Data from: A phenome-wide association study of rurality in the All of Us Research Program
Data files
Jul 21, 2026 version files 10.86 MB
-
README.md
10.08 KB
-
RUCA_zipcode_mapping_v2.xlsx
10.32 MB
-
Rurality_mapping_2010.csv
18.55 KB
-
Table_S3_Rural_vs_Urban_noWI_JAMIA_v4.csv
252.02 KB
-
Table_S4_Mixed_vs_Urban_noWI_JAMIA_v4.csv
255.38 KB
Abstract
This study used de-identified medical records with three-digit ZIP code prefix data in the All of Us Research Program to compare health and disease profiles between rural and urban residents in the United States. The rural participants were more likely to have obesity and related medical conditions, including joint problems, high blood cholesterol, high blood pressure, and diabetes. On the other hand, although the rural and urban participants had similar rates for cancer diagnosis, the cancer screening records were less frequent among the rural participants. The rural participants had a slightly higher risk for heart failure. This study suggested that national participant-driven de-identified health record databases such as the All of Us Research Program are useful to understand health and disease conditions of the rural populations and to guide healthcare resources to improve health outcomes in rural areas of the United States.
Objective: Disease profiles and access to care differ between rural and urban populations. The aim of this study is to examine the association of rurality with Electronic Health Records (EHR)-derived disease profiles and heart failure risk.
Materials and Methods: We conducted a retrospective observational study of 242,758 participants enrolled in the All of Us Research Program. Rurality status was estimated using 2010 Rural-Urban Commuting Area codes from participant home three-digit zip code prefixes, and participants were grouped into rural, mixed and urban arms. We compared the disease profiles with phenome-wide association studies (PheWAS) and the risks of heart failure between arms.
Results: The final cohort included 5,581 participants in the rural arm, 62,287 in the mixed arm, and 174,890 in the urban arm. Compared to the urban setting, 70 phecodes were significantly enriched in the rural arm including obesity (odds ratio [OR], 1.42, 95% confidence interval [CI], 1.34 - 1.51) and related diagnoses. Cancer screening-, dermatological-, and eye care-related diagnoses were significantly depleted in the rural arm. The rural arms also showed a greater risk of heart failure (adjusted hazard ratio [HR], 1.20, 95% CI, 1.10-1.31).
Discussion and Conclusion: With the national dataset in All of Us, we found that rural participants had significantly higher risks for obesity-related diseases and heart failure in this study. Depleted phecodes in the rural participants suggested a lack of access to cancer screening, stressing the potential importance of targeted cancer disease management in rural communities.
Dataset DOI: 10.5061/dryad.7sqv9s57m
Description of the data and file structure
Files and variables
File: Rurality_mapping_2010.csv
Description: This file is the rurality mapping with zip3 prefix used in the manuscript. The main exposure studied was the classification of each participant into rural, urban, or mixed (intermediate between rural and urban) arms. Similar to previous studies, we employed the 2010 Rural-Urban Commuting Area (RUCA) codes (as the most recent available version upon the submission of this manuscript), and mapped it to University of Washington Rural Health Research Center's C rural grouping categorization. All of Us includes only home three-digit zip code prefixes to protect privacy. To map the All of Us participant data by three-digit zip code prefix with the RUCA data by five-digit zip codes, we used 2010 Census data to calculate the percentage of rural population within the first three-digit zip code areas and specified 50% or more as the rural arm, >0% to <50% as the mixed arm, and 0% as the urban arm. (Figure 1)
Variables
- Zip3: The three-digit zip code prefix that can be matched to the deidentified All of Us demographic data.
- Rural_Proportion: The population in the rural full five-digit zip code areas divided by the population of the zip 3 prefix area.
- Grouping: Assigned arm in the manuscript.
