1 Definition

The purpose of population screening is to divide eligible subjects into two groups: those with a low risk of having cancer and those with a sufficiently high risk to warrant further diagnostic examination. Screening is applied to asymptomatic individuals who meet the eligibility requirements of the screen. The idea is to identify individuals whose disease is in an earlier stage so that it can be more effectively treated.

2 Purpose

As indicated in the definition, screening tests are applied to populations of asymptomatic individuals and represent a public health strategy. The object, therefore, is to improve the health of the population and the value of cancer screening is measured by its affect on the population.

3 Attributes

An ideal cancer screening program will have the following attributes: 

• The cancer will be a major health problem (serious and common)    
• The cancer will be more treatable if detected early    
• The test should be acceptable to those eligible    
• The test should be inexpensive    
• The test should have high sensitivity (the bulk of subjects with the cancer should test positive)    
• The test should have high specificity (the vast majority of subjects without the cancer should test negative)    
• Screening will have been shown to reduce mortality in randomized controlled trials    
• Screening will have been shown to be a cost-effective means of controlling this cancer

Any screening test applied in practice has most of the attributes listed above although many of them are fairly expensive and complex and this is justified by the seriousness of the disease. Most cancers can potentially develop over many decades of life so that a cancer screening test is usually repeatedly administered as long as subjects remain eligible.

4 Impact 

Many of the characteristics of screening can be illustrated by consideration of the following table which cross-tabulates the results of a single screen versus the true state for a hypothetical group of screened persons:

Thus:

• A= number of persons who do not have cancer and test negative by the screen, these will be the majority of people screened since cancer is a relatively rare disease, these subjects will be minimally affected by undergoing screening (e.g. inconvenience of procedure, personal costs),    
• B= number of persons who do not have cancer but screen positive, this will be the second largest group of subjects, all will be adversely affected by the screen (e.g. anxiety associated with positive test, follow-up diagnostic investigations)    
• C= number of persons who do have cancer but screen negative, this should be the smallest group of subjects, they do not benefit from being screened and are potentially adversely affected if future signs and symptoms of cancer are not promptly investigated because of the false reassurance of the screen result    
• D= number of persons who have cancer which is detected by the screen, this group contains the main beneficiaries of screening since they have the potential to have their cancer treated earlier and more effectively

Basic measures are defined as follows:

False Negative Rate = 1-Sensitivity = Proportion of cancers missed by the screen

False Positive Rate = 1-Specificity = Proportion of cancer free subjects testing positive by the screen

The False negative rate is almost impossible to calculate under field conditions. Persons who test negative by the screen are not investigated further and whether they are truly free of cancer is not known with certainty. Estimates are usually made by assuming persons developing disease within a time interval after a negative screen (2-5 years typically) had disease present at the screen and are thus false negatives.

The False positive rate is more straightforward to calculate since the persons screening positive are usually thoroughly investigated so that the number of false positives are recorded.

Another commonly used measure is the Positive Predictive Value (PPV) where

5 Screening - What, Who and When 

What - Generally there are not multiple different screening technologies to choose from for screening for a single cancer although exceptions exist. More often, the question usually centres on how a procedure will be done (e.g. what views will be used in mammographic screening, how will cervical smears be taken and prepared). Decisions about these questions are based on accuracy, ease, cost, etc.

Who - Eligibility for cancer screening is based on age, sex and history/symptoms of the disease. History and symptoms are used to exclude subjects because their risk of disease is much greater and more definitive follow-up procedures are indicated. Age and sex are used since they determine the likelihood of cancer and are often related to the performance of the test in order that those being screened have a reasonable probability that they will benefit from being screened.

When   - The screening test should be repeated at intervals determined by the natural history of the cancer. Cancer develops throughout life and the screen should be used frequently enough to provide a reasonable expectation that the test will be applied at least once while the cancer is in a pre-clinical phase. The specification of the optimum frequency for screening represents a judgment regarding the benefits of more frequent screening (e.g. increased disease detection, better outcomes) versus the costs (e.g. greater numbers of false positive tests, increased medical costs).

The progressive nature of cancer would suggest that screening longer (starting earlier in life and continuing later) and more frequently will result in better disease specific outcomes when an effective screen is available. It is also likely that more screens will result in greater numbers who are adversely affected by screening (e.g. larger numbers of false positives, increased diagnostic tests) and those who are adversely affected can come to form the largest group in the population. The determination of screening recommendations will therefore always represent a balance between the individual and societal benefits and costs. 

6 Follow-up

Performance of a screening test will have no effect on disease unless individuals with abnormal results are adequately investigated and treated. Experience in several countries has shown that the value of screening can be "lost" if follow-up of positive screenees is not effective. Therefore screening programs must include  

• Recommendations for follow-up of abnormal results    
• Monitoring and tracking systems to measure follow-up    
• Routine publication of results of the effectiveness (field performance) of screening

8 Recommendations for the Clinician in using Screening Tests

Questions to be answered: 

1. Is there evidence from randomized trials supporting the use of the test in a screening scenario?    
2. Is the potential screenee eligible by relevant guidelines (e.g. asymptomatic, age sex etc.)?    
3. Should the subject be screened now?    
4. Are there guidelines available for the disposition of screen positives? 

9 References

1. Evaluation of Cancer Screening, Eds., J. Chamberlain and S. Moss, Springer Verlag, London, 1996.    
2. Advances in Cancer Screening, Ed. A. B. Miller, Kluwer Academic, Boston, 1996.