| Abstrakt: |
Background: Identifying factors that influence the length of hospital stay for suspected stroke patients is crucial for optimizing the utilization of hospital resources. This study aimed to determine the factors associated with the length of hospital stay for suspected stroke patients transferred to Qaem Hospital in Mashhad through emergency services using survival analysis. Methods: In this historical cohort study, general information was gathered for all suspected stroke patients who sought emergency services in Mashhad, the largest city in northeast Iran, from March 21, 2018, to March 20, 2019, and were then transferred to the Emergency Department of Qaem Hospital. Pre-hospital emergency data were integrated with hospital records using the mission ID. The primary outcome assessed in the study was the length of hospital stay, with model implementation carried out using the statistical software Stata. Results: The median hospitalization time until patients' recovery was seven days. Out of the 578 participants, 386 cases (66.8%) recovered, while the remaining 190 cases (33.2%) were censored (83 individuals had died during the study, and 107 individuals had exited the hospital for other reasons). The average age of patients at the time of hospitalization was 71.13±13.01 years. Statistical analysis employing Log-rank and Breslow tests identified a significant difference in hospitalization duration among patients receiving various levels of care and based on their insurance status. During multivariate analysis, the Cox regression model was considered unsuitable due to some variables not meeting the proportional hazards assumption, leading to the utilization of AFT models. Following the evaluation of AFT models, including Log-normal, Log- logistic, Exponential, and Weibull, the log-normal model emerged as the most suitable choice, exhibiting AIC and BIC values of 1273.909 and 1356.740, respectively. Significant variables influencing length of stay included patient admission priority, insurance status, season, and residency status. Conclusion: The study suggests that parametric survival models are effective for analyzing lifetime data. Additionally, in light of the significant variables identified, enhancing facility readiness and resource allocation could facilitate more efficient planning and implementation. [ABSTRACT FROM AUTHOR] |
