Screening for disease
DR.K.ARULANANDEM
LECTURER/COORDINATOR
Application of principles and methods of epidemiology
- Prevent diseases - Improve the outcome of illness Screening
Screening
Screening is the early detection of disease , precursors to disease or susceptibility to disease in individuals who do not show any signs of disease.
In ordinary medical practice ,the patient initiates an encounter because of a troubling symptom.
The physician does his best to help but is not responsible if the symptom turns out to represent something beyond the ability of current medical practice to cure
A screening test is usually initiated by physician or indirectly by professional or advocacy groups done on apparently well.
In this situation, there is an implied promise that it will do more good than harm.
This adds an ethical dimension to screening.
In screening ,test or procedures are applied to asymptomatic people for the purpose of dividing them in to two groups
-Those who Probably have a disease or condition -Those who Probably do not have a disease or condition
A screening test is not intended to be diagnostic Person with a positive findings must be referred for diagnosis
Why screen?
To reduce morbidity and mortality Therefore, early diagnosis alone does not justify a screening program Early diagnosis and treating early must lead to a measurable improvement in outcome
Certain conditions needs to be fulfilled to implement a screening programme. -Disease should be appropriate -a suitable screening test should be available - should be feasible to implement as a programme - should be able to conduct the programme effectively
Disease should be appropriate
Should be serious-cost/discomfort should be worthwhile E.g.: congenital hypothyroidism vs. gallstones
Treatment given before symptoms develop should be beneficial in reducing morbidity/mortality than that given after they develop. Prevalence of preclinical disease should be high among the population screened.
Disease should be appropriate
Treatment given before symptoms develop should be beneficial in reducing morbidity/mortality than that given after they develop.
Evaluated by looking at the natural history of disease.
Natural history
A B C D
susceptibility
Sub clinical disease
Clinical disease
Recovery/disability /death
A biological onset B disease detected by screening C symptoms develop D - usual time of diagnosis B to Cdetectable preclinical phase
Treatment during detectable preclinical phase must result in better prognosis than treatment after symptoms develop
E.g. cervical cancer vs. lung cancer
Disease should be appropriate
Prevalence of preclinical disease should be high among the population screened
-cost of the program relative to the number of cases detected -can be increased by screening high risk groups breast cancer among family history
Hypertension meets all criteria
Sometimes all criteria not met-PKU
Certain conditions needs to be fulfilled to implement a screening programme -Disease should be appropriate - a suitable screening test should be available - should be feasible to implement as a programme - should be able to conduct the programme effectively
A suitable screening test should be available
Inexpensive Easy to administer Impose minimal discomfort Results-valid,reliable,reproducible
Valid of screening test
-Ability to do what is supposed to do Correctly categorize persons who have preclinical disease as test positive and those without preclinical disease as test negative
Outcomes of a Screening Test
True Disease Status Screening Test Positive
True Positives (a)
False Negatives (c)
a+c
Negative
False Positives (b)
True Negatives (d)
b+d
Total
a+b
Positive
Negative
c+d
Total
a+b+c+d
Measures of validity
Sensitivity
Specificity
Measures of validity
Sensitivity-probability of testing positive if the disease is truly present a/(a+c) A high sensitive test rarely incorrectly classify persons with the disease as negative
Sensitivity increase-false negatives decreases
Measures of validity
Specificity-probability of screening negative if the disease is truly absent d/(b+d) A high specific test will rarely be positive in the absence of the disease
Specificity increase-false positives decreases
Measures of validity
Ideally -high sensitive -highly specific
Cut off between normal and abnormal -tradeoff between sensitivity & specificity
Altering the criterion Decision based on consequences of leaving cases undetected against erroneously classifying healthy person as diseased
Sensitivity should be increased at the expense of specificity -when it is crime to miss a case -definite treatment available -disease can spread -subsequent diagnosis -easy, less costly, not risky
Specificity should be sensitivity
-cost and risk of diagnosis high -when it is dangerous to give a false sense of security
Strategies to increase sensitivity /specificity
Use results if several tests together -in parallel -in series In parallel -same time if any positive - a positive results e.g.DVT Increase sensitivity In series - if one positive administer another if both positive - a positive results eg:syphilis increase specificity
Reliability
Consistency of results in repeated measurements on same persons under same conditions -biologic variation in the actual measurement -method-instrument -observer-intra/inter
Evaluation of screening programme
Even is disease is appropriate and a valid test is available Should we implement as a widespread programme? - feasibility - efficacy
Feasibility
-Acceptability of the programme
Quick,easy,minimal discomfort
- Cost effectiveness -Availability of subsequent diagnosis and treatment -Yield of cases
Yield
Number of cases detected by a screening programme -predictive value of the test
Whether or not an individual actually has the disease, given the results if a screening test
Outcomes of a Screening Test
True Disease Status Screening Test Positive
True Positives (a)
False Negatives (c)
a+c
Negative
False Positives (b)
True Negatives (d)
b+d
Total
a+b
Positive
Negative
c+d
Total
a+b+c+d
Positive predictive value
Probability that a person actually has the disease given that he/she tests positive
a/a+b True positives/all test positives
Negative predictive value
Probability that a person who tested negative not to have the disease
d/c+d True negatives/all test negatives
High if the disease is rare
Predictive values do not depend only on the sensitivity/specificity
But also on prevalence of the preclinical disease
More sensitive-less chances to false negatives
High negative predictive value
More specific-less false positives
High positive predictive value But if the disease is rare irrespective of the specificity-the positives are mostly false positive
To increase Yield -increase prevalence? Screen high risk group
screening
Done on apparently healthy people
Diagnostic test
Done on those who show signs of disease
Less expensive Less accurate Applied to groups Based on one criterion or cut-off point
More expensive More accurate Applied to single person Based on symptoms,signs,laboratory investigations,etc
Initiative comes from the investigator
Initiative comes from the patient
Test results are arbitrary and final
Diagnosis is not final but modified in light of new evidence, diagnosis is the sum of all evidence
Not a basis for treatment
Used as basis for treatment
THANK YOU
