Online Decision Support Tool for Personalized Cancer Symptom Checking in the Community (REACT): Acceptability, Feasibility, and Usability Study

Background: Improving cancer survival in the UK, despite recent significant gains, remains a huge challenge. This can be
attributed to, at least in part, patient and diagnostic delays, when patients are unaware they are suffering from a cancerous symptom
and therefore do not visit a general practitioner promptly and/or when general practitioners fail to investigate the symptom or
refer promptly. To raise awareness of symptoms that may potentially be indicative of underlying cancer among members of the
public a symptom-based risk assessment model (developed for medical practitioner use and currently only used by some UK
general practitioners) was utilized to develop a risk assessment tool to be offered to the public in community settings. Such a tool
could help individuals recognize a symptom, which may potentially indicate cancer, faster and reduce the time taken to visit to
their general practitioner. In this paper we report results about the design and development of the REACT (Risk Estimation for
Additional Cancer Testing) website, a tool to be used in a community setting allowing users to complete an online questionnaire
and obtain personalized cancer symptom-based risk estimation.

Objective: The objectives of this study are to evaluate (1) the acceptability of REACT among the public and health care
practitioners, (2) the usability of the REACT website, (3) the presentation of personalized cancer risk on the website, and (4)
potential approaches to adopt REACT into community health care services in the UK.
Methods: Our research consisted of multiple stages involving members of the public (n=39) and health care practitioners (n=20)
in the UK. Data were collected between June 2017 and January 2018. User views were collected by (1) the “think-aloud” approach
when participants using the website were asked to talk about their perceptions and feelings in relation to the website, and (2)
self-reporting of website experiences through open-ended questionnaires. Data collection and data analysis continued simultaneously,
allowing for website iterations between different points of data collection.

Results: The results demonstrate the need for such a tool. Participants suggest the best way to offer REACT is through a guided
approach, with a health care practitioner (eg, pharmacist or National Health Service Health Check nurse) present during the
process of risk evaluation. User feedback, which was generally consistent across members of public and health care practitioners,
has been used to inform the development of the website. The most important aspects were: simplicity, ability to evaluate multiple
cancers, content emphasizing an inviting community “feel,” use (when possible) of layperson language in the symptom screening
questionnaire, and a robust and positive approach to cancer communication relying on visual risk representation both with affected
individuals and the entire population at risk.

Conclusions: This study illustrates the benefits of involving public and stakeholders in developing and implementing a simple
cancer symptom check tool within community. It also offers insights and design suggestions for user-friendly interfaces of similar
health care Web-based services, especially those involving personalized risk estimation.