BACKGROUND
About one out of every four people in the United States lives with multiple chronic conditions including cancer, hypertension, diabetes, and heart disease. Attempts to personalize care must take into account the individual’s comprehensive needs rather than focusing on one condition in isolation. Collection and presentation of person generated health data such as symptoms, medication use, physical activity, and health goals are important complements to clinical data in personalized care. There are challenges in integrating and making sense of these various data types in order for both patients and clinicians. User-centered design applied to data visualization has the potential to address this challenge by offering well-designed information displays that fit users’ needs and preferences.
OBJECTIVE
The aim of this study was to assess the perceptions of and feedback regarding visualizations developed to support care of individuals with multiple chronic conditions engaging in cancer care in one important user group, healthcare practitioners.
METHODS
Medical doctors (MDs; n=4) and Registered Nurses (RNs; n=4) providing cancer care at an academic medical center in the western United States provided feedback on visualization mockups. Mockup designs were guided by current health informatics and visualization literature and Munzner’s Nested Model for Visualization Design. Visualizations included: a four-week calendar view summarizing four measures; temporal line graphs of blood glucose, blood pressure, and weight; longitudinal symptom ratings. Visual encodings included color, lines or bands indicating expected normal ranges for various measures, and icons. User-centered design methods, a mock patient persona, and a scenario were used to elicit insights from participants. Directed content analysis was used to identify themes from session transcripts. Means and standard deviations were calculated for health care providers’ rankings of overview visualizations.
RESULTS
Themes identified were data elements, supportive elements, confusing elements, interpretation, and use of visualization. Overall, participants found the visualizations useful and with potential to provide personalized care. Use of color, reference lines, and familiar visual presentations (calendars, line graphs) were noted as helpful in interpreting data. Participants were confused by: the meaning of white spaces, the miniature line graphs indicating symptoms, and the different ways in which symptoms were quantified. Participants were able to see trends and spot recurrent issues using the calendar view and line graphs. The visualizations were perceived of as useful to: better understand the patient outside the clinic; empower patients engaging in self-management, and support provision of personalized care.
CONCLUSIONS
Visualizations guided by a framework and literature can support healthcare providers’ understanding of data for individuals with multiple chronic conditions engaged in cancer care. This understanding has the potential to support the provision of personalized care.