Date of Award

Summer 5-2019

Degree Name

Master of Public Health (MPH)


Dr Rebecca Wells PHD

Second Advisor

Dr Mary Frances Mcaleer MD, PHD


PURPOSE: Excessive medical treatment at end-of-life is an indicator of poor quality care. While radiation therapy (RT) is effective for palliation, some patients die shortly after or even during treatment. Any treatment that requires terminal patients to spend significant time in the hospital setting contradicts palliative goals. This study investigates patterns of end-of-life RT to inform quality improvement initiatives. METHODS: All patients who died within 6 months of starting RT at a single large academic cancer center between 2015 through 2018 were identified through our institutional databases on an Institutional Review Board-approved protocol. Clinical factors including age, treatment service, number of fractions, diagnosis, treatment site, and treatment date were evaluated for associations with endpoints, 30-day mortality and on-treatment mortality (mid-course), via logistic regression analysis. RESULTS: 1,855 patients died within 6 months of initiating RT at our center. Of these, 619 patients (33%) died within 30 days of starting RT, and were most commonly treated by thoracic (26%), CNS (21%), and hematologic (13%) services. Commonly treated sites included brain/spine (27%), bone (26%), and mediastinum/thorax (10%). On logistic regression, both extended radiotherapy prescription fractionation (>10 fractions/course) [OR 0.50, p<0.001] and advanced stereotactic treatment technique (OR 0.61, p=0.002) were associated with decreased likelihood of 30-day mortality, reflecting appropriate clinical rationale of treating providers. Neither age (≥70 vs. <70 years) [OR 0.93, p=0.538] nor treatment year (2017-2018 vs. 2015-2016) [OR 0.97, p=0.744] were associated with 30-day mortality. Of the 619 patients, 142 (23%) died midway before completion of RT course. Patients treated for emergent palliative mediastinal/thoracic indications (OR 11.4, p<0.001) were more likely to die midway through RT than those treated for bone metastases. Notably, 2 out of every 3 patients treated for emergent palliative mediastinal/thoracic indications (e.g. airway obstruction, hemoptysis) died on-treatment, comprising 27% of all on-treatment deaths (p<0.001). CONCLUSION: Palliative RT remains an important therapeutic tool at the end-of-life. However, careful consideration of RT for emergent mediastinal/thoracic indications should be used, given the high potential for on-treatment mortality. Taken together, these data may help inform physician decision-making and facilitate treatment consistent with palliative goals at the end-of-life.