Agent-based Simulation for Online Mental Health Matching

Yuhan Liu,Anna Fang,Glen Moriarty,Robert Kraut,Haiyi Zhu

2023-03-21

Abstract:Online mental health communities (OMHCs) are an effective and accessible channel to give and receive social support for individuals with mental and emotional issues. However, a key challenge on these platforms is finding suitable partners to interact with given that mechanisms to match users are currently underdeveloped. In this paper, we collaborate with one of the world's largest OMHC to develop an agent-based simulation framework and explore the trade-offs in different matching algorithms. The simulation framework allows us to compare current mechanisms and new algorithmic matching policies on the platform, and observe their differing effects on a variety of outcome metrics. Our findings include that usage of the deferred-acceptance algorithm can significantly better the experiences of support-seekers in one-on-one chats while maintaining low waiting time. We note key design considerations that agent-based modeling reveals in the OMHC context, including the potential benefits of algorithmic matching on marginalized communities.

Human-Computer Interaction,Artificial Intelligence,Machine Learning

What problem does this paper attempt to address?

This paper attempts to solve the problem of the imperfect user - matching mechanism in online mental health communities (OMHCs). Specifically, current OMHC platforms mainly rely on simple methods (such as first - come - first - served or only based on discussion topics) when matching users seeking support and support providers, without taking into account the unique characteristics and preferences of users. This has led to problems such as poor user experience, a decrease in long - term relationships, and a reduction in member participation, especially having a more significant impact on marginalized groups.

### Core issues of the paper
1. **Insufficient matching mechanism**: The existing matching methods cannot fully meet the needs and capabilities of users, resulting in a poor user experience.
2. **Special needs of marginalized groups**: The lack of a targeted matching mechanism may have an adverse impact on marginalized groups (such as adolescents, gender minorities, etc.).
3. **Balancing different goals**: When designing a matching algorithm, it is necessary to balance multiple goals, such as the trade - off between matching accuracy and speed, and between supporting vulnerable populations and other groups.

### Solutions
To address these problems, the authors developed an agent - based simulation framework and cooperated with one of the world's largest OMHC platforms to use real - data to verify the effectiveness of the simulation. Through this framework, researchers can compare different matching algorithms and observe their impact on various outcome indicators.

### Main contributions
1. **Developing a simulation framework**: Created a "sandbox" environment that allows community stakeholders to try different matching algorithms and helps designers consider complex design trade - offs.
2. **Experimenting with new matching strategies**: Proposed five different algorithm - matching strategies and evaluated the effects of these strategies through simulation.
3. **Revealing design considerations**: Demonstrated the performance differences of different matching algorithms in different user groups, especially the potential benefits for marginalized groups.
4. **Open - source project**: Released the simulation framework as an open - source project, enabling OMHC designers to freely explore and understand the trade - offs in designing actual matching algorithms.

### Experimental results
- **Deferred - Acceptance Algorithm**: Can significantly improve the experience of support - seekers in one - to - one chats while maintaining a low waiting time.
- **Effects of different matching strategies**: Different algorithm choices can achieve different goals, such as optimizing chat scores and reducing blocking rates.
- **Protecting vulnerable groups**: Some matching strategies not only improve the overall user experience but also specifically protect the interests of vulnerable groups.

In conclusion, this paper provides new tools and insights for the design of online mental health communities by introducing an agent - based simulation method, especially in the design of matching algorithms, helping to achieve better user experiences and support effects.