'Are They Doing Better In The Clinic Or At Home': Understanding Clinicians? Needs When Visualizing Wearable Sensor Data Used In Remote Gait Assessments For People With Multiple Sclerosis

Title of the Paper

‘Are They Doing Better In The Clinic Or At Home?’: Understanding Clinicians’ Needs When Visualizing Wearable Sensor Data Used In Remote Gait Assessments For People With Multiple Sclerosis

Paper Information

• Subject Area: Healthcare Visualization and Gait Assessment for Multiple Sclerosis (MS) Patients
• Keywords: Multiple Sclerosis, Gait Assessment, Telemedicine, Clinical Practice Translation, Data Visualization, Wearable Devices
• Publication Year: 2022
• Conference: CHI 2022 (ACM Conference on Human Factors in Computing Systems)
• DOI: 10.1145/3491102.3501989

Research Background and Problem

Identified Problems or Challenges:

• Patients with Multiple Sclerosis (MS) often experience gait impairments, but existing clinical assessments primarily rely on in-person observation, lacking high-precision, long-term quantification methods.
• Although studies using wearable sensors to collect gait data have demonstrated potential benefits, integrating these technologies into clinical practice faces numerous barriers, such as technology acceptance and the complexity of data interpretation.
• The transition of evidence-based practices (EBP) into clinical application typically takes 17 years on average, with only 50% of new practices achieving widespread adoption.

Significance:

• Changes in gait may reflect early progression of MS, and quantitative assessments could facilitate earlier detection of disease changes and tracking of intervention outcomes.
• Remote gait monitoring can provide clinicians with real-world gait data, beyond clinical settings, helping to improve the quality of patient rehabilitation.

Research Motivation and Related Work:

• This study aims to use data visualization design to facilitate the translation of wearable device-based gait analysis from research to routine clinical application.
• Existing research indicates that data visualization plays a crucial role in interpreting complex medical data, but further exploration is needed to tailor it to the needs of different clinical roles.

Solution

Proposed Methods or Solutions:

• Designed and developed a low-fidelity prototype interface to display wearable device-based quantitative gait data, such as stride length and walking speed.
• Through multidisciplinary collaboration (HCI, neurology, rehabilitation), the prototype integrated heatmaps, spider plots, box plots, and key information prompts to highlight significant trends and patient background information.
• Conducted interviews with 10 clinicians to explore how to improve these data-driven tools to support gait assessments for MS patients.

Innovations:

1. Intuitively presented sensor data in a visual interface, reducing the technical barrier to data interpretation.
2. Introduced features such as "long-term tracking" and "key event marking" to enhance the clinical relevance and usability of the data.
3. The co-design approach ensured that the interface effectively supported clinicians' actual needs and aligned with their typical workflows.

Implementation Steps:

1. Collected gait data using motion sensors (e.g., RunScribe system).
2. Designed prototypes using Figma, presenting gait data in relative trends, such as heatmaps and spider plots.
3. Conducted user testing and interviews with 10 clinicians specializing in MS rehabilitation to gather feedback.
4. Applied qualitative analysis methods (e.g., thematic analysis) to summarize clinicians' needs and propose further design optimization suggestions.

Research Outcomes

Specific Findings:

1. Identified clinicians' primary needs and challenges regarding gait data, including balancing quick access to key information with detailed analysis.
2. Proposed visualization design recommendations to support clinical goals, such as dynamically tracking disease progression through specific metrics (e.g., walking speed, community ambulation ability).
3. Emphasized the importance of integrating patient background information (e.g., living environment, medication changes) with gait data to help clinicians construct a "holistic patient narrative."

Advantages:

• Compared to traditional gait assessment tools (e.g., 25-foot walk test), the interface provides multi-dimensional, long-term trend analysis.
• Intuitive visualized data aids in fostering patient engagement and alignment of goals during clinician-patient interactions.
• Supports continuity of assessment from clinical settings to real-world environments.

Experimental or Evaluation Results:

• Feedback from 10 clinicians revealed that the majority found heatmaps and trend lines useful for quickly identifying patterns and making decisions.
• Clinicians noted that the visualization design needs to balance usability and information density, and adding background information and data comparison features would further enhance its clinical utility.

Limitations and Future Directions:

• Limitations:

  • The study included only 10 clinicians, with a limited sample size from a single medical center, requiring further validation of the design's generalizability.
  • The current prototype is designed solely for clinician use; future iterations need to incorporate patient perspectives.
  • Integration with existing electronic health record (EHR) systems has not been considered, which could significantly impact clinical adoption.

• Future Directions:

  1. Validate the prototype's adaptability in broader clinical environments.
  2. Develop patient-friendly interactive interfaces to support clinician-patient collaboration on data.
  3. Explore predictive and analytical capabilities of gait data using machine learning models and design intuitive explanatory interfaces.