Population Health Assessment

Population Health Assessment

The Use of 2 by 2 Tables

Healthy

UH

Healthy UnHealthy

HOW TO SEPARATE THE TWO SUBPOPULATIONS?

Health AssessmentHealth Assessment

Health AssessmentHealth Assessment

SURVEYS

SURVEILLANCE

ON-GOING EPIDEMIOLOGIC INTELLIGENCE

GATHERINGUSUALLY LONG TERM

SYSTEMATIC and ORGANIZED

Health AssessmentHealth Assessment

Unhealthy

HEALTHY

H

UH

POPULATION SAMPLE

SAMPLING

MUST BE: 1. UNBIASED

2. REPRESENTATIVE

THIS WILL BE THE ONLY WAY TO MAKE INFERENCES

SAMPLING

NON RANDOM.....CONVENIENCE

RANDOM

1. SIMPLE RANDOM

2. STRATIFIED RANDOM

3. STAGE BY STAGE RANDOM

4. SYSTEMATIC RANDOM

SAMPLING

NON RANDOM.....CONVENIENCERANDOM

#1: SAMPLING PROCEDURE

#2: SAMPLE SIZE Determination

WITH THESE TWO CONSIDERATIONS...ONE CAN PROCEED

TO ASSESS THE HEALTH OF A POPULATION UNDER STUDY

Need to Have the Knowledge Base and Skills in:

Healthy

UH

Healthy UnHealthy

HOW TO SEPARATE THE TWO SUBPOPULATIONS?

Health AssessmentHealth Assessment

The Use of TEST SYSTEMS

TEST SYSTEMS

CLINICAL

IMMUNO-DIAGNOSTIC

PHYSIO-CHEMICAL

OTHERS

ONE COMMON CHARACTERISTICS OF TEST SYSTEMS ?

FALLACIES AND INACCURACIES

Health AssessmentHealth Assessment

TEST SYSTEMS

FALLACIES AND INACCURACIESTEST POSITIVES

TEST NEGATIVES

WHAT DOES IT MEAN ? COULD MEAN ?

TEST SYSTEMS

FALLACIES AND INACCURACIES

TEST POSITIVES

TEST NEGATIVES

WHAT DOES IT MEAN ? COULD MEAN ?

HOW TO ASSESS OR ESTIMATE INACCURACIES.....

TEST SYSTEMS

TEST SENSITIVITY AND TEST SPECIFICITY

AND OTHER MEASURES

Se

Sp

F+

F-

PREDICTIVE VALUES (+ or -)

DIAGNOSABILITY

TEST SYSTEMS

TEST

SYSTEM

+

-

TOTALS

Refer to Exercise

Test Negatives

Test Positives

The 2 By 2 Table: How to Create it??

How to Create a 2 By 2 Table???

Why do we need it? Where do we Get the data to insert in these Tables??

TEST SYSTEMS

TEST

SYSTEM

+

-

SUB-TOTALS

SUB-TOTALS

D+ D-

STATE OF NATURE

Note: How to Cross Classify to Create the 2 By 2 Table.

TEST SYSTEMS

TEST

SYSTEM

+

-

SUB-TOTALS

SUB-TOTALS

D+ D-

STATE OF NATURE

a b

c d

a+b

c+d

a+c b+d n=a+b+c+d

Note: How to Cross Classify to Create the 2 By 2 Table

TEST SYSTEMS

TEST

SYSTEM

+

-

SUB-TOTALS

SUB-TOTALS

D+ D-

STATE OF NATURE (Gold Standard)

a b

c d

a+b

c+d

a+c b+d n=a+b+c+d

Se= P(T+/D+)= a/(a+c) Sp= P(T-/D-)= d/(b+d)

TEST SYSTEMS

TEST

SYSTEM

+

-

SUB-TOTALS

SUB-TOTALS

D+ D-

STATE OF NATURE (Gold Standard)

a b

c d

a+b

c+d

a+c b+d n=a+b+c+d

Se= P(T+/D+)= a/(a+c) Sp= P(T-/D-)= d/(b+d)

F+= P(T+/D-)= b/(b+d) F-=P(T-/D+)=c/(a+c)

