Identifying Household Diabetes Risk for Family Diabetes Prevention Using Electronic Health Records
Tainayah W. Thomas, Holly Finertie, Arjun Silverberg, Maher Yassin, James E. Patino, Lisa G. Rosas, June Tester, Luis A. Rodriguez, O. Kenrik Duru, Julie A. Schmittdiel
TL;DR
This study uses health records to assess diabetes risk in households of adults with prediabetes to help prevent diabetes in families.
Contribution
The study introduces a novel approach to diabetes prevention by analyzing household-level risk factors using EHR data.
Findings
Household members of adults with prediabetes show increased diabetes risk.
Shared household environments may contribute to diabetes risk among family members.
Abstract
This cohort study uses electronic health record (EHR) data to evaluate diabetes risk factors among household members of adults with prediabetes.
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
| Characteristic | Adults, No. (%) | ||
|---|---|---|---|
| Overall (N = 356 626) | One-person household (n = 171 615) | Multiperson household (n = 185 011) | |
| Demographic | |||
| Age, mean (SD), y | 50.5 (12.0) | 50.7 (13.4) | 50.4 (10.6) |
| Sex | |||
| Female | 184 297 (51.7) | 95 081 (55.4) | 89 216 (48.2) |
| Male | 172 256 (48.3) | 76 491 (44.6) | 95 765 (51.8) |
| Other or unknown | 73 (0.02) | 43 (0.03) | 30 (0.02) |
| Race and ethnicity | |||
| Asian American | 99 322 (27.9) | 41 982 (24.5) | 57 340 (31.0) |
| Black or African American | 32 298 (9.1) | 19 882 (11.6) | 12 416 (6.7) |
| Hawaiian or Pacific Islander | 4017 (1.1) | 1827 (1.1) | 2190 (1.2) |
| Hispanic or Latino | 94 728 (26.6) | 46 997 (27.4) | 47 731 (25.8) |
| Native American or American Indian | 1422 (0.4) | 685 (0.4) | 737 (0.4) |
| White | 107 517 (30.1) | 51 623 (30.1) | 55 894 (30.2) |
| Multiracial | 2729 (0.8) | 1393 (0.8) | 1336 (0.7) |
| Unknown | 14 593 (4.1) | 7226 (4.2) | 7367 (4.0) |
| Insurance type | |||
| Commercial | 321 654 (90.2) | 137 805 (80.3) | 183 849 (99.4) |
| Medi-Cal | 34 972 (9.8) | 33 810 (19.7) | 1162 (0.6) |
| Membership type | |||
| Adult subscriber | 285 616 (80.1) | 166 895 (97.2) | 118 721 (64.2) |
| Adult dependent | 67 220 (18.8) | 4170 (2.4) | 63 050 (34.1) |
| Child or grandchild | 3610 (1.0) | 519 (0.3) | 3091 (1.7) |
| Disabled dependent | 143 (0.04) | 22 (0.01) | 121 (0.1) |
| Other dependent | 37 (0.01) | 9 (0.005) | 28 (0.02) |
| BMI | 32.0 (6.6) | 32.4 (6.9) | 31.5 (6.2) |
| Overweight | 148 142 (41.5) | 68 965 (40.2) | 79 177 (42.8) |
| Obese grade 1 | 111 892 (31.4) | 52 244 (30.4) | 59 648 (32.2) |
| Obese grade 2 | 54 433 (15.3) | 27 077 (15.8) | 27 356 (14.8) |
| Obese grade 3 | 42 159 (11.8) | 23 329 (13.6) | 18 830 (10.2) |
| Clinical | |||
| HbA1c, mean (SD), % | 5.9 (0.2) | 5.9 (0.2) | 5.9 (0.2) |
| Count of HbA1c tests | 340 363 (95.4) | 164 116 (95.6) | 176 247 (95.3) |
| FPG, mean (SD), mg/dL | 106.1 (5.9) | 106.2 (6.0) | 106.0 (5.8) |
| Count of FPG tests | 16 263 (4.6) | 7499 (4.4) | 8764 (4.7) |
| Household | |||
| Household size (including index participant), No. of persons | |||
