Can recommendation systems help people find health-related peer support online? Human-computer interaction researchers love to imply that they can. I wrote a paper exploring this question.

In the last few years I noticed a bunch of papers that:

  • Study an online health community
  • Find some really interesting things
  • Identify design implications

Ridiculously often, one of their design implications is improving peer recommendation systems. I even wrote a paper like that too!

But, empirically, I couldn’t find any research actually using recommendation systems to connect peers (and thereby to increase their health-related social support). Recommendation systems are ubiquitous on the big social media platforms, but the smaller, health-focused communities often have no recommendation system or very basic “recent activity” summaries.

I thought the lack of research on this topic was weird, and I wanted to know why people weren’t using or studying peer recommendation systems, so I built one. With a team of student researchers, I deployed a peer recommendation system and convinced 79 people to use it for 12 weeks. And as with any socio-technical system, it turns out there are a lot of barriers to implementing and evaluating this kind of thing.

Take a look in the paper and I bet you’ll find something that interests you. We’ve got:

  • A quantitative survey asking people what they’re look for in potential peer connections
  • The real impacts of actually using a peer recommendation system
  • A qualitative investigation of how peers interacted with the complete strangers we recommended

Here’s the full citation:

Zachary Levonian, Matthew Zent, Ngan Nguyen, Matthew McNamara, Loren Terveen, and Svetlana Yarosh. 2025. Peer Recommendation Interventions for Health-related Social Support: a Feasibility Assessment. Proc. ACM Hum.-Comput. Interact. 9, 2, Article CSCW146 (April 2025), 59 pages. https://doi.org/10.1145/3711044

A preprint is also available on arXiv: http://arxiv.org/abs/2209.04973