Number Needed to Screen

The number of individuals who would need screening for a condition to prevent one adverse outcome, based on prevalence and programme effectiveness.

Source: Rembold 1998

The number needed to screen is how many people must be screened for a condition to prevent one adverse outcome, given how common the condition is and how effective the programme is. It tells a screening programme the scale of effort required for each benefit gained, which helps judge whether the resources, and the harms of screening the many who will not benefit, are justified. A large number warns that great effort buys little, a small one that screening is efficient. Weighing effort against benefit is its use. Gordis (2014) describes this measure.

Source: Gordis 2014

The number needed to screen is calculated from the absolute risk reduction in the adverse outcome achieved by the screening programme, taking its reciprocal, where that risk reduction reflects the prevalence of the condition, the effectiveness of the screening test in detecting it, and the effectiveness of treatment for those detected. So the number needed to screen is calculated as one divided by the absolute reduction in the adverse outcome attributable to screening the population, which incorporates how common the condition is and how well screening and treatment work, translating these into the number who must be screened for one prevented outcome.

Source: Rembold 1998

The number needed to screen relates to the number needed to treat as its extension to whole-population screening: the number needed to treat concerns patients receiving a treatment, whereas the number needed to screen concerns people undergoing screening, and it is typically much larger because it includes the many screened who do not have the condition. So the number needed to screen builds on the same reciprocal-of-risk-reduction logic as the number needed to treat but applies it to the screening of a population, reflecting that screening involves testing many to benefit few, which makes the numbers involved considerably larger.

Source: Rembold 1998

The number needed to screen is useful because it expresses the yield of a screening programme in an interpretable way, showing how many people must be screened to prevent one adverse outcome, which aids judgements about a programme's value and the balance of benefits against harms and costs. So the number needed to screen is useful for evaluating and communicating the efficiency of screening, since a very large number needed to screen may indicate that a programme prevents few outcomes relative to the effort and potential harms of screening many people, informing decisions about whether and how to screen.

Source: Rembold 1998

The limitations of the number needed to screen include its dependence on prevalence, screening and treatment effectiveness, and the time horizon, so it varies between populations and periods; that it captures only the prevented adverse outcome and not the harms of screening, such as false positives and overdiagnosis; and that it requires reliable estimates of the inputs. So the number needed to screen is interpreted alongside the harms and costs of screening and with attention to the population and assumptions, since it summarises only the benefit side and depends on estimates that vary by setting, meaning it informs but does not by itself determine the value of a screening programme.

Source: Rembold 1998