Risk factors for unexpected hospital attendances and protocol modifications in patients on systemic cytotoxic anticancer treatments

Background Systemic anticancer cytotoxic therapies (SACT) often have complications that can lead to unexpected hospital attendances usage (UHA) and protocol modifications (PM). UHA have been associated with poorer median survival and increased economic burden on healthcare systems. Our study aims to identify any significant predictors for UHAs and PMs. Methods Our single centre retrospective cohort study aims to identify any significant predictors for UHAs and PMs. Patient demographics, details of UHA and SACT regimen data were collected for each patient between March 2024 to March 2025. For our univariate analysis of possible predictors, we used chi squared for any association between categorical variables and independent t tests for continuous variables. While for multivariate analysis, binary logistic regression was conducted to identify any significant predictors. Results 534 patients were analysed with 218 (40.8%) having stage IV disease. 314 patients (58.8%) had at least one UHA and 378 patients (70.8%) had PMs. There were no significant predictors for UHAs. While patients who progressed on SACT was a predictor for PMs (HR 2.514, 95%CI 1.216–5.198, p = 0.013). Furthermore, progression on SACT (HR 0.441, 95%CI 0.232–0.841, P = 0.013) and line of therapy (HR 0.719, 95%CI 0.567–0.912, P = 0.0006) were predictors for completing all SACT cycles. Finally, age (HR 1.026, 95% CI 1.002–1.051, p = 0.036) and line of therapy (HR 0.679, 95% CI 0.530–0.870, p = 0.002) were predictors for dose reduction. Discussion Line of therapy was a predictor for patients completing their SACT and undergoing a dose reduction. Potentially, older patients could have an upfront dose reduction to improve their quality of life. Our limitations consisted of our study being a single centre and using retrospective data collection. Future multi centre prospective studies are needed to understand the scale of the problem and validate this study results.