Unplanned hospital admissions as an early surrogate indicator of patient (pt) attrition in phase I trials

Abstract

2516 Background: We have previously reported a high Royal Marsden Hospital (RMH) prognostic score (RPS) of 2- 3 predicts 90-day mortality and reduced overall survival (Arkenau et al BJC & EJC 2008). In this study, we explored the significance of unplanned hospital admissions (UHA) as a potential surrogate indicator of poor clinical outcome. Methods: All pts admitted to RMH Phase I Unit, UK, during 2-month intervals over 3 consecutive years were included in this analysis (2005–2007). We collated pt baseline characteristics, demographic and laboratory profiles, reasons for hospital admissions and relevant clinical trial data. Results: A total of 172 pts accounting for 310 admissions were seen on the Phase I unit during the stipulated time periods (amounting to 6 months in total). Median age: 61 years (range: 19–84); male to female admissions ratio 1.3:1. Pts were on trials of single-agent targeted therapies (69%), cytotoxic combinations (26%), vaccine/viruses (3%) and hormonal modulation (2%). Reasons for planned admissions (n=246) included treatment commencement, PK/PD sampling, paired pre/post treatment biopsies and insertion of central lines. 20.6% (64/310) of overall admissions were unplanned: 50 (78%) were due to disease-associated symptoms/complications and 14 (22%) treatment-related toxicities (TRT). 71% of pts with TRT were on cytotoxic combination trials. Median duration of UHA was 2 days (range:1–20) and there was no relation between length of stay and predicted outcome. 78% of pts in the UHA cohort had a high RPS of 2–3 (i.e. poor outcome) vs 43% in patients whose admission was planned (p=0.001). Of pts who required UHA, only 27% resumed their trial drug after recovery. The main statistically significant risk factors for UHA include >2 metastatic sites, (RR 2.6 [1.64 - 4.38], p=0.001), poor performance status (RR 2.47 [1.48 - 3.72] p=0.003), low albumin (RR 2.17 [1.36 - 3.52] p=0.001) and cytotoxic combination trials (RR 1.7 [1.09 - 2.86] p=0.025). Conclusions: Unplanned admissions constitute 20.6% of Phase I inpatients, with the majority being disease rather than treatment-related. Regardless of length of stay, UHA portend poor outcomes for patients who are on treatment, with a risk profile underscoring the importance of pt and trial selection. No significant financial relationships to disclose.