The past couple of decades we witnessed a (r)evolution in digital technology, opening new vistas for managing human health. For example, the field of wearable technology has evolved tremendously, resulting in, among others, a myriad of mobile health (mHealth) technologies that have the potential to support management of human health under very diverse conditions.
Notwithstanding their promising value for supporting human health and well-being, there are also important challenges to be overcome when developing these technologies. One such challenge is that the accuracy of most commercially available wearables today are hardly evaluated,1 affecting the quality of the decisions taken on the basis of these wearables and/or of the algorithms developed using the data produced by these wearables. Another important challenge is the analysis of the data generated by these wearables. Data analysis in general and artificial intelligence (AI) more specifically is radically changing the landscape of health management these days and leads to unprecedented possibilities for quantifying health conditions and transitions.2 However, the call for developing explainable AI (opening the black box) becomes louder at 2 levels.3 First, there is the need for methods that allow developers and users of AI to understand how complex algorithms, such as those used in deep learning, are exactly working and how their output is generated. Second, there is also a need for explainability in understanding how the resulting model structures and parameters can be interpreted in biologically relevant terms.
Typically, mHealth technology is developed by engineers who are skilled in areas such as hardware, data management, electromechanics, and AI but much less in medical and/or biological aspects. Hence, users are experiencing many of the wearables, such as smartwatches and fitness trackers, more as gadgets with little added value in actionable advice.
So, besides solving the hardware and software issues mentioned earlier, there is a third challenge that needs to be addressed by the engineering community, namely developing health technology in such way that the users (eg, patients, caregivers, and coaches) are willing to keep using it because they are convinced about the added value for their application. In other words, there is a need to actively involve the user in the development process of such mHealth devices.
In their article in this issue of Mayo Clinic Proceedings: Digital Health, Honinx et al4 describe a systematic review on meditation and breathing devices for stress reduction. In their work, they start from the observation that these devices are without no doubt promising for alleviating unwanted mental load, but suffer from lack of evidence-based information and development and at the same time are mostly developed without involvement of target-users. The authors convey the message that currently health technology is typically developed top-down considering aspects such as commercial strategy, previous innovations, and methods of intuitive selection, resulting in products that are not necessarily addressing the users’ needs. This is especially problematic in the area of mental health where personal engagement is key in successfully applying such technology.
Another issue that is addressed by the authors is the fact that different users might have different preferences, making a one-fits-all approach less obvious or even impossible. As suggested by the authors, a first step to solve this is to stratify the users resulting in different technology designs depending on the targeted population. Although this is certainly an important step in optimizing health technology development, we believe that, in many applications, mHealth technology should be personalized up to the individual’s level to fully reach its potential.5
One of the key points that is addressed by Honinx et al4 is the need of a bottom-up approach in the development of (mental) health technology, considering the needs of the targeted patients. We argue that this bottom-up approach should involve as many relevant stakeholders as possible (not only patients), and ideally, this mindset could already be introduced during the training of the health professionals (including human health engineers). As suggested by Schokkaert et al,6 one might even consider to introduce course modules in the curricula of the care professionals of the future where individuals learn to collaborate in transdisciplinary teams across the boundaries of their own discipline. This does promote not only the interactions between caregivers and technology developers but, consequently, also the cocreation of new health technology involving all relevant stakeholders.
To conclude, the human health engineers of the future should not develop health technology in their ivory towers but link to the biological process involved (ie, patients) and the biologically trained people (ie, doctors, nurses, and coaches) to develop (mobile) health technology by “going from biology to technology” and not vice versa as is still the case today for many health products.
Potential Competing Interests
Dr Aerts is a scientific advisor of the KU Leuven spin-off company BioRICS NV that develops applications for monitoring mental load in individuals.
Acknowledgments
The author acknowledges the support of KU Leuven via a faculty position in human health engineering.
References
- 1.Peake J.M., Kerr G., Sullivan J.P. A critical review of consumer wearables, mobile applications, and equipment for providing biofeedback, monitoring stress, and sleep in physically active populations. Front Physiol. 2018;9:743. doi: 10.3389/fphys.2018.00743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Aerts J.-M. Editorial—special issue on human health engineering. Appl Sci. 2020;10(2):564. doi: 10.3390/app10020564. [DOI] [Google Scholar]
- 3.Rudin C. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nat Mach Intell. 2019;1(5):206–215. doi: 10.1038/s42256-019-0048-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Honinx E., Broes S., Roekaerts B., Huys I., Janssens R. Existing meditation and breathing devices for stress reduction and the incorporated stimuli: a systematic literature review and competition analysis. Mayo Clin Proc Digit Health. 2023;1(3):395–405. [Google Scholar]
- 5.Piette D. Faculty of Bioscience Engineering; Leuven, Belgium: 2020. Depression and Burnout—A Different Perspective: Investigating the Differences and Similarities Between These 21st Century Epidemics Through Data-Based Transfer Function Modelling. PhD Thesis. [Google Scholar]
- 6.Schokkaert E., Aerts J.-M., Callens S., et al. The Health and Care Professions of the Future. Metaforum Position Paper 20. Working group Metaforum. 2023:1–112. [Google Scholar]