Patient-centered modeling of the breast biopsy experience

Area of Science:

• Medical Informatics
• Patient Experience Research
• Oncology

Background:

• Rising patient volumes and time constraints challenge effective patient distress management in breast cancer care.
• Limited resources hinder proactive anticipation and management of patient anxiety during procedures like biopsies.

Purpose of the Study:

• To model key components of the breast biopsy procedure and their impact on patient experience.
• To predict patient anxiety levels using pre-procedural assessments and procedural events.
• To analyze caregiver-patient communication to identify moderators of patient experience.

Main Methods:

• Integrated real-world data from 236 patients, including patient-reported outcomes, psycho-social assessments, and workflow annotations.
• Utilized linear mixed models and machine learning to predict anxiety.
• Employed natural language processing to analyze patient expressions of pain and distress alongside workflow data.

Main Results:

• Psychological pre-assessments significantly correlated with median anxiety during biopsy (e.g., IES β = 0.91, p < 0.001).
• Higher baseline stress strongly predicted greater anxiety.
• Patient anxiety exhibited a temporal pattern, increasing until local anesthesia and decreasing thereafter (βt² p = 5.43e-06).

Conclusions:

• Combining diverse data sources offers a powerful approach to modeling patient experience during medical procedures.
• Developing digital twins of medical procedures can support clinicians in providing proactive care.
• Mitigating patient distress through data-driven insights is crucial for improving the patient journey.