It is widely known that the true success of digital health initiatives ultimately depends on end-user acceptance, motivation and engagement. The harsh reality is that only a fraction of eHealth solutions reaches the user, a smaller fraction of those is being adopted, while an even smaller fraction passes the line of “initial adoption” and achieves sustainability. This reality holds equally for less or more technologically advanced environments. For example, Norway, which is considered an eHealth role model, revealed as part of a national survey in 2013 that despite the wide availability of telehealth across hospitals and a large fraction of the population living remotely, only about 1% of consultations were conducted digitally.
One of potentially many remedies to that challenge is a closer and more careful consideration of patient preferences, primarily when implementing new eHealth infrastructures. Preferences, if extracted correctly, provide information on how technologies should be developed in order to be perceived as useful and valuable. Advocating for a proper integration of patient preferences is easier said than actually done. In reality, resources are scarce, time is limited and the required effort higher than expected. Once the barriers are overcome, the question of when and how patient preferences are to be identified and integrated constitutes a challenge on its own.
Focusing on the “when”, a generic, universally accepted answer does not exist. From the standpoint of person-centred practice, perceived to be the golden standard of care, identifying and responding to user preferences should occur as early as possible in the life-cycle of an eHealth product. From a critical standpoint, one may claim that previous evidence and validated theories provide adequate guidance for designing effective health technologies. Is sound evidence-based of “what works” practically enough? The answer is no. No, because preferences are highly context and time specific and cannot be derived from generalisations. There are numerous examples suggesting that exploring user preferences early enough impacts use and perceived usefulness. A recent smoking cessation study aimed to develop a gamification-based health app. Using a theoretically grounded process, the team carefully integrated expert input, existing evidence and early user preferences, all combined with behaviour change techniques. The results: an app that was perceived as engaging, motivating and very helpful.
Focusing on the “how”, the answer is that preferences can be elicited in multiple, and often complex ways. The most widely accepted and robust approaches are being increasingly adopted from economic and market research. Discrete choice experiment surveys are one such approach. Being considered highly rigorous and statistically sound, discrete choice experiments repeatedly provide users with two or more alternative scenarios of a product, asking them to choose the alternative with the most favourable characteristics. A second widely used approach is that of best-worst scaling surveys, where participants receive single scenarios of single products with multiple characteristics, repeatedly requested to choose the best and worst among those characteristics.
A recent study from Australia used a discrete choice experiment to explore the preferences of older patients regarding telehealth, focusing on (a) how much of actual care is desired digitally (b) the distance to the nearest hospital when using telehealth, (c) the ideal attitude of clinicians towards telehealth, (d) the required level of previous experience with technology, and the (e) costs of telehealth sessions. Part of the study’s conclusions state that patients prefer inexpensive telehealth sessions, that target most aspects of care and are provided by technology-enthusiastic providers. In that case, exploring user preferences revealed that success goes beyond technology itself and is determined by the interest of those who provide them.
Preferences are highly heterogeneous and vary greatly from context to context. That heterogeneity provides rich information on what is perceived as valuable by patients and how that differs across places and populations. What constitutes a good eHealth solution will likely differ between a 65-year old diabetic and 25-year old healthy person. What we need to keep in mind is that the future of eHealth has to be truly patient-centred and increasingly sensitive to the preferences of people, whether those are older diabetics that rely on technology for managing their disease, or younger age groups who just want to self-track their health and well-being.
Vasileios Nittas is interested in the development, implementation and evaluation of digital health solutions for disease surveillance, prevention and control, and is currently conducting his doctoral studies in Epidemiology at the University of Zurich.