File: RUCA_zipcode_mapping_v2.xlsx
Description: This file is the calculation of the "Rurality_mapping_2010.csv". It is a reproduced implementation of Figure 1 with Excel. We verified that the “Rurality_final_2010” sheet is identical to the one we used in the manuscript.
Variables
- Useful final assignment for grouping: “Rurality_final_2010” and “Rurality_final_2020” sheets. Both sheets were calculated with 2010 RUCA codes, but with different census data to assign proportions of Zip 5 population to Zip 3 population.
- The most useful calculated sheet that was used in this publication (Rurality_final_2010) is exported as "Rurality_mapping_2010.csv"
Data sheets and fields
- "Data" Sheet: Five-digit zip code level data.
- ZIP_CODE: Five-digit zip code
- STATE: State abbreviations
- ZIP_TYPE: Not used.
- RUCA1: Primary RUCA code.
- RUCA2: Secondary RUCA code (used for categorization).
- zip3: The first three digits of the zipcode, derived from ZIP_CODE
- Categorization: Rural/urban mapped from "Rural_Urban" sheet with RUCA2
- Zip5_pop_2010: Populations of the zip 5 areas according to 2010 Census, looked up from "Census_2010" sheet
- Zip3_pop_2010: Populations of the zip 3 prefix areas according to 2010 Census, looked up from "GroupBy_2010" sheet.
- Proportion_2010: Zip5_pop_2010 / Zip3_pop_2010
- Urban_Proportion_2010: If the zip 5 area is urban, the Proportion_2010 is awarded here.
- Rural_Proportion_2010: If the zip 5 area is rural, the Proportion_2010 is awarded here.
- Zip3_pop_202, Proportion_2020, Urban_Proportion_2020, and Rural_Proportion_2020: Similar to the 2010 counterparts, with 2020 Census data (Census_2020).
- "Rurality_final_2010" Sheet: Final assignment of rurality based on 2010 RUCA code and 2010 census data. (used in the manuscript and exported as Rurality_mapping_2010.csv)
- Fields: See Variables in Rurality_mapping_2010.csv above.
- "Rurality_final_2020" Sheet: Final assignment of rurality based on 2020 RUCA code and 2010 census data.
- Fields: See Variables in Rurality_mapping_2010.csv above.
- Census_2010 and Census 2010:
- Geography: Not used.
- Geographic Area Name: From Census.gov. It contains five digit zip codes.
- Zip5: Five digit zip codes derived from Geographic Area Name.
- Total: Populations of the zip 5 areas.
- GroupBy_2010 and GroupBy_2020:
- Zip3: Zip 3 prefix.
- Zip3_pop: Aggregated populations of the zip 3 areas by summing up the zip 5 areas under the zip 3 areas.
- Rural_Urban: UWRUCA (https://depts.washington.edu/uwruca/ruca-uses.php) Categorization C
- RUCA code description: 2010 RUCA Codes
Calculation
- Step 1: “Rural_Urban” sheet
This mapping is described in UWRUCA (https://depts.washington.edu/uwruca/ruca-uses.php) Categorization C and in the Supplement eTable 2 of Turecamo SE, Xu M, Dixon D, et al. Association of Rurality With Risk of Heart Failure. JAMA Cardiol. 2023;8(3):231-239. doi:10.1001/jamacardio.2022.5211
- Step 2: “Census_2010” (used in the manuscript) and “Census_2020” sheets
The population of each five-digit zip Code area.
Downloaded from:
https://data.census.gov/table/DECENNIALSF12010.P1
https://data.census.gov/table/DECENNIALDHC2020.P1
The “Zip5” column was derived from the “Geographic Area Name” column.
- Step 3: “Data” sheet
2010 Rural-Urban Commuting Area Codes, ZIP code file
Download from: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes
“zip3” column was transformed from the “ZIP_CODE” column.
“Categorization” column was looked up with the “RUCA2” column in sheet “Rural_Urban” with the “RUCA2” column to obtain values from “Categorization” (column 2) using the VLOOKUP function.
