Getty Images/iStockphoto

Racial Disparities in EHR Family Health Information Present CDS Concern

Interventions to improve EHR family health information collection in historically underserved groups could help address informative presence bias in CDS.

Members of racial and ethnic minority groups had less available family health information (FHI) in the EHR, including cancer FHI, compared to White patients, according to a JAMA Network Open study that raises concern for identification disparities in clinical decision support (CDS) algorithms.

Researchers used EHR data from two healthcare systems to inform the potential impact of a CDS algorithm that aims to identify patients eligible for genetic evaluation for hereditary breast, ovarian, prostate, pancreatic, and/or colorectal cancers based on cancer FHI available in the EHR.

Spanish-speaking patients had less cancer FHI available in the EHR, and when available, it was less comprehensive compared to cancer FHI of English-speaking patients.

“These patterns strongly suggest that patients from demographic minority groups in medical care are less likely to be identified as needing specialty healthcare services or with tailored disease prevention recommendations if identification relies on FHI,” the study authors noted.

“Findings such as these, which show the potential of an algorithm to exacerbate health disparities, are essential as part of continuous quality improvement efforts and allow solutions to be developed before integration into a healthcare system so that patients are not overlooked owing to missing or unavailable data,” they added.

For instance, automated procedures could help avoid potential biases based on missing data. However, a recent review of generated prediction algorithms studies found that only 54 percent of studies accounted for missing data in the EHR.

The authors noted that interventions to improve FHI collection, including interventions that target historically underserved groups, are imperative to address informative presence bias.

“This work also highlights how the potential impact of new technologies on disparities should be embedded into the process of development and inform the decision to deploy based on the estimated impact on health inequities,” the researchers emphasized.

Limited prior research has examined the availability of FHI across patient subgroups. Barriers to collecting FHI include limited time, competing demands, reimbursement criteria, and clinician and staff training.

“Prior studies have also shown underuse of FHI owing to incomplete or inaccurate information, lack of awareness about hereditary cancers, lack of awareness of evidence-based guidelines, and time constraints,” the authors pointed out.

Some stakeholders have launched efforts to improve FHI collection through EHR patient portals.

Moving from patient intake forms to electronic formats has improved the completeness, processing, and storage of patient information. However, the authors noted that not all patients have access to a patient portal system.

“As was seen in the transition to telehealth during the pandemic, the shift to digital FHI data collection may reinforce disparities driven by the digital divide,” they wrote. “In addition, this puts the onus on patients to input comprehensive information about their own FHI, and patient-clinician discussions are often still needed to supplement patient input while addressing missing data.”

“Thus, interventions for both patients and clinicians are needed to improve the collection of FHI across patient subgroups,” the researchers said. “While FHI collection is taught in professional schools and emphasized in residency training for many clinicians, there are noted differences in clinical practice patterns across sites and specialties of care.”

The authors said that training and continuing education efforts with a cultural lens could help address how to collect complete family history.

Next Steps

Dig Deeper on Health IT optimization

xtelligent Rev Cycle Management
xtelligent Patient Engagement
xtelligent Healthtech Analytics
xtelligent Healthtech Security
xtelligent Virtual Healthcare
Close