NL
Since the start of breast cancer screening, teams of radiologists who assess screening examinations have been visited every three years. The positive predictive value (PVW) referral diagram, introduced in 2001, helps visiting teams to analyze screening results. Tanya Geertse's study shows that the diagram provides insight into the relationship between referral rates, detection rates, and PPV. This reveals differences in performance and enables inspection teams to make targeted recommendations to further improve the quality of assessments by screening radiologists.
Since the start of Dutch breast cancer screening, quality assurance has been an integral part of the program. Every three years, an assessment takes place, during which teams of screening radiologists – the assessment units (AUs) – are evaluated. For radiologists, insight into their performance is important in order to know where improvement is possible. Important parameters are the referral rate, detection rate, and positive predictive value (PPV). Because these figures are interrelated, it is useful to view them in conjunction with each other.
That is why Robert Blanks[1] introduced the PPV referral diagram in 2001. This diagram plots the PPV against the referral rate and shows the detection rate as ‘isobars’. Target values and national averages can also be added. This gives assessment teams a clear picture of performance.
In our study, we investigated whether this diagram is suitable for monitoring performance and making recommendations. We analyzed visitation data from 12 BEs from 2010–2019. The diagrams showed large differences between BEs. In initial screening examinations, the PVW varied from 4.9% to 23.7%, and in follow-up screening examinations from 21.2% to 54.3%. Target values were achieved more often in follow-up screening examinations (2010: 8 BEs; 2019: 10 BEs) than in initial screening examinations (2010: 0 BEs; 2019: 5 BEs). This led to more recommendations for initial screening examinations (24 versus 13). All recommendations concerned adjusting the referral rate. In 17 of the 24 cases, this improved in the desired direction, but not always enough to achieve target values.
Conclusion: the PVW referral diagram is valuable for identifying underperformance and providing targeted advice. Feedback at BE level helps, but is not always sufficient for individual radiologists. Group feedback combined with personal feedback can increase effectiveness.
Blanks RG, Moss SM, Wallis MG (2001) Monitoring and evaluating the UK National Health Service Breast Screening Programme: evaluating the variation in radiological performance between individual programmes using PPV-referral diagrams. Journal of medical screening 8(1): 24–28
Tany Geertse, Eric Tetteroo, Maartje Smid-Geirnaerdt, Lucien Duijm, Ruud Pijnappel, Daniëlle van der Waal D, Mireille Broeders. Applying the "positive predictive value-recall diagram" to monitor performance and provide recommendations for screening radiologists. Eur Radiol. 2025 Sep 4. doi: 10.1007/s00330-025-11978-3.