Disease Screening

Screening

Session 12 [email protected]

27 March 2009

1

Introduction 

 

Screening refers to the application of a test to people who are as yet asymptomatic for the purpose of classifying them with respect to their likelihood of having a particular disease. The screening procedure itself does not diagnose illness. Those who test positive are sent on for further evaluation by a subsequent diagnostic test or procedure to determine whether they do in fact have the disease. 27 March 2009

2

Introduction 



An implicit assumption underlying the concept of screening is that early detection, before the development of symptoms, will lead to a more favourable prognosis because treatment began before the disease become clinically manifest will be more effective than later treatment. Screening has played an important role in public health over the years. 27 March 2009

3





There are often risks and costs associated with screening. Before a screening program is undertaken, there is need to evaluate the benefits against risks and cost involved.

27 March 2009

4

Diseases appropriate for screening    



Some diseases are not suitable candidates for the application of a screening program. To be appropriate for screening: The disease should be serious Treatment given before symptoms develop should be more beneficial in terms of reducing morbidity or mortality than that given after they develop The prevalence of the preclinical disease should be high among the populations screened. 27 March 2009

5

Diseases appropriate for screening 





The expenditure of resources on screening must be justifiable in terms of eliminating or ameliorating adverse health consequences. The consequences of failing to diagnose and treat early must be sufficiently grave to warrant undergoing the risks and discomforts of the screening procedure itself. Life threatening diseases such as breast cancer and those known to have serious and irreversible consequences if not treated early such as congenital hypothyroidism meet the criterion of seriousness. 27 March 2009

6



  



One question of importance is whether treatment of preclinical disease is more effective than treatment begun after the development of symptoms. To evaluate this, it is necessary to consider the natural history of a disease. There is a period after biologic onset during which the disease is asymptomatic but detectable by screening. The interval of time between the point at which the disease can be detected by screening and the point at which the individual becomes symptomatic and seeks medical attention has been termed the detectable preclinical phase(DPCP). For screening to be of benefit, treatment given during the detectable preclinical phase must result in a better prognosis than therapy given after symptoms develop.

27 March 2009

7







For example ,cancer of the uterine cervix develops slowly ,taking may be more than a decade for the cancer cells ,which are initially confined to the outer layer of the cervix ,to progress to a phase of invasiveness. During the pre-invasive stage, the cancer is usually asymptomatic but can be detected by screening using the papanicolaou smear. It is far preferable to begin treatment during this stage than when the cancer has become invasive. 27 March 2009

8







On the other hand, if early treatment makes no difference because the prognosis is equally good (or equally bad) whether treatment is begun before or after symptoms develop Then the application of a screening test will be neither necessary nor effective. For example Lung cancer has a very poor prognosis regardless of when treatment is initiated. 27 March 2009

9

Screening tests A

screening test should ideally be inexpensive  Easy to administer  And impose minimal discomfort on the patients.  Results of a screening test must be valid, reliable and reproducible. 27 March 2009

10

Validity  The

validity of a screening test is measured by its ability to accomplish what it is suppose to do.  That is, categorise persons who have preclinical disease as test-positive and those without preclinical disease as testnegative.

27 March 2009

11

Validity 



The table below summarises the relationship between the results of a screening test and the actual presence of disease as determined by the results of an appropriate subsequent diagnostic test. In this table

 a=no. of individuals screening test positive and the individual 





has the disease(True positives) b=screening test positive but the individual has no disease(False positive) C=screening test is negative but the individual do have the disease. (False negatives) d=Screening test negative and the individual does not have the disease (true negatives) 27 March 2009

12

Positive

Disease Status Negative Total

Result of Screening test

Positive

a

b

a+b

Negative

c

d

c+d

Total

a+c

b+d

27 March 2009

13

Definitions  False

Positive: Healthy person incorrectly receives a positive (diseased) test result.

 False

Negative: Diseased person incorrectly receives a negative (healthy) test result.

27 March 2009

14

Validity 







Sensitivity,Specificity,Positive Predictive Value (PPV) and Negative Predictive Value (NPV) are measures of the validity of a screening test. Sensitivity is defined as the probability of testing positive if the disease is truly present and is calculated as a/a+c As the sensitivity of a test increases ,the number of persons with the disease who are missed by being incorrectly classified as testnegative (false negatives) will decrease. Specificity is defined as the probability of screening negative if the disease is truly 27 March 2009and is calculated by d/(b+d) 15 absent

Validity: Sensitivity and Specificity 



Sensitivity ≡ probability a case will test positive (e.g., a test that is a 100% sensitive will have no false negatives) Specificity ≡ probability that noncases will test negative (e.g., a test that is a 100% specific will identify no false positives) 27 March 2009

Gold Standard

D+

D−

Total

T+

TP

FP

TP+FP

T−

FN

TN

FN+TN

Total

TP+FN

FP+TN

N

Test

SEN = TP / (those with disease) = TP / (TP + FN) SPEC = TN / (those free of disease) = TN / (TN + FP) 16

A

highly specific test will rarely be positive in the absence of disease and will therefore result in a lower proportion without disease who are incorrectly classified as test-positive (false positives).

