Objectives: To compare whether treatment with self-expanding metal stents (SEMS) is more cost-effective than treatment with conventional modalities in patients with inoperable oesophageal cancer. Quality of life effects were also considered.
Design: A multicentre pragmatic, randomised controlled trial with health economic analysis.
Setting: Seven NHS hospitals selected to represent a cross-section of UK hospitals in terms of facilities and staffing.
Participants: All patients attending the centres with oesophageal cancer deemed unsuitable for surgery were assessed for inclusion in the main trial; 217 patients were randomised. A health state utilities substudy was also performed in 71 patients who had previously received curative surgery for oesophageal cancer.
Interventions: Eligible patients were randomised to one of four treatment groups within two study arms. Assessments were performed at enrolment, 1 week following treatment and thereafter at 6-weekly intervals until death, with prospective data collection on complications and survival. Structured interviews to elicit patient preferences to health states and treatments were performed in a substudy.
Main outcome measures: Dysphagia grade and quality of life were examined at 6 weeks. Survival, resources consumed from randomisation to death and quality-adjusted life-years were also considered.
Results: There was no difference in cost or effectiveness between SEMS and non-SEMS therapies, and 18-mm SEMS had equal effectiveness to, but less associated pain than, 24-mm SEMS. Rigid intubation was associated with a worse quality of swallowing and increased late morbidity. Bipolar electrocoagulation and ethanol tumour necrosis were poor in primary palliation. A survival advantage was found for non-stent therapies, but there was a significant delay to treatment. The length of stay accounts for the majority of the cost to the NHS. Patients were found to have distinct individual treatment preferences.
Conclusions: It was suggested that rigid tubes and 24-mm SEMS should no longer be recommended and bipolar electrocoagulation and ethanol tumour necrosis should not be used for primary palliation. The choice in palliation would between non-stent and 18-mm SEMS treatments, with non-stent therapies being made more available and accessible to reduce delay. A multidisciplinary team approach to palliation is also suggested. A randomised controlled clinical trial of 18-mm SEMS versus non-stent therapies with survival and quality of life end-points would be helpful, as would an audit of palliative patient admissions to determine the reasons and need for inpatient hospital care, with a view to implementing cycle-associated change to reduce inpatient stay. A study of delays in palliative radiotherapy treatment is also suggested, with a view to implementing cycle-associated change to reduce waiting time.