← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

Evidotes: Integrating Scientific Evidence and Anecdotes to Support Uncertainties Triggered by Peer Health Posts

Shreya Bali , Riku Arakawa , Peace Odiase , Tongshuang Wu , Mayank Goel

DOI PDF Program page Page JSON

Evidotes is a thoughtful reframing of peer-health support: instead of optimizing only for relevance or accuracy, it treats posts as uncertainty triggers and augments them with complementary evidence and anecdotes. The contribution is strongest as a system and interaction design paper with promising mixed-methods evidence, though the evaluation remains exploratory.

Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Review Summary

Evidotes is a strong CHI-style contribution because it does more than propose another retrieval or credibility filter for health content. The paper’s central move is conceptual: it argues that peer health posts should be understood as uncertainty triggers that produce both informational and emotional needs, and that the right response is information augmentation rather than only verification. That framing is a meaningful departure from the common-sense assumption that better ranking or fact-checking is sufficient. The system contribution is also coherent: it combines scientific evidence and anecdotes at the level of individual posts and gives users three lenses—Dive Deeper, Focus on Positivity, and Big Picture—to steer how uncertainty is addressed. This makes the novelty less about a single algorithmic trick and more about a design architecture for co-presenting complementary source types. The validation is appropriate for an early-stage CHI paper: a mixed-methods study with 17 chronic illness patients, baseline browsing comparison, and reported gains in information satisfaction and emotional cost. That said, the evidence is still bounded. The study is small, single-session, and exploratory, so it supports perceived usefulness and interaction value more than durable behavior change or real-world deployment claims. The authors also acknowledge important risks around citation verification and over-reliance on positivity framing, which keeps the paper grounded. Overall, this reads as a credible and well-scoped honorable-mention paper: conceptually sharp, systemically novel, and empirically promising, but not yet validated at scale.

What Changed

Canon before

Prior CHI work on peer health content largely emphasized relevance, accuracy, credibility, or misinformation handling; this paper shifts the framing toward uncertainty support through augmentation of posts with complementary evidence and anecdotes.

Departure from common sense

The paper argues that peer health posts should not be treated mainly as an accuracy or relevance filtering problem. Instead, it reframes them as uncertainty triggers whose value lies in supporting readers' diverse informational and emotional needs through augmentation rather than simple verification.

Actual novelty

Evidotes combines scientific evidence and anecdotal information at the level of individual posts and exposes three user-selectable lenses—Dive Deeper, Focus on Positivity, and Big Picture—to retrieve and present complementary sources. The novelty is in the integrated support model and the lens-based presentation of evidence, not just in retrieval alone.

Evidence

The paper presents a system for augmenting peer health posts with scientific and anecdotal sources and evaluates it in a mixed-methods study with 17 chronic illness patients. Reported outcomes include improved self-reported information satisfaction and reduced emotional cost relative to baseline browsing, alongside qualitative evidence that co-presented sources helped users interpret uncertainty.

Limits

Method limits

The evaluation is exploratory and relatively small in scale, relying on a single-session mixed-methods study with 17 participants. The evidence supports usability and perceived benefit, but not long-term behavioral change, clinical outcomes, or broad generalization across health communities.

Deployment limits

The paper notes that safeguards may be insufficient if users do not verify citations or if they over-rely on positivity-oriented framing. Deployment would also depend on reliable retrieval and careful handling of potentially hallucinated or misleading citations.

Boundary conditions

Findings are bounded to chronic illness patients browsing Reddit in a study setting, with lens-based interactions over a limited set of posts. The approach is most applicable where users want both scientific grounding and peer experience, and less certain where source quality, moderation, or domain norms differ substantially.

Manual check: none

Position in field

This work sits at the intersection of health informatics, social computing, and AI-supported information augmentation. It extends prior efforts on credibility and relevance by proposing a dual-source, uncertainty-centered interaction model that explicitly integrates scientific evidence with lived-experience anecdotes.

Abstract

Peer health posts surface new uncertainties, such as questions and concerns for readers. Prior work focused primarily on improving relevance and accuracy fails to address users' diverse information needs and emotions triggered. Instead, we propose directly addressing these by information augmentation. We introduce Evidotes, an information support system that augments individual posts with relevant scientific and anecdotal information retrieved using three user-selectable lenses (dive deeper, focus on positivity, and big picture). In a mixed-methods study with 17 chronic illness patients, Evidotes improved self-reported information satisfaction (3.2->4.6) and reduced self-reported emotional cost (3.4->1.9) compared to participants' baseline browsing. Moreover, by co-presenting sources, Evidotes unlocked information symbiosis: anecdotes made research accessible and contextual, while research helped filter and generalize peer stories. Our work enables an effective integration of scientific evidence and human anecdotes to help users better manage health uncertainty.