International comparison of performance measures for screening mammography: can it be done?

Objective: Published screening mammography performance measures vary across countries. An inter-national study was undertaken to assess the comparability of two performance measures: the recall rate and positive predictive value (PPV). These measures were selected because they do not require identification of all cancers in the screening population, which is not always possible.

Setting: The screening mammography programs or data registries in 25 member countries of the International Breast Cancer Screening Network (IBSN).

Methods: In 1999 an assessment form was distributed to IBSN country representatives in order to obtain information on how screening mammography was performed and what specific data related to recall rates and PPV were collected. Participating countries were then asked to provide data to allow calculation of recall rates, PPV and cancer detection rates for screening mammography by age group for women screened in the period 1997-1999.

Results: Twenty-two countries completed the assessment form and 14 countries provided performance data. Differences in screening mammography delivery and data collection were evident. For most countries, recall rates were higher for initial than for subsequent mammograms. There was no consistent relationship of initial to subsequent PPV, although PPV generally decreased as the recall rate increased. Recall rates decreased with increasing age, while PPV increased as age increased.

Conclusion: Similar patterns for mammography performance measures were evident across countries.However, the development of a more standardized approach to defining and collecting data would allow more valid international comparisons, with the potential to optimize mammography performance. At present, international comparisons of performance should be made with caution due to differences in defining and collecting mammography data.