| 1 | 171 615 (48.1) | 171 615 (100.0) | 0 |
| 2 | 84 573 (23.7) | 0 | 84 573 (45.7) |
| 3 | 43 234 (12.1) | 0 | 43 234 (23.34) |
| ≥4 | 57 204 (16.0) | 0 | 57 204 (30.9) |
| Household composition (not including index participant) | |||
| Households with ≥1 household member <18 y | NA | NA | 78 217 (42.3) |
| Households with ≥1 household member with diabetes risk factors | NA | NA | 140 398 (75.9) |
| Characteristic | Household members, No. (%) | ||
|---|---|---|---|
| Overall (N = 364 563) | Adult (≥18 y) (n = 238 247) | Child (<18 y) (n = 126 316) | |
| Demographic | |||
| Age, mean (SD), y | 30.7 (20.2) | 41.5 (16.4) | 10.2 (4.8) |
| Sex | |||
| Female | 187 233 (51.4) | 125 765 (52.8) | 61 468 (48.7) |
| Male | 177 257 (48.6) | 112 431 (47.2) | 64 826 (51.3) |
| Other or unknown | 73 (0.02) | 51 (0.02) | 22 (0.02) |
| Race and ethnicity | |||
| Asian American | 106 645 (29.3) | 68 315 (28.7) | 38 330 (30.3) |
| Black | 20 780 (5.7) | 13 594 (5.7) | 7186 (5.7) |
| Hawaiian or Pacific Islander | 4501 (1.2) | 2809 (1.2) | 1692 (1.3) |
| Hispanic or Latino | 95 891 (26.3) | 62 016 (26.0) | 33 875 (26.8) |
| Native American or American Indian | 1353 (0.4) | 968 (0.4) | 385 (0.3) |
| White | 102 034 (28.0) | 74 068 (31.1) | 27 966 (22.1) |
| Multiracial | 6285 (1.7) | 2587 (1.1) | 3698 (2.9) |
| Unknown | 27 074 (7.4) | 13 890 (5.8) | 13 184 (10.4) |
| Membership type | |||
| Adult subscriber | 63 949 (17.5) | 63 945 (26.8) | 0 |
| Adult dependent | 96 858 (26.6) | 96 851 (40.7) | 0 |
| Child or grandchild | 202 813 (55.6) | 76 675 (32.2) | 126 138 (99.9) |
| Disabled dependent | 811 (0.2) | 727 (0.3) | 84 (0.1) |
| Other dependent | 132 (0.04) | 49 (0.02) | 94 (0.1) |
| BMI | 25.5 (7.3) | 28.5 (6.6) | 20.5 (5.5) |
| Normal | 122 428 (34.1) | 53 894 (22.6) | 68 534 (56.6) |
| Overweight | 78 126 (21.7) | 59 971 (25.2) | 18 155 (15.0) |
| Obese | 93 953 (26.1) | 70 804 (29.7) | 23 149 (19.1) |
| Missing | 64 809 (18.0) | 53 578 (22.5) | 11 231 (9.3) |
| Clinical | |||
| HbA1c, mean (SD), % | 5.9 (1.1) | 6.0 (1.2) | 5.5 (0.5) |
| Count of HbA1c tests | 154 288 (42.3) | 137 058 (57.5) | 17 230 (13.6) |
| FPG, mean (SD), mg/dL | 101.3 (35.1) | 102.3 (35.5) | 87.8 (26.3) |
| Count of FPG tests | 36 043 (9.9) | 33 488 (14.1) | 2555 (2.0) |
| No. of primary care visits | |||
| 0 | 87 625 (24.0) | 69 132 (29.0) | 18 493 (14.6) |
| 1 | 98 157 (26.9) | 60 943 (25.6) | 37 214 (29.5) |
| ≥2 | 178 781 (49.0) | 108 172 (45.4) | 70 609 (55.9) |
| No. of glucose screening visits | |||
| 0 | 205 859 (56.5) | 97 443 (40.9) | 108 416 (85.8) |
| 1 | 85 171 (23.4) | 70 469 (29.6) | 14 702 (11.6) |
| ≥2 | 73 533 (20.2) | 70 335 (29.5) | 3198 (2.5) |
| Diabetes risk factors | 179 449 (57.7) | 153 996 (64.6) | 25 453 (35.0) |
| Overweight or obese | 156 101 (50.2) | 130 775 (54.9) | 25 326 (34.8) |
| Gestational diabetes | 1219 (0.5) | 1219 (0.5) | NA |
| Hypertension | 45 278 (19.0) | 45 278 (19.0) | NA |
| Dyslipidemia | 38 416 (16.1) | 38 416 (16.1) | NA |
| Cardiovascular disease | 3848 (1.6) | 3848 (1.6) | NA |
