Digital phenotyping as a surrogate for Patient-Reported Outcomes (PROs) in Rheumatic Disease Management

 

Digital phenotyping as a surrogate for Patient-Reported Outcomes (PROs) in Rheumatic Disease Management (pilot study will be conducted in healthy recruits)

 

Lead: Vinaika Maruvada

Team Members:  Dr.  Anjani Reddy Marri, Thanoj Aila, Dr. Jayaraman Tharmalingam, Sofia Cantu  

Collaborators:  Dr. Hien Nguyen

Project Summary:

Lupus nephritis (LN) and other rheumatic diseases are characterized by fluctuating disease activity, requiring  prompt changes in therapy. Interestingly, patient-reported outcomes (PROs) are highly effective for monitoring disease by capturing patients’ personal experiences with symptoms, functioning, and quality of life. However, current patient-reported outcome (PRO) measures are typically collected only during clinic visits, and are subjective. This project aims to develop and validate a digital phenotyping platform that enables continuous, remote monitoring of behavioral, physiological, and psychological parameters relevant to rheumatic disease activity and quality of life.

We plan to conduct a pilot study in healthy University of Houston students to investigate stress and cognitive status using continuous digital phenotyping. This pilot will evaluate the feasibility, data quality, and analytic workflows required for secure data capture, integration, and longitudinal analysis of multimodal behavioral and physiological data.

What is already known in the field?

  • PROs are essential for understanding patient well-being and have demonstrated predictive value for disease activity and flares in rheumatic diseases; however, they are typically collected intermittently and retrospectively, with significant subjectivity.
  • Wearable sensors and smartphones can capture continuous behavioral and physiological data, yet their integration into rheumatology care remains limited.
  • There is a critical need for scalable, home-based telemonitoring systems to support early flare detection and personalized disease management in rheumatic diseases such as LN.

What is new?

  • This study integrates multimodal digital data streams, including activity, sleep, heart-rate variability, breathing rate, skin temperature, screen use, GPS-derived mobility, and menstrual metrics, using wearable devices and smartphone applications.
  • Weekly validated stress surveys and brief audio journals are incorporated to capture subjective experiences alongside objective sensor data.
  • Advanced analytic workflows are developed to securely integrate and longitudinally analyze these heterogeneous data streams, creating individualized digital health profiles.

Why is this important?

  • In the pilot study, continuous digital phenotyping will be used to detect subtle, early changes in stress and cognitive status in healthy participants, enabling validation of data quality, analytic pipelines, and longitudinal tracking.
  • In future patient studies, this same digital framework will be applied to disease-relevant physiological and behavioral markers, with the goal of enhancing routine PRO-based monitoring and early detection of clinical deterioration.
  • This platform lays the groundwork for real-time telemonitoring in LN and rheumatic disease patients, supporting early intervention, reduced disease flares, and improved quality of life.
  • The approach enables a shift toward precision, PRO-guided care, integrating patient-generated data directly into clinical decision-making.

Ongoing/Future Steps

  • Conduct a pilot study to evaluate the feasibility, data quality, participant adherence, and analytic workflows of the digital phenotyping framework in a controlled, non-clinical population.
  • Refine and validate analytic pipelines, including machine learning approaches, using pilot data to identify robust digital features relevant to stress, cognition, and physiological variability.
  • Deploy the optimized digital phenotyping framework in LN and other rheumatic disease patients receiving routine clinical PRO assessments.
  • Integrate digital phenotyping outputs with established clinical indices (e.g., SLEDAI, physician global assessment) to assess clinical relevance and utility.
  • Develop scalable digital tools, such as LupusTracker, to support long-term remote monitoring and disease surveillance.