RESUMEN
OBJECTIVE: The objective of this study was to examine the effect of time spent viewing images and level of confidence on a screening mammography test set on interpretive performance. MATERIALS AND METHODS: Radiologists from six mammography registries participated in this study and were randomized to interpret one of four test sets and complete 12 survey questions. Each test set had 109 cases of digitized four-view screening screen-film mammograms with prior comparison screening views. Viewing time for each case was defined as the cumulative time spent viewing all mammographic images before recording which visible feature, if any, was the "most significant finding." Log-linear regression fit via the generalized estimating equation was used to test the effect of viewing time and level of confidence in the interpretation on test set sensitivity and false-positive rate. RESULTS: One hundred nineteen radiologists completed a test set and contributed data on 11,484 interpretations. The radiologists spent more time viewing cases that had significant findings or cases for which they had less confidence in their interpretation. Each additional minute of viewing time increased the probability of a true-positive interpretation among cancer cases by 1.12 (95% CI, 1.06-1.19; p < 0.001) regardless of confidence in the assessment. Among the radiologists who were very confident in their assessment, each additional minute of viewing time increased the adjusted risk of a false-positive interpretation among noncancer cases by 1.42 (95% CI, 1.21-1.68), and this viewing-time effect diminished with decreasing confidence. CONCLUSION: Longer interpretation times and higher levels of confidence in an interpretation are both associated with higher sensitivity and false-positive rates in mammography screening.
Asunto(s)
Neoplasias de la Mama/diagnóstico por imagen , Competencia Clínica , Mamografía/estadística & datos numéricos , Pautas de la Práctica en Medicina/estadística & datos numéricos , Reacciones Falso Positivas , Femenino , Humanos , Modelos Lineales , Sistema de Registros , Sensibilidad y Especificidad , Encuestas y Cuestionarios , Factores de Tiempo , Estados UnidosRESUMEN
PURPOSE: To develop criteria to identify thresholds for minimally acceptable physician performance in interpreting screening mammography studies and to profile the impact that implementing these criteria may have on the practice of radiology in the United States. MATERIALS AND METHODS: In an institutional review board-approved, HIPAA-compliant study, an Angoff approach was used in two phases to set criteria for identifying minimally acceptable interpretive performance at screening mammography as measured by sensitivity, specificity, recall rate, positive predictive value (PPV) of recall (PPV(1)) and of biopsy recommendation (PPV(2)), and cancer detection rate. Performance measures were considered separately. In phase I, a group of 10 expert radiologists considered a hypothetical pool of 100 interpreting physicians and conveyed their cut points of minimally acceptable performance. The experts were informed that a physician's performance falling outside the cut points would result in a recommendation to consider additional training. During each round of scoring, all expert radiologists' cut points were summarized into a mean, median, mode, and range; these were presented back to the group. In phase II, normative data on performance were shown to illustrate the potential impact cut points would have on radiology practice. Rescoring was done until consensus among experts was achieved. Simulation methods were used to estimate the potential impact of performance that improved to acceptable levels if effective additional training was provided. RESULTS: Final cut points to identify low performance were as follows: sensitivity less than 75%, specificity less than 88% or greater than 95%, recall rate less than 5% or greater than 12%, PPV(1) less than 3% or greater than 8%, PPV(2) less than 20% or greater than 40%, and cancer detection rate less than 2.5 per 1000 interpretations. The selected cut points for performance measures would likely result in 18%-28% of interpreting physicians being considered for additional training on the basis of sensitivity and cancer detection rate, while the cut points for specificity, recall, and PPV(1) and PPV(2) would likely affect 34%-49% of practicing interpreters. If underperforming physicians moved into the acceptable range, detection of an additional 14 cancers per 100000 women screened and a reduction in the number of false-positive examinations by 880 per 100000 women screened would be expected. CONCLUSION: This study identified minimally acceptable performance levels for interpreters of screening mammography studies. Interpreting physicians whose performance falls outside the identified cut points should be reviewed in the context of their specific practice settings and be considered for additional training.