Appendix 2: Screening:

Screening has been defined as the presumptive identification of unrecognised disease or defect by the application of tests or examinations. It involves the performance of a relatively simple test on a large population to separate from it a subpopulation with a high likelihood of having a treatable disease.

The operational objective of screening is the accurate classification of persons who do or do not have the disease in question. The goal or outcome objective of screening is the reduction of the consequences of the disease by incidence, morbidity or mortality.

Test characteristics to consider:

The following test characteristics should be considered when selecting appropriate tests for medical surveillance (Cochrane and Holland criteria):

• Simple - both in application and in interpretation.
• Acceptable - the methods employed should cause minimum discomfort, and avoid clashes with belief systems.
• Cost-effective - the result should be worth the cost.
• Precise (repeatable) - the result should be reliable, so that repeat testing of the same sample will produce the same result.
• Accurate - the results should be a true reflection of the condition being sought.
• Sensitive - the ability of the test to identify the condition, even if it means including a lot of false positives.
• Specific - the ability of the test to exclude conditions that are the false positives.

The positive predictive value of a test is a combination of sensitivity and specificity; it’s a test’s ability to accurately identify what it is looking for (deafness, blindness, etc.); this means without identifying cases incorrectly (false positives), and also without missing any (false negatives). It can be calculated, by multiplying the sensitivity by the specificity. In general, although not an absolute rule, tests with a high degree of sensitivity are best for screening, and tests with a high degree of specificity are best for confirmatory diagnosis.

However, note that some tests have high specificity but poor sensitivity, which means that, whilst the cases identified may be reliably regarded as positive, these tests may miss a significant proportion of positive cases, due to the poor sensitivity. These are therefore only useful when positive - when the result is negative, one cannot be certain that this is true. A good example is the (cell diameter test for alcohol consumption. When it is positive, it is generally regarded as a reliable marker of excessive alcohol consumption (high specificity). However, a significant proportion of cases are reported by the test as normal, even when circumstantial evidence indicates the contrary (poor sensitivity). To continue with the same example,the enzyme gamma glutamyl transferase (GGT) is often used as a screening test for alcohol consumption, because it is fairly sensitive. However, it must be remembered that its specificity is poor: there are many conditions that give rise to a raised GGT. To overcome this, tests are combined in a battery so that the combination improves the overall positive predictive value of the examination. Hence the combination of a good history, collateral evidence, GGT, MCV and CDT all contribute to improve the combined positive predictive value in establishing the presence of excessive alcohol consumption.

What Makes Screening A Viable Option?

The Disease

The diseases being screened for must have serious health consequences in terms of morbidity and mortality.

The natural history of the disease needs to be reasonably well understood.

The disease must be treated more successfully in the screen-detected stage than when symptoms have led to the diagnosis. There is no point in screening for a disease that can be successfully treated after symptoms appear, nor is there any point in screening for an untreatable disease.

The detectable pre-clinical phase of the disease should have a high prevalence among the people screened. Otherwise too few cases will be detected to justify the expense of screening.

The Test

The screening test should designate people with early disease as positive and those without as negative.

Success of a screening programme is at meeting these expectations with sensitivity and specificity. In designing a screening programme one often has to trade between sensitivity and specificity. Sensitivity should necessarily be maximised. One needs to balance the consequences of low sensitivity (missed cases or false negatives) and low specificity (economic, social and psychic costs of false positives).

A useful indicator of the capabilities of the screening test is the positive and negative predictive values.

The test must be acceptable to both the employee and tester. The unpleasantness of the test should be as minimal as possible. and not have potential hazardous consequences. These characteristics assist in compliance with the programme.

The test must be cost-effective.

Sensitivity = a/a+c

Specificity = d/b+d

Positive Predictive Value = a/a+b

Negative Predictive Value = d/d+d

Evaluation

Proper choice of outcome measure. For cancer and other chronic conditions the objective should be to reduce mortality.

Problems of comparing survival times of screen detected cases against symptom-detected cases: almost invariably, screen detected cases have a longer survival time. Is this due to the effect of screening or the following biases?

• Lead time bias: screening has advanced the diagnosis of the disease.
• Selection bias: screening may tend to select cases destined to have a benign clinical course. One will detect those diseases with the longest detectable pre-clinical phase. These lesions tend to also have long clinical phases.

Thus people with relatively benign lesions and correspondingly good survival tend to be over-represented among cases detected by screening.

The only valid outcome variable for assessment of the results of a screening programme is the mortality rate from the disease in the total population offered screening in comparison with the mortality rate that would be expected in the same population if screening had not been offered.

Screening should only be done if there is conclusive evidence that the natural history of the disease will be altered in a significant proportion of people screened.

Monitoring

Monitoring is the repetitive performance of an observation or measurement used to detect an unfavourable trend, which may be altered by appropriate intervention.

Screening by definition is cross-sectional and monitoring longitudinal.