Dr Ioannis Sarafis presents the REBECCA project at the EHMA PIE session on Improving healthcare access and outcomes
Dr Ioannis Sarafis, a postdoctoral researcher at the Aristotle University of Thessaloniki – the REBECCA project leader, represented REBECCA with an EHMA PIE presentation at the EHMA 2021 Annual Conference. The presentation titled Moving beyond RCTs: Tapping the potential of Real-World Data for advancing clinical research and improving post-treatment in breast cancer underlined the innovative idea of the REBECCA project, namely using Real-World Data (RWD) to support clinical research and improve existing clinical workflows by combining clinical data with data describing patients’ real-life behaviour including their physical activity, eating habits, sleep, and online interaction.
The EHMA 2021 Annual Conference, hosted by the European Health Management Association – one of the REBECCA partners, was held digitally from 15-17 September 2021 under the theme Health management: Managing the present and shaping the future. EHMA PIE is a new initiative launched this year for the EHMA Annual Conference, designed to give space to discussions that are Provocative, Innovative and Encouraging (PIE) and help shape the future of health management. Dr Sarafis held his presentation at the EHMA PIE session under the sub-theme Improving healthcare access and outcomes.
He started by presenting the goal of the REBECCA project, which is, to improve the quality of life of breast cancer patients and survivors. To achieve this goal, the REBECCA project has two streams of work: clinical research, that is, analysing data to better understand how breast cancer treatment affects patients’ quality of life, and patient management, that is, improving interventions or care choices at the individual level.
Dr Sarafis then introduced the REBECCA Consortium, referencing, in particular, the work of his research group, Multimedia Understanding Group at the Aristotle University of Thessaloniki, which have expertise in developing commercial devices, such as a smartwatch, to monitor patient behaviour and health parameters in real-life conditions. Such devices have been previously used to monitor patients with overweight and obesity, eating disorders and Parkinson’s disease. In REBECCA, they will be used to monitor breast cancer patients.
He proceeded to explain that cancer has several associated comorbidities and disabilities. Cancer patients can suffer from multiple Complex Chronic Conditions (CCCs), which is a specific characteristic of cancer. The REBECCA project will focus on three CCCs in particular: breast cancer treatment-induced peripheral neuropathy, breast cancer-related fatigue, and adjuvant treatment-induced osteopenia/osteoporosis.
Dr Sarafis followed up by presenting the concept of Randomised Control Trials (RCTs), which is the gold standard in clinical research. RCTs are tools used to monitor the causal relationship between interventions and their effect on patients. The first step in RCTs methodology is formulating a hypothesis about a causal relation. Study participants are then assigned to interventions or control groups. The data collected from the study is then used to verify the initial hypothesis. He presented some of the limitations of RCTs, such as combinatorial explosion, the gap between RCTs and clinical practice, and the feasibility of RCTs.
He asserted that the way to overcome these limitations is by including RWD in clinical research. RWD is observational data, generated and collected for purposes other than research. Some examples include patient diagnoses, prescriptions, examination records, visit records, discharge notices, etc.
REBECCA project, in particular, will use RWD from multiple sources that monitor behaviour, lifestyle and emotional status of the patient: electronic health records, wearable devices such as smartwatch and fitness bands, mobile apps, web browser plug-ins, and local environment of the patient. RWD will then be processed to extract indicators, providing an insight into patients’ behaviour and lifestyle, online behaviour, emotional status, and clinical status. This will infer a causal relationship between Patient-Reported Outcome Measures (PROMs) and CCCs.
Dr Sarafis then presented the causal model technique which will be used by REBECCA. An example of a causal model technique is Direct Acrylic Graphs, constructed by experts or generated from health data, where arrows indicate causality. Once the models are created, they can be used to assess the safety and effectiveness of breast cancer treatments, to improve clinical outcomes and PROMs, as well as to develop personalised recommendations and treatments based on the data collected by each patient.
He provided an overview of the REBECCA system, which will help clinicians develop personalised and improved care plans through adjustments in medication and dosage, recommendations for physical activity and diet, and measuring how well patients adhere to prescribed lifestyle guidelines.
Lastly, Dr Sarafis presented some of the caveats and challenges that present themselves in the causal analysis, primarily data heterogeneity from using different measurement sources, errors in the estimation of indicators, and incomplete or missing values in RWD. He also presented some of the measures to counter these challenges such as estimating errors in characteristics, adopting ‘error in variables’ models, and estimating and removing bias.