27 March 2009

17

Positive and Negative Predictive Values (PPV and NPV)  Positive

predictive ,is the probability that a person actually has the disease given that he or she tests positive.  PPV= a/(a+b)  Negative predictive (NPV) is the probability that an individual is truly disease-free given a negative screening test  NPV= d/c+d 27 March 2009

18

Screen test (physical exam and mammography

Positive Negative Total

27 March 2009

Breast Cancer Cancer confirmed

Cancer not confirmed

Total

132 45 177

983 1115 63,650 63,695 64,633 64,810

19

 Sensitivity=

a/a+c=132/177=74.6%  Specificity=d/b+d=63,650/64,633=98.5%  PPV=a/a+b=132/1115=11.8%  NPV= d/c+d=63,650/63,695=99.9%

27 March 2009

20







When the prevalence of preclinical disease is low, the Positive Predictive Value will be low even when using a test with high sensitivity and specificity. The PPV of a screening test can thus be increased by either increasing the specificity of the test or prevalence of preclinical disease in the in the screened population. This can be achieved by targeting the screening program to groups of individuals at high risk of developing the disase of interest. 27 March 2009

21

Example: SEN & SPEC low prevalence population An HIV screening test is used in one million people with HIV Prevalence of 0.1% SEN and SPEC are 99%.Calculate PPV and NPV.

D+

D−

Total

T+

TP

FP

TP+FP

T−

FN

TN

FN+TN

Total

1000

999,000

1,000,000

Since Prevalence = (no. w/ disease) / N, then (no. w/ disease) = Prevalence × N For this illustration, (no. w/ disease) = Prevalence × N = 0.001× 1,000,000 = 1000 27 March 2009

22

Example: SEN & SPEC

low prevalence population (cont.)

D+

D−

Total

T+

TP

9,990

TP+9,990

T−

FP

989,010

FP+989,010

Total

1000

999,000

1,000,000

Since SPEC = TN / (those w/out disease), then TN = SPEC × (those w/out disease). For this example, TN = 0.99 × 999,000 = 989,010 27 March 2009

23

Example: SEN & SPEC

low prevalence population (cont.)

D+

D−

Total

T+

990

9,990

10,980

T−

10

989,010

989,020

Total

1000

999,000

1,000,000

Since SEN = TP / (no. with disease), then TP = SEN × (no. with disease) For this example: TP = SEN × (no. with disease) = 0.99 × 1000 = 990 27 March 2009

24

Example: PPV, NPV low prevalence population

D+

D−

Total

T+

990

9,990

10,980

T−

10

989,010

989,020

Total

1000

999,000

1,000,000

Prevalence = 1000 / 1,000,000 = 0.001 (or 0.1%) PPV = TP / (those w/ positive test) = 990 / 10,980 = 0.090 NPV = TN / (those w/ negative test) = 989010 / 989020 = 0.99 27 March 2009

25

Example 2 : SEN & SPEC high prevalence population

An HIV screening test is used in one million people. Prevalence in population is now 10%. SEN and SPEC are again 99%. D+

D−

Total

T+

99,000

9,000

108,000

T−

1,000

891,000

892,000

Total

100,000

900,000

1,000,000

Prevalence = 100000 / 1,000,000 = 0.10 = 10% SEN = 99000 / 100,000 = 0.99 SPEC = 891,000 / 900,000 = 0.99 27 March 2009

26

Example: PPV, NPV high prevalence population

D+

D−

Total

T+

99,000

9,000

108,000

T−

1,000

891,000

892,000

Total

100,000

900,000

1,000,000

Prevalence = 100000 / 1,000,000 = 0.10 = 10% PPV = 99,000 / 108,000 = 0.92 (better PPV in high prev pop.) NPV = 891,000 / 892,000 = 0.999 27 March 2009

27

Cutoff Points for Positive Tests  Sensitivity

and specificity are influenced by the cutoff point used to determine results  e.g., HIV screening test detects color change (optical density ratio)  At what point do we say the color change is sufficient for a positive test? 27 March 2009

28

Low Cutoff A low cutoff point will have few false negatives (high sensitivity) but many false positive (low specificity)

27 March 2009

29

In-Between Cutoff An intermediate cutoff point will have some false negatives (moderate sensitivity) & some false positive (moderate specificity)

27 March 2009

30

High Cutoff A high cutoff point will have many false negatives (low sensitivity) and few false positive (high specificity)

27 March 2009

31

Evaluation of screening programs  Evaluation

of a potential screening program involves consideration of two issues:  Is the program Feasible  Is the program Effective

27 March 2009

32





No matter how effective a screening procedure is in reducing subsequent morbidity and mortality, it will not be accepted if it can not be conducted efficiently, with minimal inconvenience and discomfort and at a reasonable cost. Conversely, no matter how cost-effective the screening program can be, it will not be warranted if it does not reduce morbidity and mortality. 27 March 2009

33

Feasibility  Determined

by a number of factors related to program performance.  Measure the acceptability of the program to the screenees,  cost effectiveness,  subsequent diagnosis and treatment of subjects who test positive and the yields of cases. 27 March 2009

34

 The

screening program should be acceptable to the population being screened  Means that it must be quickly and easily administered with minimum discomfort.  Cost must also be considered.

27 March 2009

35







With respect to yield, or number of cases detected by a screening program, one measure that is commonly considered is the predictive value of a screening test. Predictive value measures whether or not an individual actually has the disease, given the results of the screening test. The low the prevalence of the disease, the lower is the positive predictive value. 27 March 2009

36

Effectiveness  There

is need to determine how effective the screening programme.  Is it effective in reducing morbidity and mortality from the disease.

27 March 2009

37