Measuring Health-Related Quality of Life with Multimodal Data

One of the main advantages of digital health is that measurements of performance and behaviour can be made objectively and often continuously. Advances in digital health tools mean it is now possible to simultaneously measure several modalities. Machine learning technologies are also being applied to convert multimodal data into measures of Quality of Life (QoL), such as fatigue, stress and depression. These measures have the potential to be used to create personalised medicines and improve care plans for patients. However, there are several barriers that need to be overcome before these technologies can be used in the real world.  

I was part of a workshop on “Measuring Quality of Life with Multimodal Data” as part of the 17th IEEE-EMBS International Conference on Wearable and Implantable Body Sensor Networks. There were two sessions within the workshop, the first focused on disease detection and the second on the measurement of wellbeing.  

Session one: Disease detection 

During session one, representatives from multiple organisations spoke about their own experiences with using person-generated health data (PGHD) across different conditions, including Parkinson’s disease, multiple sclerosis and sleep disorders. There were also discussions about the advantages and disadvantages of the different methods of data collection: through Bring Your Own Device (BYOD), via an app or by provisioned devices. For example, while BYOD means that participants are likely to have greater adherence, provisioned devices allow more uniform data collection and are more accessible for those who do not have their own devices.  

Session two: Measuring Quality of Life and wellbeing 

Session two focused on how QoL and wellbeing can be measured. During this session, I presented our work on the measurement of fatigue. Fatigue is complex and highly heterogenous, so our approach is to capture data through multiple means including active tests, voice biomarkers and passive data collection.  

A key part of the overall discussion was how to measure wellbeing as a whole versus measuring specific aspects of QoL. While measuring specific aspects of QoL can be relevant for clinical development and treatment, measuring general wellbeing can be very useful for public health.  

Conclusions 

The discussions emphasised the point that digital measures are to be used as additional tools that complement patent-reported outcomes. The tools should be as unobtrusive and low-burden as possible, meaning the focus should be on measures that matter when defining patients’ health or general wellness.  

The need to adopt digital products and devices in clinical trials was also discussed. Defining outcome measures and adopting these in clinical trials will help to develop and validate endpoints. The use of digital products and devices in clinical trials has increased rapidly but mostly in observational studies. However, clinical applications of multimodal digital data have begun to be used in areas such as Parkinson’s disease, cognitive decline in Alzheimer’s disease and COVID-19. There is also increasing interest in using digital measures for drug discovery through capturing ‘digital phenotypes’ in real-world settings.  

Challenges 

However, several challenges remain. Slow and limited collaboration and reluctance to share data means it can be difficult to develop digital measures at-scale. Another challenge is that multimodal sensor data lacks broadly accepted and adopted common data models. Creating these models is needed to integrate and synchronize data, which is often a significant technical challenge for individual studies. Wearables also have short life cycles and the digital health market is moving forwards rapidly, which creates a challenge for data integration and reproducibility. 

Future steps 

The key takeaways from the workshop were:

• It is critical to incorporate practical, representative and systemic approaches to involving patients in everyday health decisions, including when it comes to digital health.
• Decision support systems or outcomes for clinical development are important, and regulators in this space should be engaged early.
• Clinicians need to confident in using digital health measures in order to bridge the gap between measures and medicine.
• There has been a great amount of progress in the field of and we are optimistic about the potential for digital health measures of QoL in the future.  

Full author list 

Ieuan Clay, Francesca Cormack, Szymon Fedor, Luca Foschini, Giovanni Gentile, Chris van Hoof, Priya Kumar, Florian Lipsmeier, Akane Sano, Benjamin Smarr, Benjamin Vandendriessche, Valeria De Luca 

Tags : cantab | digital health | digital tools | patient-centric | quality of life