Health equity and artificial intelligence in nephrology

Excerpt:

"Racial and ethnic disparities in the prevalence rates of ESKD and transplantation remain a major challenge for society and the field of nephrology. The introduction and conscientious use of artificial intelligence (AI) will increasingly drive nephrology care practices and nephrology-related public health/health system surveillance with the promise of improving patient outcomes and advancing health equity. The implementation of AI for predictive modeling, clinical decision support system (eCDSS), generative AI, genomics, and more can offer nephrologists targeted and timely diagnostic, prognostic, and personalized treatment advice to improve both population and individual health and attenuate disparities in kidney diseases.¹ Using various patient and community-level health-related data, AI can provide insights that encompass and exceed electronic health record (EHR) or clinical trial data alone. Using natural language processing to collect data from EHR-free text and elsewhere can advance health services research and nephropathology and has improved patients' 5-year risk prediction for developing kidney failure. However, natural language processing has inherent sources of bias, including the data source, the annotation process, and word/character inputs."

doi: doi.org/10.2215/CJN.0000000673 PMID: 39874085