Back

The Role of Regional and Practice Trial Sites in Distorted Randomized Cancer Trial Enrollment

Shah, S. J.; Manz, C. R.; Balthis, B.; Raman, H. S.; Abaluck, J.; Keating, N. L.; Agha, L.

2025-02-24 epidemiology
10.1101/2025.02.24.25322776 medRxiv
Show abstract

Representative trials are critical to advancing cancer treatment, yet little is known about how geographic siting contributes to non-representative enrollment. Using patient-level data, we determined how the choice of trial-enrolling regions and practices impacts representativeness. We created a SEER-Medicare cohort of people [≥]65 years old with lung, breast, pancreatic, or renal cancer (2014-2019). We identified randomized cancer drug trial participants and determined the prevalence of age [≥]75, sex, race, ethnicity, and rural residence in the full cohort, trial-enrolling regions, trial-enrolling practices, and trials. The choice of region and practice contributed to >50% of the under-enrollment of Black, Hispanic, and rural patients. Cancer trials enrolled 45% fewer Black patients than expected with proportional representation. Trial recruitment in regions and practices with proportionately fewer Black patients accounted for 27% and 35% of this disparity, respectively. These findings suggest that diversifying cancer trials requires changing the regions and practices referring and enrolling patients.

Matching journals

The top 10 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.