Facility-level JSMO specialist availability and reconstructable treatment-process measures in colorectal cancer using...

C-CAT lacks patient-level data on Japanese Society of Medical Oncology (JSMO) specialist involvement. We assessed facility-level registry-listed JSMO specialist availability and C-CAT-reconstructable treatment-process and endpoint-ascertainment measures in colorectal cancer.

This nationwide retrospective C-CAT facility-level analysis included 11,906 patients at 261 facilities. The primary exposure was facility-level JSMO specialist count (0-3 vs 4 +); the secondary exposure was gastrointestinal-domain JSMO specialist presence. The primary endpoint was time from systemic therapy start to recorded second-line treatment end. Facility-cluster robust Cox models were used. Endpoint ascertainment, missingness-focused sensitivity analyses, and overall survival (OS) were supportive/contextual.

Overall, 5349 patients at 181 facilities were in the 0-3 group and 6557 patients at 80 facilities in the 4 + group. Median time to recorded second-line treatment end was 20.2 versus 23.1 months (log-rank p < 0.001). In facility-cluster robust Cox models, the unadjusted HR for 4 + versus 0-3 specialists was 0.840 (95% CI 0.761-0.927; p < 0.001), attenuating after facility-category adjustment (HR, 0.948; 95% CI 0.853-1.055; p = 0.329) and clinical adjustment (HR, 0.968; 95% CI 0.866-1.082; p = 0.565). Second-line end-date availability was lower in the 4 + group (61.6% vs 69.1%). Supportive OS showed no specialist-associated survival advantage (median, 50.2 vs 51.3 months; clinically adjusted HR, 0.962; 95% CI 0.816-1.135; p = 0.648).

C-CAT can reconstruct facility-level treatment-process and endpoint-ascertainment measures, but current C-CAT data do not support specialist-attributable clinical interpretation. Direct evaluation requires chemotherapy-specific national database elements.