JMIR Publications

ABSTRACT

Background:

The internet is an increasingly important source for health information, but information quality varies substantially. Disparities in one’s ability to identify reliable information and to make sense of such contribute to negative psychosocial outcomes and aggravate digital divide-related health disparities. The ever-growing amounts of health information available online are increasing demand for tools that provide personalized and actionable health information. Such tools include symptom checkers, which provide users with a potential diagnosis after responding to a set of probes about their symptoms. Although the potential for their utility is great, little is known about actual usage and effects of such tools.

Objective:

We sought to understand who is using an online AI-powered symptom checker and for what purposes; how they evaluate the experiences of online information seeking and the interactive interview and the quality of the information, and what they intend to do with the recommendation.

Methods:

Cross-sectional survey of online health information seekers immediately following completion of symptom checker visit (N=2,437). Symptoms and diagnoses were coded using CDC’s National Ambulatory Medical Care Survey.

Results:

Buoy users were well-educated (81% some college or more), mostly white (72%) and female (85%). Most had insurance (89%), a regular healthcare provider (76%), and reported good health (59%). Just three types of symptoms – pain (35%), gynecological issues (13%), and masses or lumps (8%) – accounted for 55% of site visits. Buoy’s primary recommendation was more evenly split across the less-serious triage categories: Primary Care Doctor in 2 weeks (35%), Self-Treatment (21%), and Primary Care in 1-2 days (17%) as the top three recommendations received by Buoy. Common diagnoses were musculoskeletal (12%), gynecological (12%), and skin conditions (12%), and infectious diseases (12%). Users generally reported high levels of confidence in Buoy (Mean (SD)=3.47(0.97)), found it useful (Mean (SD)=4.18(0.81)), easy to understand (Mean (SD)=4.64(0.53)), and said Buoy made them feel less anxious (Mean (SD)=3.60(1.05)) and more empowered to seek medical help (Mean (SD)=3.75(0.96)). Users had strong intention to follow recommended action: 76% intended to follow Buoy’s recommendations, while 66% intended to discuss Buoy’s advice with a physician. Users for whom Buoy recommended “Waiting/Watching” (Mean (SD)=4.38 (0.90)), or “Self-Treatment” (Mean (SD)=4.33 (0.93)), had the strongest intentions to comply, whereas those who were advised to seek primary care had weaker intentions.

Conclusions:

Results demonstrate the potential utility of an online health in formation tool to empower people to seek appropriate care and reduce health-related anxiety. An interactive symptom checker might provide more personalized and potentially reliable medical information compared to other forms of online health information seeking. Yet despite encouraging results suggesting that the online tool may fill unmet health information needs among women, African Americans, and Latinos, analyses of the user base illustrate persistent second-level digital-divide effects.