Background
The sensitivity of endoscopic ultrasound (EUS) in gastro-oesophageal cancer (GOC) staging has been studied; however, the cost-effectiveness of EUS staging has not been evaluated until the inception of the HTA-funded COGNATE trial (ISRCTN1444215). This thesis aimed to explore the cost-effectiveness of EUS in GOC staging alongside the COGNATE trial, the economic evidence and current practice of the utilisation of EUS in GOC staging in the UK, and the potential use of disease-specific measures in cost-utility analysis in clinical trials.

Method
From an NHS perspective, a cost-effectiveness analysis alongside the COGNATE trial (Chapter 2), using QALY as the measure of effect, was conducted to evaluate whether adding EUS to the usual staging strategy is cost-effective in staging patients with GOC.
A systematic review was undertaken retrospectively (Chapter 3) following the completion of the COGNATE trial to identify the economic evidence of EUS staging in patients with GOC.
An online survey of UK GOC healthcare professionals (Chapter 4) was conducted through Bristol Online Survey (BOS) platform to explore the utilisation of EUS in GOC staging and the current clinical practice in the UK.
Given that disease-specific measures are usually collected alongside EQ-5D, a generic preference-based health-related quality of life measure, in clinical trials and seeing that EQ-5D has been argued to be not always sensitive enough to pick up changes in individual’s quality of life, the potential use of disease-specific measures in cost-utility analysis was explored (Chapter 5), aiming to further exploring the transferability/generalisability of the hybrid QALY technique first tried in the MORTISE trial (on which I was a Research Officer in Health Economics working with the Trial Statistician, Dr Daphne Russell) in other disease areas (e.g. cancer and ophthalmology) in cost-utility analysis. Data from two large clinical trials, each with a 12-month follow-up (the COGNATE trial and the NIHR-EME-funded CLARITY trial (ISRCTN32207582)), were used. Regression models between disease-specific measures and conventional QALYs (measured solely by EQ-5D) for both trials were established, and results were compared.

Results
In the COGNATE trial, on average, EUS was found to cost £2,860 less per patient (95% bootstrapped CI: –£7,987 to £2,192) (2008 price year) [2019 price year: £3,490 less per patient, 95% bootstrapped CI: –£9,746 to £2,675] and improved QALYs by 0.1969 years (equivalent to 72 days in perfect health). Combining these savings and benefits showed that the probability of EUS being cost-effective exceeds 95% at the NICE threshold range of £20,000 to £30,000 per QALY (Chapter 2).
The systematic review findings (Chapter 3), based on the six identified economic articles, suggested that use of EUS as a complementary staging technique to other staging techniques for GOCs appears to be cost saving and offers greater QALYs.
From the healthcare professionals survey (Chapter 4), results showed that the majority support the use of EUS in GOC staging (n=89; 90.8%), have experience in the field of EUS either by requesting, performing or both (n=81; 82.7%) with most of them felt that EUS is more useful for staging oesophageal (n=78; 96.3%) and gastro-oesophageal junction cancer (n=78; 96.3%) than gastric cancer (n=58; 71.6%). Interestingly, ‘attend Upper GI cancer MDT meeting’ and ‘clinician’s age’ were found to be important factors (p<0.05) associated with referral for EUS. Attendance at MDT meetings is likely to increase EUS referral, and younger clinicians are less likely to refer GOC patients for EUS.
From the exploratory study of assessing the potential use of disease-specific measure in QALY estimations (hybrid QALY approach) for cost-utility analysis in clinical trials (Chapter 5), disease-specific measures collected in trials were found to be potentially useful for cost-utility analysis. In both trials, cost-utility analysis results showed not only more certainty around the estimates of incremental cost-effectiveness ratios when conventional QALY was replaced with disease-specific measures guided QALY but also a shifting towards the direction of the respective disease-specific measures.

Conclusion
To date, EUS has been found to be not available at every hospital in the UK despite the widespread adoption of EUS, and the economic evidence in this area is scant. This thesis offers various novel contributions to the economic evidence and evaluation of EUS in GOC staging, and the insights into the current practice of the utilisation of EUS in GOC staging in the UK and the advancement of health economics assessment in clinical trials. These novel contributions from this PhD study not only would facilitate policy makers and commissioners in evidence-based decision making but also provide unique insights for future research and policy in this area and other health technology assessment areas.