Instead, she takes a nap while her children eat dinner and watch television. Her brother arrives to watch the kids and wakes her up in time to go to her night job. Maria’s family does not live near affordable grocery stores, where they could purchase quality food. They do, however, pass multiple fast-food restaurants coming and going to their homes.

In this scenario, we would ideally know Maria’s health goals, dietary restrictions, shopping list, context, and schedule. We envision a mashup system that merges these personal inputs with publicly available data streams, such as public transportation and maps, to recommend that in her 30-min- ute break, she could walk three blocks and purchase good produce from a store near her work.

Challenges: Understanding
all the sources of data that
might be necessary to incorpo-
rate into the system that not
only support the user’s health
goal, but also facilitate steps
necessary to achieving that
goal. Another challenge is that
although human-centered com-
puting research is well suited
to address this challenge, we
still must build on social sci-
ence methodologies to identify
the needs of the diverse target
populations participating in
wellness activities.
Photographs from top by Brian Talbot, Sylvain Mercier
The end user (the person stay-
ing well) is the primary user
of the information. In many
health informatics fields, the
end user is a source of data
(e.g., input into clinician-used
systems). By contrast, in well-
ness, the end user is not just a
producer of data, but also the
primary consumer of informa-

tion. For example, consider how socioeconomic status and culture affect the design of technologies. A nutrition system designed for a particular individual must account for how easy it is to find inexpensive, healthy food locally (or the time/money costs of leaving the neighborhood to shop) [ 3]; and whether the food recommendations fit the food-consumption practices of that person [ 4]. Finally, those applications must present his information in ways that make sense, such as being actionable [ 4, 5].

Example: Extending our previous example, Maria explained that the only time she thinks about herself is when she puts on her make-up each morning. Thus, she designed a make-up clamshell case where the mirror could look at her and let her know what she needs (in this case, more water) [ 6]. Although she constantly thinks about the wellness of her family, she cannot always be there because of work constraints. Her solution to this problem was a portable game system that would provide children points on their game for each task they completed—in this case, drinking more milk. The system could also remind the child of other family and wellness-based tasks, such as eating as a family to engender communication.

Challenges: Once the hurdle of
accurately identifying data and
processing the data in context
is cleared, we must address the
issue of presenting this aggre-
gate and contextualized data.
It is not just enough to provide
all the necessary information;
it needs to be given in such
a way that the end user can