Presentation (CTI)

OBJECTIVES: Prescription days supply is used to estimate treatment duration or persistence for oral anticancer therapies. In cyclic oral chemotherapy, days supply may reflect intended on-drug days rather than full calendar-time cycle coverage, so gap-based definitions could register planned rest as uncovered time. This analysis evaluated whether refill intervals characterize cycle timing in gastric cancer real-world data.
METHODS: Medication prescription records for capecitabine and S-1 (tegafur/gimeracil/oteracil) were extracted from a hospital clinical data warehouse for gastric cancer patients during 2023; oxaliplatin records were additionally extracted to identify a CAPOX subset. Same-patient/same-drug/same-date oral records were collapsed into patient-date events using maximum days supply. Refill transitions were defined between consecutive events with positive refill intervals ≤120 days and positive days supply. Days supply, refill interval, apparent uncovered interval, and coverage ratio were calculated. In the CAPOX subset, defined by same-day capecitabine and oxaliplatin records, oxaliplatin intervals served as an intravenous cycle anchor.
RESULTS: The analysis included 270 capecitabine-treated and 114 S-1-treated patients (1,513 and 388 events). Capecitabine showed median 14-day supply (IQR 14-14) and 21-day refill interval (IQR 21-25), yielding median apparent uncovered interval of 7 days and coverage ratio of 0.67. S-1 showed median 28-day supply (IQR 14-28) and 42-day refill interval (IQR 21-43), with median apparent uncovered interval of 14 days and coverage ratio of 0.67. Schedule-consistent patterns were frequent: 14-day capecitabine supply with 19-23-day intervals in 61.5% of transitions, and 28-day S-1 supply with 40-44-day intervals in 40.1%. In the CAPOX subset, capecitabine refill intervals aligned with oxaliplatin intervals, with median absolute difference of 0 days across 944 matched pairs.
CONCLUSIONS: Days supply and refill interval captured distinct information: intended on-drug days versus calendar-time cycle timing. Coverage ratios near 0.67 were consistent with cyclic on/off schedules, supporting refill intervals as a complement to days supply when defining projected exposure for cyclic oral therapies.