The HACLab

The Human Algorithm Collaboration Laboratory (HACLab) administers research projects that involve health care technology, data, and cancer care delivery.  Based at the Perelman School of Medicine at the University of Pennsylvania and led by Director and CHIBE affiliate Ravi Parikh, MD, MPP, the HACLab develops, evaluates, and implements machine learning-based interventions to inform clinical decisions, improve risk-stratification, and ameliorate disparities in care.

Fiscal year 2023 included multiple highlights for the lab. Findings from PROStep – a project highlighting our Patient-Generated Health Data portfolio – were published in JMIR and JCO-OP. In this project, we used Fitbits and symptom text messages to generate dashboard summaries for clinicians to review. Our mixed methods workflow, incorporating qualitative research methods via semi-structured interviews, were published in JPM.  These findings detailed clinician perspectives in designing a virtual palliative care program, suggesting that clinicians are open to specialty services including symptom management, physical therapy, and mental health counseling in a palliative care program but had reservations regarding automated referrals. The HACLab was also pivotal in establishing internal and external partnerships resulting in pragmatic, real-world interventions utilizing algorithms to increase palliative care referrals in Pennsylvania and Tennessee. 

The HACLab was active outside academia throughout fiscal year 2023.  We released a series of two white papers, blogs summarizing our broad portfolio (i.e. performance drift, algorithmic bias, PGHD) for the lay audience, and accepted podium talks at major conferences including the 2024 ASCO Annual Meeting, 2024 AcademyHealth Annual Research Meeting, and the 2024 The Symposium on Artificial Intelligence for Learning Health Systems (SAIL).