Table 1 R4P Model for Digital Medicine Innovation.
From: Promoting racial equity in digital health: applying a cross-disciplinary equity framework
AI Algorithms | Wearables | Telehealth | |
|---|---|---|---|
R4P Model | |||
Remove | - correct bias in algorithms - recall applications on market dangerous to health equity | - increase health literacy/education about wearables | - increase structural access to telehealth-enabling technology, e.g. broadband, video/audio device |
Repair | - assess barriers to representation in model training data from diverse populations - train developers on sources, dangers, and prevention of biased algorithms | - assess barriers to access for wearables - increase transparency about wearable device accuracy/efficacy | - telehealth training and literacy initiatives - build cultural competence + technical skills in providers who use telehealth with diverse patients |
Restructure | - establish equity/diversity criteria for datasets and approval - establish organizational incentives, e.g. provider-developer agreements | - health equity criteria for FDA approval process - public funding or payer coverage for access to wearables | - coverage of telehealth services - racial equity as quality metric for telehealth |
Remediate | - establish liability for algorithmic bias and racism - establish protection of marginalized groups - consider new ways to recruit diverse populations | - increase research funding for inclusive wearable devices - consider new ways to recruit diverse populations | - actively assess current/ongoing adoption of telehealth from health equity standpoint |
