Key question New
AI-processed · Author-approved
How does the way health risk data is collected affect whether patients trust and act on the results?
Loads Spotify's player; Spotify may set cookies. This device remembers your choice.
Key findings
- As a research-in-progress design paper, it develops a conceptual framework that separates the source that produces risk information from the source that communicates it, theorizing the patient’s own role in generating evidence about themselves as a distinct source cue.
- Clinician-mediated evaluations are theorized to promote behavioral engagement primarily through perceived source credibility and relational trust.
- Wearable sensor tracking is proposed to increase engagement by fostering active bodily participation, personal relevance, and the perception that health risks are modifiable.
- Passive electronic health record calculations are hypothesized to produce lower engagement due to limited patient participation and opaque data provenance, though personal relevance may partially mitigate this gap.
- The study details a standardized field experiment protocol isolating data collection modality while controlling feedback formatting and treatment recommendations.
What it means for you
CIO / IT Executive
When evaluating passive EHR risk-scoring tools, consider pairing them with patient-facing communication that explains how scores are generated. This study proposes, and will test, whether opaque score provenance reduces patient engagement.
IT Manager
Review how automated, patient-facing risk scores are labeled. Plain-language statements explaining how a score was calculated are a low-cost change, and this study is designed to test whether provenance shapes patient trust and follow-through.
Business Strategist
Treat how risk data are generated as part of a digital health product, not only its accuracy. The authors hypothesize that clinician involvement and patient participation in data generation, such as wearables, may each support engagement through different routes.
Researcher
In studies of algorithmic or sensor-based risk assessment, consider separating the entity that produces evidence from the entity that communicates it. Holding communication constant while varying the evidence source offers a way to isolate the recipient’s own role in generating evidence about themselves.
Policymaker
As passive algorithmic risk scores spread in clinical care, consider whether patients receive understandable explanations of how scores are generated. Results from this study will help show whether provenance disclosure affects patient trust and engagement.
Summary
This study outlines the conceptual model and methodology for a randomized longitudinal field experiment investigating how different frailty assessment methods impact patient action. The proposed experiment assigns approximately 300 older adults to clinician-mediated assessment, passive electronic health record scoring, or wearable remote monitoring. While the data collection method varies, participants receive standardized feedback and identical offers for follow-up physical therapy care.
Problem
Digital health technologies enable scalable detection of health risks, but identifying risk does not guarantee patient adoption of recommended treatments. Frailty feedback can evoke negative age stereotypes and may lead patients to disengage from recommended interventions. Furthermore, existing theory conflates the entity that generates risk information with the entity communicating it, leaving the behavioral impact of data-generation methods poorly understood.
Transcript
Expert: That's right, Anna. Digital health tools now make it possible to screen older adults for frailty at scale. But identifying risk is only half the battle. Frailty feedback can evoke negative age stereotypes, which can lead patients to resist the label and disengage from recommended care.
Host: And that creates a major gap between risk identification and risk reduction. How does this study address that challenge?
Expert: As a research-in-progress design study, it outlines a randomized longitudinal field experiment involving approximately 300 older adults through Wake Forest University School of Medicine. The participants will be assigned to one of three data-collection methods: a clinician-mediated assessment, passive electronic health record scoring, or wearable remote monitoring.
Host: And do these different groups get different medical advice?
Expert: No, the clinical feedback and the follow-up recommendations for physical therapy are identical. Only the explanation of how the data was gathered varies. By holding communication constant and varying only how the evidence was produced, the design separates the source that produces risk information from the source that communicates it. This allows the authors to theorize the patient’s own role in generating evidence about themselves as a distinct source cue.
Host: What are the authors hypothesizing about how patients will react to these different modalities?
Expert: They theorize that clinician-mediated evaluations will foster engagement through perceived source credibility and relational trust. Wearables are proposed to increase engagement by fostering active bodily participation, personal relevance, and the perception that health risks are modifiable. Conversely, passive electronic health record calculations are hypothesized to produce lower engagement due to limited patient participation and opaque data provenance, though personal relevance may partially mitigate this.
Host: In other words, how risk data are collected may shape whether patients act on them.
Expert: Exactly. The authors propose that how risk data are generated may act as a cue that shapes whether patients trust and act on the results. If the hypotheses are supported, organizations relying solely on passive algorithmic scoring could see lower follow-through on recommended care, while approaches that involve patients in generating the data, such as wearables, could support engagement.
Host: What is the broader operational takeaway for healthcare organizations and digital health practitioners?
Expert: The study suggests that leaders implementing technology-mediated risk assessments should consider how the evidence source may influence patient behavior, not only whether the algorithm correctly identifies risk. The authors expect this logic to extend to other systems that make evaluative claims about people from their own data.
Host: A critical distinction between data accuracy and patient action. Thank you, Alex, and thank you to our listeners for joining us on A.I.S. Insights — powered by Living Knowledge.
Frailty assessment, digital health, source credibility, patient engagement, remote patient monitoring, electronic health records