Key Takeaways
- Race-based clinical algorithms are a widespread but hidden problem. These tools remain embedded across California electronic health record (EHR) workflows, labs, clinical guidelines, and vendor tools, with no consistent system-level oversight or inventory.
- Clinical algorithms have a direct impact on care and cost. These tools delay diagnosis, restrict access to treatment, and drive high costs for health systems and payers.
- Fragmented responsibility for clinical algorithms sustains the status quo. Responsibility for these tools is distributed across clinicians, vendors, specialty societies, and payers, making coordinated action difficult, and allowing harmful tools to persist.
- Accelerating AI adoption poses a growing risk. Without intervention, AI-enabled clinical decision tools risk embedding and scaling existing biases.
- California needs a statewide strategy to build a path to equitable care. A multidisciplinary, cross-sector body can build a statewide strategy for eliminating harmful, race-based tools from clinical practice in California.
For decades, race has been improperly embedded as a biological variable in clinical tools used across American medicine — shaping who gets diagnosed, referred, and treated. In this issue brief, Digital Medicine Society (DiMe), a global nonprofit that drives adoption of digital approaches to improve health care, examines how these tools continue to operate across California’s health systems and outlines a path toward eliminating them.
While some high-profile, race-based tools in nephrology, and pulmonology have been addressed, the brief finds that a “long tail” of lesser-known algorithms remains embedded in EHR workflows, vendor software, lab reference ranges, and departmental protocols — with no consistent oversight or inventory. Responsibility for these tools is fragmented across clinicians, vendors, specialty societies, and payers, making coordinated action difficult. As AI-enabled clinical decision tools expand, there is a growing risk that legacy biases will scale further and become harder to detect and correct.
The Path Forward
The DiMe brief calls on California health system leaders to take three immediate steps:
- Conduct comprehensive inventories of race-based tools active in their systems.
- Assess the clinical harm and regulatory exposure each tool creates.
- Engage in building a statewide coordinating body that connects health system findings to specialty society validation, aligns vendors and payers around shared standards, and makes the removal of harmful algorithms replicable at scale.
