Project Summary
Bipolar disorder (BD) is strongly associated with financial instability. Symptomatic periods in BD often
manifest in poor financial decision-making. For example, 70% individuals with BD have reported impulsive
spending during hypomania. Such problematic financial behaviors during symptomatic periods can lead to
serious long-term financial instability, which can severely impact the quality of life for individuals with BD and
their care partners. Maintaining financial stability is, thus, a critical challenge to ensure the long-term
wellbeing for individuals with BD. As such, there has been an increasing focus on understanding financial
behaviors of individuals with BD. However, there remains a knowledge gap regarding how idiosyncratic,
context-driven, and illness-specific factors impact financial decision-making in BD. Furthermore, the lack of
granular, in-situ assessment methods is a key challenge against developing just-in-time and personalized
interventions focusing on financial stability for this population. Given the importance of financial stability for
individuals with BD, this remains a serious knowledge gap with broad practical and societal implications.
In recent years, there has been a considerable progress toward more open and accessible financial data. We
argue that the granular and real-time access to financial activity data can lead to a paradigm shift in the domain
of financial wellbeing and BD. That is, it can provide a unique opportunity to explore the nuanced relationship
between financial behaviors and BD, uncover financial patterns indicative of early-warning signs, and develop
preemptive interventions to sustain financial stability and improve long-term quality of life for this population.
This project aims to advance this vision of using financial activity data as an objective behavioral marker in BD.
We will develop a prototype to collect privacy-preserving, customized financial data. The prototype will also
collect symptom and illness trajectory data. We will follow the Center for eHealth Research (CeHRes) roadmap
for implementation. We will use the prototype to retrieve financial data from individuals with BD (N=50) for
the last 24 months (i.e., 1200 months of financial data in total). We will use the data to develop and evaluate
privacy-preserving machine learning models to identify early-warning signs in BD. We will also conduct focus
group interviews to collect in-depth usability and acceptance data from individuals with BD.
This project will establish a preliminary evidence base regarding the feasibility and acceptability of using
financial activity data as an objective behavioral marker in BD. Given the association between BD and financial
instability, the project will provide crucial and urgent insights into assessment and intervention methods to
support financial stability and overall wellbeing in BD, including for those living in remote and rural areas.
