Extended Data Table 3 Diabetes Sample Characteristics
From: Access to care affects electronic health record reliability and AI-driven disease prediction
- Sample includes All of Us participants that answered the Healthcare and Access survey and had linked EHR data. Survey responses were used to classify participants into three access groups: participants with cost-constrained and/or delayed care indicated they had trouble affording care and/or had delayed care in the past 12 months (participants can be represented in both the cost-constrained and delayed care groups), and participants that indicated neither were classified as having standard care. All sample characteristic information is based on self-reported information found in the All of Us baseline survey. The Other Race category includes people who answered, ‘More than one population’,’None of these’,’Middle Eastern or North African’,’Native Hawaiian or Other Pacific Islander’. Unknown race and ethnicity categories include participants who skipped or selected that they preferred not to answer. AI/AN = American Indian or Alaska Native. Medicare insurance includes participants that are dually enrolled in Medicaid and those with supplemental insurance. Other insurance includes participants with directly purchased health insurance, TRICARE, Veterans Affairs (VA), and other military insurance. Those with Indian Health Service (IHS) were included in the uninsured category following the categorization used by the American Community Survey.
- Source data
