The effect of volume on esophageal cancer resections: What constitutes acceptable resection volumes for centers of excellence?

Graph of in-hospital mortality rates for individual hospitals by annual hospital esophageal resection volume. These data represent the annual hospital mortality rates of in-hospital death at 1506 different hospitals. The data points have been staggered to illustrate the trend because of the high frequency of overlapping values.

Graph showing the different mortality rates above and below each volume threshold. Squares indicate the mortality rates of hospitals with annual esophagectomy volumes for esophageal cancer less than the volume threshold. Triangles indicate the mortality rates of hospitals with annual esophagectomy volumes for esophageal cancer greater or equal to the volume threshold. The difference between each pair of mortality rates at a given annual hospital volume is statistically significant for all volume thresholds except more than 29 and 34. Associated data are listed in Table 2.

Graph of goodness of fit versus annual hospital resection volume. Each point represents one resection volume. McFadden's pseudo r² is shown as percent. Dashed line represents “baseline” McFadden's pseudo r² from the model without resection volume (3.23%).