Clarifying the Concepts of Personalization and Tailoring of eHealth Technologies: Multimethod Qualitative Study.

Abstract

Background: Although personalization and tailoring have been identified as alternatives to a "one-size-fits-all" approach for eHealth technologies, there is no common understanding of these two concepts and how they should be applied.

Objective: This study aims to describe (1) how tailoring and personalization are defined in the literature and by eHealth experts, and what the differences and similarities are; (2) what type of variables can be used to segment eHealth users into more homogeneous groups or at the individual level; (3) what elements of eHealth technologies are adapted to these segments; and (4) how the segments are matched with eHealth adaptations.

Methods: We used a multimethod qualitative study design. To gain insights into the definitions of personalization and tailoring, definitions were collected from the literature and through interviews with eHealth experts. In addition, the interviews included questions about how users can be segmented and how eHealth can be adapted accordingly, and responses to 3 vignettes of examples of eHealth technologies, varying in personalization and tailoring strategies to elicit responses about views from stakeholders on how the two components were applied and matched in different contexts.

Results: A total of 28 unique definitions of tailoring and 16 unique definitions of personalization were collected from the literature and interviews. The definitions of tailoring and personalization varied in their components, namely adaptation, individuals, user groups, preferences, symptoms, characteristics, context, behavior, content, identification, feedback, channel, design, computerization, and outcomes. During the interviews, participants mentioned 9 types of variables that can be used to segment eHealth users, namely demographics, preferences, health variables, psychological variables, behavioral variables, individual determinants, environmental information, intervention interaction, and technology variables. In total, 5 elements were mentioned that can be adapted to those segments, namely channeling, content, graphical, functionalities, and behavior change strategy. Participants mentioned substantiation methods and variable levels as two components for matching the segmentations with adaptations.

Conclusions: Tailoring and personalization are multidimensional concepts, and variability and technology affordances seem to determine whether and how personalization and tailoring should be applied to eHealth technologies. On the basis of our findings, tailoring and personalization can be differentiated by the way that segmentations and adaptations are matched. Tailoring matches segmentations and adaptations based on general group characteristics using if-then algorithms, whereas personalization involves the direct insertion of user information (such as name) or adaptations based on individual-level inferences. We argue that future research should focus on how inferences can be made at the individual level to further develop the field of personalized eHealth.