A study of the pattern of hospital admissions in a specialist Phase i oncology trials unit: Unplanned admissions as an early indicator of patient attrition

Background: Unplanned hospital admissions (UHAs) in the context of oncology Phase I trials are important, yet rarely reported. Methods: All patients admitted to the Royal Marsden Hospital Phase I clinical trials unit during February and March of 2005-2007 were included. The patient-, admission- and trial-related variables were collected. Correlations were sought between the occurrence of UHAs and the baseline patient/trial-related characteristics. Results: Of the 308 admissions involving 177 patients, UHAs constituted 21% of all the admissions and 38% of the total bed occupancy. The majority of UHAs were cancer related (78%) and their occurrence was associated with a significant early patient attrition. Using multivariate analysis, the factors significantly associated with UHAs included age >60 years (RR 2.32, confidence interval (CI)-95% 1.12-4.81), ≥3 metastatic sites (RR 3.26, CI-95% 1.54-6.90) and LDH > ULN (RR 2.18, CI-95% 1.06-4.46), with albumin <35 g/dL trending to significance (p = 0.052). The trials that contained cytotoxic chemotherapy incurred disproportionately higher rates of admissions (69.5%) than the trials that did not. Conclusions: UHAs constitute a substantial workload and impact on the speed and cost of, as well as resource allocation in Phase I oncology trials. The majority of UHAs are cancer rather than treatment related. The risk stratification to guide patient selection may help reduce the incidence of UHAs.