A Deloitte report found that more than 80% of health care and life sciences executives see improving health equity as a top priority. Nearly half of survey respondents plan to increase spending on health equity, and drivers for health equity identified by survey participants include regulatory pressure, patient engagement and rising awareness of social drivers of health.

Epic Systems plans to include generative AI in its EHR that will summarize information in patients’ charts and offer clinicians improved research capabilities, in addition to launching its Best Care Choices for My Patient initiative, which seeks to help doctors choose appropriate treatments for patients. Oracle Health will focus on its AI-based Oracle Clinical Digital Assistant that targets patient interactions, with the goal of also eliminating time-consuming documentation tasks, and plans to release a cloud-based EHR system, while Meditech will use Google’s AI tools in its EHR to automate clinical documentation and will adopt HL7 Fast Healthcare Interoperability Resources guidelines.

EHR vendor Meditech plans to focus on AI next year by further integrating Google Health into its Expanse EHR, while it also works to advance interoperability by adopting HL7 Fast Healthcare Interoperability Resources guidelines and supporting users with the Trusted Exchange Framework and Common Agreement and other health information exchange initiatives. Optimization of its EHR is also on tap for 2024, with patient registries and business and clinical analytics among the vendor’s efforts to help users improve efficiency and safety.

Kaiser Permanente is funding five health care AI development projects with grants of up to $750,000 each through its Augmented Intelligence in Medicine and Healthcare Initiative. Projects involve using machine learning or AI for pediatric asthma care, identification of cardiac amyloidosis, precision resuscitation with crystalloids in sepsis, point-of-care diabetic retinopathy screening, and improving diagnoses using randomized trials, algorithmovigilance and proactive risk assessment.

Systematically biased AI models reduce diagnostic accuracy, although accuracy improves with standard AI models, according to a study in the Journal of the American Medical Association. Clinicians were shown vignettes of patients’ health without using AI predictions to determine the likelihood they would develop certain conditions and to create a baseline for accuracy, and clinicians’ accuracy rose when they were shown a standard AI model’s prediction and how the model reached its prediction, but accuracy dropped when clinicians were shown systematically biased AI model predictions with and without explanations.

Data can show hospitals where they can direct their focus to improve health equity, says Lawrence Moss, CEO of Nemours Children’s Health, which has worked to address disparities among disadvantaged groups. Patient data can highlight whether some groups are experiencing worse outcomes than others, and Moss urges health systems to prioritize one area where they can make progress rather than trying to tackle all problems at once.