271 OESOPHAGO-GASTRIC CANCER SURVIVAL IN SCOTLAND: LESSONS LEARNT FROM ESTABLISHING A NATIONAL NETWORK

Oesophageal cancer is the 9th most common cancer in Scotland, with 972 new cases in 2017. It remains the 4th most common cause of cancer death, with an estimated 5 year survival of 12.1%. Scotland’s oesophago-gastric cancer network consists of 3 regional divisions, serving a population of 5.44 million. In 2017, survival data indicated survival differences across the 3 networks. This work analyses these outcomes and provides insight into the challenges faced by national collaboration. Methods Upper GI Cancer Quality Performance Indicators (QPIs) for patients diagnosed between January 1st 2013 and December 2015 were collected by clinical audit staff in each NHS Board and submitted centrally to the Information Services Division (ISD). Twelve QPIs are assessed and include: % discussed at MDT, % undergoing neo-adjuvant chemotherapy/CRT, 30/90 day mortality, length of stay, and R1 resection rates. Quality assurance of the dataset was assessed by analysing 20% of records submitted. Both univariate (log rank testing and Kaplan–Meier survival curves) and multivariate survival analysis (Cox’s proportional hazards model) were then performed on all patients across the networks. Results —Quality assurance of the dataset was high—96% accuracy of those records analysed.—Univariate analysis showed that increased age, higher deprivation scores, poor performance status, high tumour grade, an R1 resection, and a poor Charlson comorbidity index, had worse survival. Gender was not a significant survival factor. There was no difference at either network or health board level for all patients with oesophageal cancer.—multi-variate analysis revealed survival differences across the networks for radical treatment of oesophageal adenocarcinoma. In addition, the survival of patients receiving radical radiotherapy in one network appeared better than chemoradiotherapy for SCC. Conclusion This body of work highlights the challenges and pitfalls of establishing a national clinical network. Importantly, data collection and accuracy are high. Moreover, unlike other national collaboratives, data submission is compulsory. Whilst survival differences were detected across the clinical networks, further in depth analysis revealed confounding factors e.g. 23% of oesophageal cancers were stage “unknown”. The algorithm used at ISD did not capture stage T4a or b accounting for most of these patients.