is the aggregate data not

individual level.

• It is usually be conducted as the first step

study for research.

• The result is difficult to interpret because of

confounding and bias.

Ecological or correlation study Ecological or correlation study

43

Study design Observational

Experim ental

Descriptive Case and serial case Cross-sectional Ecological Analytic

Case-control Cohort Quasi-experim ental Experim ental Cross-sectional Ecological

Summary of the study design Summary of the study design

Analytic study

Disease among exposed? Disease among non-exposed? Usually Usually prospective prospective Population at risk

Cohort study Cohort study

Exposed Not Exposed and

46

Distribution of illness according Distribution of illness according to exposure in a cohort study to exposure in a cohort study

Exposed Not exposed I ll Not ill a b c d a+ b c+ d Risk Risk a+ b c+ d a c

Relative risk = Risk Relative risk = Risk exposed

exposed / Risk

/ Risk not exposed

not exposed

47

Key features

Should have follow up period (even in

retro-cohort)

Have confine population Comparison group should be as similar as

possible

48

Cohort study Cohort study

Advantage

Advantage

Determine multiple effect of single

exposure

Able to estimate incidence Minimize bias for prospective study Disadvantage

Disadvantage

Expensive and time consumed Validity of result depend on follow up

Disease (Case) Exposure

?

No disease (Control)

? Case Case-

• control study

control study

50

Distribution of cases and controls according Distribution of cases and controls according to exposure in a case to exposure in a case -

• control study

control study

Exposed Not exposed Total Cases a c a+ c

Odds of exposure a/ c Odds of exposure a/ c

Controls b d b+ d

b/ d b/ d

Odds ratio = OddsE ( cases) / OddsE ( controls)

Odds

Team A plays 10 games:

8 wins 2 defeats

Risk (probability) of losing = 2/10 Odds of losing = 2/10 = 2/8

8/10

Probability of event Probability of non-event

Key features

Study only some part of the population,

especially among non-case

Case and control should come from the same

source population, to allow possibility that control could receive the exposure

Case and control are not necessary to be the

same

53

Advantage

Advantage

• Cheap, quick and efficient for rare

Cheap, quick and efficient for rare disease disease

• Potential role for testing multiple risk

Potential role for testing multiple risk factors factors

Disadvantage

Disadvantage

• Susceptible to bias

Susceptible to bias

• Inefficient for rare exposure

Inefficient for rare exposure

• Sometime difficult to establish

Sometime difficult to establish temporal relationship between exposure temporal relationship between exposure and disease and disease

Case Case-

• control study

control study

54

Advantage: Advantage:

• Feasible and practical especially for

Feasible and practical especially for ethical concern ethical concern

• Study population is more represent of

Study population is more represent of the target population the target population Lim itation: Lim itation:

• Difficult to replicate

Difficult to replicate

• Less control of extraneous factors that

Less control of extraneous factors that will lead to distortion of the result will lead to distortion of the result

• Less secure to make generalization

Less secure to make generalization

OBSERVATIONAL STUDY OBSERVATIONAL STUDY

55

Experimental or Experimental or intervention study intervention study

56

Experimental study Experimental study

• Put afford to control the situation of study

factors and observe effect of them

• Control extraneous factors by holding those

factors fix or “randomization” to make two groups have the same distribution of extraneous factors.

• One group of subject is given experimental

treatment and another group (control) is given either none or less preferable. After a period effect of exposure is measured and compared between two groups.

57

• Random ization

Random ization is used to allocate subject: subject get treatment by chance

• One important method to make the

result more valid is called “ “double double blinding blinding” ”: : neither investigator nor subjects know the treatment to which they have been allocated

Experimental study Experimental study

58

Lim itation: Lim itation:

• Effect is observed in artificial setting

because the controlling extraneous factors

• Randomization may not be ethical
• Apart from pharmaceutical intervention,
• ther intervention is difficult to do double

blinding

Experimental study Experimental study

59

Study design Observational Experim ental Descriptive Case and serial case Cross-sectional Ecological Analytic Case-control Cohort Quasi-experim ental Experim ental Cross-sectional Ecological

Summary of the study design Summary of the study design

Thank you, Khob Khun ka

61

ECOLOGI CAL STUDY

• Exploratory study

Exploratory study: : if there is no specific exposure of interest

• r the exposure of potential

interest is not measure

• Analytic study

Analytic study: : if the primary exposure variable is measured and included in the analysis

62

63

Main feature of each type of epidemiologic research

• Experimental :

Experimental :

• artificial

artificial m anipulation of study factor m anipulation of study factor with with random ization random ization

• Quasi

Quasi-

• experimental study :

experimental study :

• artificial manipulation of the study factor

artificial manipulation of the study factor w ithout random ization w ithout random ization

• Observational

Observational : :

• no

no artificial m anipulation artificial m anipulation of the study

• f the study

factor factor