Helping People to Control Their Everyday Data for Care: A Scenario-Based Study

Abstract

With the advent of pervasive sensing devices, data captured about one’s everyday life (e.g., heart rate, sleep quality, emotion, or social activity) offers enormous possibilities for promoting in-home health care for severe chronic care, such as can be found in Spinal Cord Injury or Disorders or the like. Sharing these Everyday Data for Care (EDC) allows care team personnel (e.g., caregivers and clinicians) to assist with health monitoring and decision-making, but will also create tension and concerns (e.g., privacy) for people with health conditions due to the detailed nature of the data. Resolving these tensions and concerns is critical for the adoption and use of a pervasive healthcare environment.

We examine data sharing of EDC to determine how we can better manage the tradeoffs between privacy on one hand and the pro-active sharing of data that one needs for better care. In this paper, we target one critical aspect of using EDC, the problem of sharing an overwhelming number of sensor outputs with numerous care team recipients. We report the results of a scenario-based study that examined ways to reduce the burden of setting policies or rules to manage both the pro-active data sharing and the privacy aspects of care with EDC. In summary, we found that our participants were able to use self-generated groupings of EDC data, and more importantly, largely kept those groupings when creating to share data with potential recipients and when dealing with changes in their health trajectory. These findings offer hope that we can reduce the burden of authoring and maintaining data sharing and privacy policies through semi-automatic mechanisms, where the system suggests policies that are consistent with the users’ preferences - especially as health changes.

A Appendix

Fig. 1.

Participants created groupings of EDC by the level of comfort from the most comfortable (bin 1 on the left) to the least comfortable (bin 5 on the right).

Table 2. Participant description: “Caregiver” (C) refers to a participant who has caregiving experience (including as a nursing professional), “person”(P) refers to “a person with a chronic condition and/or a disability”, and “PFC” refers to “a person with a close family member who has a chronic condition”.

Table 3. EDC groupings shared with different care team roles: primary caregiver (primary), secondary caregiver (secondary), hired or paid caregiver (hired), primary care physician (PCP), psychotherapist (psych), physical therapist (PT), and healthcare system IT specialist (IT). The value represents the highest grouping shared. The last column shows whether a participant adjusted groupings in the process of creating sharing settings. (P20 was omitted since their data was partial.)

Table 4. EDC groupings shared with care team roles increase when health situation deteriorates. Roles include: primary caregiver (PC), hired caregiver (HC), primary care physician (PCP), and emergency room doctor (ERD). The health situations included a regular day (Normal), when symptoms begin to emerge (Symptom), and an emergency requiring a trip to the ER (Emergency). We did not include the hired caregiver role in the emergency situation in our study; we omit participants for which we have only partial data.

Fig. 2.

Heatmap showing how frequently participants assigned an EDC type to each bin. Very light pink 1–5, light pink 6–10, medium pink 5–10, dark pink 11–15, red 16–20, dark red >20 (n = 24). Some participants omitted because of partial data. (Color figure online)