TEST SYSTEMS

TEST

SYSTEM

+

-

SUB-TOTALS

SUB-TOTALS

D+ D-

STATE OF NATURE (Gold Standard)

a b

c d

a+b

c+d

a+c b+d n=a+b+c+d

Predictive value(+)=a/(a+b)

Predictive value(-)=d/(c+d)

True Prevalence=(a+c)/n

Apparent Prevalence=(a+b)/n

TEST SYSTEMS

TEST

SYSTEM

+

-

SUB-TOTALS

SUB-TOTALS

D+ D-

STATE OF NATURE

a b

c d

a+b

c+d

a+c b+d n=a+b+c+d

DIAGNOSABILITY= P(D+/T+)=(P(D+)*P(T+/D+))/(P(D+)*P(T+/D+))+((P(D-)*P(T+/D-))

LINK

EXERCISE: CASE PROBLEM #1

TEST SYSTEMS EXERCISE          Given

True prevalence 38%True prevalenceIHA test positives 250

IHA test positives

IHA positives 260IHA positives Given

True prevalence 38%Infected/diseased 250IHA test positives 260IHA test negatives 478

DATA GIVEN

IHA test - 478 65%IHA test + 260 35%

738 100%

EXERCISE: CASE PROBLEM #1

  D+ D- Marginal totalsT+ a b a+b

T- c d c+d

Marginal totals a+c b+d n

True Prev (a+c)/n 280Apparent Hlthy (c+d)/n 478

EXERCISE: CASE PROBLEM #1

TEST SYSTEMS EXERCISE

Option #1 D+

D- Marginal totals

T+ 250 10 260

T- 30 448 478Marginal totals 280 458 738

EXERCISE: CASE PROBLEM #1TEST SYSTEMS EXERCISE

Option #1 D+

D- Marginal totals

T+ 250 10 260

T- 30 448 478

Marginal totals 280 458 738

Se P(T+/D+) a/(a+c) 89.29%

Sp P(T-/D-) d/(b+d) 97.82%

F+ P(T+/D-) b/(b+d) 2.18%

F- P(T-/D+) c/(a+c) 10.71%

EXERCISE: CASE PROBLEM #1

TEST SYSTEMS EXERCISE

D+

D- Marginal totals

T+ 250 10 260T- 30 448 478Marginal totals 280 458 738

Se P(T+/D+) a/(a+c) 89.29%Sp P(T-/D-) d/(b+d) 97.82%

F+ P(T+/D-) b/(b+d) 2.18%F- P(T-/D+) c/(a+c) 10.71%

EXERCISE: CASE PROBLEM #1Note How Data in the Table is Presented

TEST SYSTEMS EXERCISEOption #1

D+

D-

Marginal totals

T+ 250 10 260T- 30 448 478Marginal totals 280 458 738

Se P(T+/D+) a/(a+c) 89.29%Sp P(T-/D-) d/(b+d) 97.82%

F+ P(T+/D-) b/(b+d) 2.18%F- P(T-/D+) c/(a+c) 10.71%

TEST SYSTEMS EXERCISEOption #2

T+

T-

Marginal totals

D+ 250 30 280D- 10 448 458Marginal totals 260 478 738

Se P(T+/D+) 89.29%Sp P(T-/D-) 97.82%

F+ P(T+/D-) 2.18%F- P(T-/D+) 10.71%

EXERCISE: CASE PROBLEM #1Note How Data in the Table is

PresentedTEST SYSTEMS EXERCISE

Option #3

D-

D+

Marginal totals

T+ 10 250 260T- 448 30 478Marginal totals 458 280 738

Se P(T+/D+) 89.29%Sp P(T-/D-) 97.82%

F+ P(T+/D-) 2.18%F- P(T-/D+) 10.71%

TEST SYSTEMS EXERCISEOption #4

T-

T+

Marginal totals

D+ 30 250 280D- 448 10 458Marginal totals 478 260 738

Se P(T+/D+) 89.29%Sp P(T-/D-) 97.82%

F+ P(T+/D-) 2.18%F- P(T-/D+) 10.71%

Bonus QuestionBonus Question

Discuss Diagnosability as a Decision Making Exercise

Good LuckGood Luck

Refer to Online Epidemiology Course @:

http://www.onemedicine.tuskegee.edu