98 Hospitalisation Rates following Systemic Anti-Cancer Treatments for Patients with Non-Small Cell Lung Cancer

Abstract

Background Patients diagnosed with non-small cell lung cancer (NSCLC) at an advanced stage typically have poorer outcomes, making decisions over choice of systemic anticancer treatments (SACT) complex for both clinicians and patients. We are developing a risk prediction model to aid patient-clinician communication and give personalised estimations of the risk of hospitalisation following the first SACT treatment. To enable development, we firstly needed to examine the unplanned hospital attendance/admission rates by regimen. Methods This retrospective cohort study used population data for patients with NSCLC within England. National Cancer Registration, SACT and Hospital Episode Statistics Datasets were used. We included only patients with advanced disease (stages 3b and beyond) treated in England between 01/01/2015 and 31/12/2017. We calculated the number of unplanned hospital admissions and deaths with 30-days of SACT initiation and compared these between treatment groups. Pearson's Chi-squared test was used to test for significant differences in characteristics by patient factorss and SACT regimen received. Results 24,480 eligible patients were included with a median age of 68 years (IQR 61-74). 54% of patients were male (n = 13,179) and 46% were female (n = 11,301). 78% were diagnosed with stage 4 disease (n = 19,145) and 22% with stage 3 (n = 5,335). A total of 8740 (36%) of patients had an unplanned hospital admission/attendance within 30 days (n = 8,740) and 5.3% of patients died within 30 days of starting SACT (n = 1,294). 30 admissions/attendances were associated with treatment received (p < 0.001), disease stage (p < 0.001) and age (p < 0.001) but not with ethnicity (p = 0.15) and sex (p = 0.06). Table 1 shows that age, sex and performance status are all factors associated currently with SACT treatment chosen (p < 0.001). Conclusions We found hospital attendances/admission rates were high (36% 8740/24480). Data will be used to develop a tool that will enable patients and clinicians understand individualised risks of these to support shared treatment decision making. Disclosure No conflicts related to the abstract. Funding was provided by the NIHR Research for Patient Benefit.