Screening and the number needed to treat

Area of Science:

• Public Health
• Epidemiology
• Biostatistics

Background:

• Traditional metrics like the number needed to treat (NNT) may not accurately reflect the efficiency of population-based screening programs.
• Existing measures like the number needed to screen (NNS) do not fully account for participation and selection biases inherent in screening trials.

Purpose of the Study:

• To develop and validate a new metric, the number needed to be screened (NNBS), analogous to NNT but suitable for population-based screening.
• To assess the NNBS for breast and colorectal cancer screening programs using data from randomized controlled trials.

Main Methods:

• Proposed the NNBS metric, calculated as the NNT adjusted for screening participation and selection effects.
• Applied the NNBS calculation to existing data from randomized controlled trials for breast and colorectal cancer screening.

Main Results:

• For breast cancer screening, the NNS was 781 women; the NNBS was 601, representing a 23% reduction.
• For colorectal cancer screening, the NNS was 1250 people; the NNBS was 688, a 45% reduction.
• The NNBS provides a more refined estimate of screening effectiveness by incorporating participation and selection factors.

Conclusions:

• The NNBS is a more appropriate measure for evaluating population-based screening programs than traditional NNT or NNS calculations.
• The NNBS is particularly valuable for comparing screening programs with varying participation rates or when contrasting screening with other health interventions.
• Utilizing the NNBS can lead to a more accurate perception of the cost-effectiveness and efficiency of public health screening initiatives.