Melody studied Decision Sciences at the London School of Economics where she taught decision theories and practices to undergraudates and master students.
Melody's research focuses on applying decision analysis methodologies to medicine. This includes finding ways to improve clinical decison making where uncertainties abound, generating evidence when it is scarce, handling data when they are heterogeneous, performing cost-benefit/risk-benefit analyses that takes into account perspectives of multiple stake holders. She is currently working on developing effective and efficient methods for the assessment of health technologies.
Since 2014, Melody has been collaborating with a large and dynamic group of clinicians, scientists and industry partners as part of the NIHR-Diagnostic Evidence Cooperative. The aim of methodology research is to develop efficient ways of evidence generation that could quicken the porcess from "bench-to-bedside", converting exciting new ideas into tangible benefits to patients and soceity.
Melody was the educational research manager for the National Training Programme (http://www.lapco.nhs.uk/), funded by the Department of Health. Lapco trained around 150 consultant colorectal surgeons in 11 national training centres. The education team developed tailor-made assessment forms and methods based on clinical outcomes and operative videos collected directly from the trainees, lending insights into the process and assessment of skill acquisition at the post-graduate level.
Lapco also provided a fertile ground for conducting research into surgical quality. One well-known factor is surgeon caseload. In an analysis of HES data we found that 30-day mortality from gastrectomy, oesophagectomy and pancreatectomy continued to improve as more cases were being carried out. That is, performances were yet to plateau. This lends support to the notion of continued practice at the specialists level.
Uncertainties and risks are abundant in mediine. Melody's first project at Imperial was to improve the safety of feeding using nasogastric feeding tubes when no bed-side tests gave definitive answers. After evidence was retrieved from the literature, Bayesian networks were used to synthesise evidence and present the dynamics between test results and implications for decisions. This research found that using the existing test method (pH paper), under the current cutoff (5.5), feeding incidents will continue to happen. Partly built on this research, another grant from Innovate UK to DEC London and the Scottish company Ingenza supported the development and first-stage clinical study of a novel test aimed at reducing uncertainties in pH testing.
et al., Surgical versus general practitioner assessment: diagnostic accuracy in two-week wait colorectal cancer referrals (poster), Association of Surgeons of Great Britain and Ireland (ASGBI), Bournemouth, UK, 11-13 May 2011
et al., Hospital volume and survival in oesophagectomy and gastrectomy for cancer: a population-based cohort study (oral presentation), Association of Surgeons of Great Britain and Ireland (ASGBI), Bournemouth, UK, 12 May 2011