“Zip5_pop_2010” column was looked up from the “Zip5” column in sheet “Census_2010” with the “ZIP_CODE” column for values in “Total” (column 2 from the index), with FALSE for range_lookup (exact lookup only), and filled with 0 if the zip code is not present.
“Zip5_pop_2020” column was looked up from “Census_2020” with a similar method.
- Step 4: “GroupBy_2010” and “GroupBy_2020” sheets
“GroupBy_2010” sheet was derived with GROUPBY function was used to group the “Data” sheet by the “zip3” column and aggregate the “Zip5_pop_2010” column with the SUM method.
“GroupBy_2020” sheet was derived with the “Zip5_pop_2020” column in the “Data” sheet with a similar method.
- Step 5: “Data” sheet
“Zip3_pop_2010” and “Zip3_pop_2020” columns were generated using VLOOKUP to look up “zip3_pop” with the “zip3” column in the “Data” sheet.
“Proportion_2010” column was calculated with “Zip_5_pop_2010” divided by “Zip_3_pop_2010”. “Proportion_2020” was calculated similarly.
If the “Categorization” of the row (zip 5 area) is “Rural”, the “Proportion_2010” will be given to “Rural_Proportion_2010”; otherwise, 0 will be given. Similarly, if “Urban”, “Proportion_2010” will be given to “Urban_Proportion_2010”; otherwise, 0 will be given. Similarly, we can calculate the “Rural_Proportion_2020” and “Urban_Proportion_2020” columns.
File: Table_S3_Rural_vs_Urban_noWI_JAMIA_v4.csv
Description: Table S3. The complete phenome-wide association study (PheWAS) result comparing the rural and urban arms. (Participants from the EHR site in Wisconsin were excluded).
Please see: https://github.com/nhgritctran/PheTK for output description.
Variables
- phecode: PheWAS code defined in https://phewascatalog.org/phewas/
- cases: Number of participants who are positive with the phenotype (defined as 2 or more instances on separate days).
- controls: Number of participants who are negative with the phenotype and fulfill the requirement for sex(es) of the phecode.
- p_value: P values from logistic regression.
- neg_log_p_value: log 10 of the p values.
- beta: Natural log of odds ratios.
- conf_int_1: Lower 95% confidence interval of estimated beta.
- conf_int_2: Upper 95% confidence interval of the estimated beta.
- odds_ratio: Odds ratio from logistic regression.
- log10_odds_ratio: Log 10 of the odds ratio.
- converged: Whether the logistic regression is converged.
- phecode_sex_restriction: Whether the phecode is run within one sex or allows both sexes.
- phecode_string: The phenotype description of the phecode.
- phecode_category: Disease category or organ system of the phecode.
File: Table_S4_Mixed_vs_Urban_noWI_JAMIA_v4.csv
Description: Table S4. The complete phenome-wide association study (PheWAS) result comparing the mixed and urban arms. (Participants from the EHR site in Wisconsin were excluded)
Variables
- Same as Table S3.
Access information
Other publicly accessible locations of the data:
- None
Data were derived from the following sources:
-
Supplement eTable 2 of Turecamo SE, Xu M, Dixon D, et al. Association of Rurality With Risk of Heart Failure. JAMA Cardiol. 2023;8(3):231-239. doi:10.1001/jamacardio.2022.5211
-
Census data: https://data.census.gov/table/DECENNIALSF12010.P1
-
2010 Rural-Urban Commuting Area Codes, ZIP code file
Download from: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes
Human subjects data
Given that all data available to researchers in All of Us have personal identifiers removed, the All of Us IRB deemed studies in the Researcher Workbench not to be human participant research and waived the need for additional IRB approval for studies performed in the Researcher Workbench.
The files included in the Dryad submission only contain only summary statistics and do not include row (participant) level data, nor include any counts <= 20, in compliance with All of Us Data and Statistics Dissemination Policy.