• Overview
• Definition
• Health consequences
• Costs
• Contributors

Health Consequences

Harms and benefits

Overview

This section covers the possible negative consequences of test results and treatment. See Box 1 for a list of the potential harms associated with screening programmes. For example, individuals who receive a false negative result will have false reassurance (and possibly delayed presentation of symptomatic disease later).

Possible anxieties for all those with positive test results are usually temporary, if reassurance follows quickly. This is one reason why the diagnostic process should follow promptly from the screening result. ^{1 2} See Box 2 for example of harms and benefits associated with a bowel cancer screening programme.

The benefits must be weighed against all the above harms. Most people assume that to have any disease detected early must always lead to an improved outcome. ³ This is true for conditions that only present clinically when it is too late to cure or treat effectively, such as phenylketonuria or congenital hypothyroidism in the newborn, or diabetic retinopathy. In most diseases, where there is an early clinical stage, which may also be curable, there are four possible outcomes of screening:

Good outcome/prognosis: screening made a difference/saved life;

Good outcome/prognosis: but screening made no difference, progression was sufficiently slow that disease could have been treated successfully at clinical presentation;

Poor outcome/prognosis: screening made no difference; and

Apparently successful but unnecessary treatment of insignificant disease.

The success of a screening programme will depend upon the balance between these four possible outcomes.

False positives

No screening test is foolproof because of the limitations of the test itself and human error. In any screening programme, there is an irreducible minimum of false-positive results (wrongly reported as having the condition) and false-negative results (wrongly reported as not having the condition). ¹ See Table 1.

Where sensitivity = A/A+C and specificity = D/B+D.

Sensitivity and specificity

It is generally assumed that all the true positives will benefit from early detection and the true negatives from reassurance. This is so in newborn screening for phenylketonuria and congenital hypothyroidism, where tests are sensitive and specific, and clinical diagnosis is too late to prevent permanent damage to the baby. However, it is too simplistic in most other screening programmes.

For those with true negative screening results, the disease could still be developing and become symptomatic within a short time or an infectious disease could be contracted. Hence, it is important to consider that the screening test is only a snapshot in time.

Interpretation of a screening test result requires clinical opinion or visual perception. There is often a grey area of mild or borderline disease. This can lead to repeat testing or can result in unnecessary investigations, and considerable anxiety.

Predictive value

The positive predictive value of a test is the probability that an individual with a positive result actually has the condition. Likewise the negative predictive value of a test is the probability that an individual who is screened negative does not have the condition being tested for.

So from table 1: Positive predictive value = A/A+B

Negative predictive value = D/C+D

Predictive values are influenced by the prevalence of the condition in the population, and hence the pre-test risk in the person being tested. The same test will produce very much better predictive values where there is high prevalence, than where the disease is very rare. Hence a positive newborn screening test result for sickle cell disease in multiethnic London is much more likely to have identified an affected baby than the same result in Sussex, where there are few families who originate from areas where sickle cell disease is common.

Investigations and diagnosis

Investigations and diagnosis may also be inconclusive for some. Individuals who have been investigated and given the all-clear may go on to develop the condition later in life.

Screen-detected disease

Screening tests are designed to pick up early stages of disease. They therefore have a tendency to also pick up disease which is of no clinical significance, either because it is likely to regress if left alone, or would normally develop so slowly that it might not become symptomatic or life-threatening during the person’s lifetime. ³

Therefore, Table 1, could be further divided because not all the disease detected may be clinically significant or ever become life-threatening. ⁶ See Table 2. It would be better if those with inconsequential screen-detected disease had not been diagnosed, although at the time it may not be obvious who they are. It is already too late - they have been labelled as sick. The inconsequential false negatives are better-off - at least until they attend for screening again, when they may be less fortunate, as ‘disease’ may be picked up.

These harms make informed choice essential before anyone enters a screening programme. ⁷ They also explain why it is essential to have a comprehensive quality assurance system, ⁸ so that potential benefits are optimised and the likelihood of harm reduced to the bare minimum.

Antenatal screening

Antenatal screening tests for conditions such as Down’s syndrome may lead to further investigations or interventions, which pose risks for the foetus. The risk of miscarriage following amniocentesis or chrorion villus sampling is about 1%. This may be acceptable if there is a high risk that the foetus is affected and termination would then be considered. However, it could result in the loss of many healthy babies if the initial screen did not have a high predictive value.

Bias

The aim of cancer screening is to reduce mortality from that cancer, and not just to detect cases. Even extending survival is misleading, because the clock is started earlier in the disease pathway. This is called lead-time bias. Screening is also very good at picking up slowly progressing disease, whereas aggressive, rapidly growing tumours will tend to present with symptoms in between routine screens. This is because the window of opportunity to detect them by screening is shorter. Screen-detected and symptomatic cases will therefore not have the same prognosis. This is called length-time bias.