Comparing medical history data derived from electronic health records and survey answers in the All of Us Research Program

Abstract

Objective A participant’s medical history is important in clinical research and can be captured from electronic health records (EHRs) and self-reported surveys. Both can be incomplete, EHR due to documentation gaps or lack of interoperability and surveys due to recall bias or limited health literacy. This analysis compares medical history collected in the All of Us Research Program through both surveys and EHRs. Materials and Methods The All of Us medical history survey includes self-report questionnaire that asks about diagnoses to over 150 medical conditions organized into 12 disease categories. In each category, we identified the 3 most and least frequent self-reported diagnoses and retrieved their analogues from EHRs. We calculated agreement scores and extracted participant demographic characteristics for each comparison set. Results The 4th All of Us dataset release includes data from 314 994 participants; 28.3% of whom completed medical history surveys, and 65.5% of whom had EHR data. Hearing and vision category within the survey had the highest number of responses, but the second lowest positive agreement with the EHR (0.21). The Infectious disease category had the lowest positive agreement (0.12). Cancer conditions had the highest positive agreement (0.45) between the 2 data sources. Discussion and Conclusion Our study quantified the agreement of medical history between 2 sources—EHRs and self-reported surveys. Conditions that are usually undocumented in EHRs had low agreement scores, demonstrating that survey data can supplement EHR data. Disagreement between EHR and survey can help identify possible missing records and guide researchers to adjust for biases.

Comparing ancestry calibration approaches for a trans-ancestry colorectal cancer polygenic risk score

“…As we used whole genome sequence data, differences in OR across ancestries cannot be attributed to uncertainty in imputation of genotypes, but rather solely to differences in LD with causal variation. However, the availability of EHR data, and therefore the number of cases and controls, may be associated with self-identified race or ancestry, which further complicates interpretation of the results 23 .…”

“…The AOU dataset contains a rich database of ICD 9 and 10 codes, CPT codes, and Logical Observation Identifiers Names and Codes (LOINC). As the completeness of this data varies among participants 25 we attempted to widen our net by creating concept sets based on the original ICD 9 and CPT codes, incorporating the related ICD 10 codes and LOINC codes. Each concept set relates to a different table from the word files provided by the algorithm’s authors (Supplemental Tables S2-S10).…”

Kidney stone prevalence based on self-report and Electronic Health Records: insight into the prevalence of active medical care for kidney stones

“…A recent study of the All of Us research program overall found that strength of concordance between EHR-based diagnosis and survey data for diseases vary. 19 The authors suggest that survey data overall provides a useful avenue to augment EHR data through identifying missing records or diagnoses, and addressing biases in EHR data collection.…”

“…However, a prior publication examining the concordance of self-reported and EHR data in All of Us showed that traditionally underreported EHR diagnoses such as myopia are low, while traditionally highly reported diagnoses such as cancer are high, suggesting that the data are appropriately collected. 19 Furthermore, issues with missing data could represent a systematic bias in EHR based kidney stone prevalence assessment globally. Finally, income and education were used as surrogates for financial ability, but net worth, which is likely a better indicator of socioeconomic status in this older population, could not be assessed.…”

Protocol for developing a personalised prediction model for viral suppression among under-represented populations in the context of the COVID-19 pandemic

“…The AoU medical history survey includes a self-report questionnaire about diagnoses of over 150 medical conditions organised into 12 disease categories 23. We will use a combination of self-reported responses to the medical history survey and data from diagnosis codes in the EHR data to ascertain the presence of all comorbidities, such as cardiovascular risk factors, including hypertension (OMOP code 316866), hyperlipidaemia (OMOP code 432867) and type 2 diabetes mellitus (OMOP code 201826), and use self-reported data from the lifestyle survey to ascertain smoking status.…”