| Prediabetes laboratory result | 48 297 (20.3) | 48 297 (20.3) | 3174 (4.4) |
| Evidence of diabetes | 29 282 (9.4) | 28 997 (12.2) | 285 (0.4) |
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsDiabetes Management and Education · Diabetes, Cardiovascular Risks, and Lipoproteins · Mobile Health and mHealth Applications
Introduction
Diabetes is a transgenerational disease that can affect children, parents, and grandparents.^1^ Because diabetes risk clusters in families,^2^ who often receive care in the same health system, electronic health records (EHRs) could be used to identify and screen coresiding family members to improve diabetes prevention efforts. This study used health system EHR data to identify household members of adults with prediabetes and evaluate their diabetes risk factors, by age.
Methods
We identified an index cohort of adults with prediabetes (fasting plasma glucose, 100-125 mg/dL [to convert glucose to millimoles per liter, multiply by 0.0555] or hemoglobin A_1c_, 5.7%-6.4% [39-46 mmol/mol; to convert to proportion of total hemoglobin, multiply by 0.01]) and body mass index (BMI) of 25 or more (calculated as weight in kilograms divided by height in meters squared) within 1 year prior to January 1, 2023, through December 31, 2023. The index date was the earliest date in 2023 on which patients met inclusion criteria. We excluded patients with diabetes, end-stage kidney disease, or less than 12 months’ health system membership within 1 year before index date. We identified coinsured members living with the index member at index date through the index member’s subscriber medical record numbers and address history. This study was approved by the UCLA institutional review board, and a waiver of informed consent was obtained due to the minimal risk to human participants. This observational cohort study followed the STROBE reporting guideline for cohort studies.
Diabetes risk factors were calculated for those aged 10 years or older. For adults, risk factors were defined as BMI of 25 or more or a history of gestational diabetes, hypertension, dyslipidemia, or cardiovascular disease or prediabetes or diabetes. For children aged 10 to 17 years, risk factors were age- and sex-specific^3^ overweight or obesity or prediabetes or diabetes. Descriptive statistics were output for the index cohort and household members. Patients self-reported race and ethnicity. Analyses were performed in SAS, version 9.4 and R, version 4.3.1. Detailed methods are provided in the eMethods in Supplement 1.^4^
Results
The index cohort included 356 626 adults with prediabetes (mean [SD] age, 50.5 [12.0] years; 51.7% women and 48.3% men; 27.9% Asian American, 9.1% Black or African American, 1.1% Hawaiian or Pacific Islander, 26.6% Hispanic or Latino, 0.4% Native American or American Indian, 30.1% White, 0.8% multiracial, and 4.1% unknown (Table 1). Obesity was present in 58.5% of the index cohort. Household composition varied (48.1% 1-person households and 51.9% multiresident households). A total of 75.9% of multiresident households had at least 1 additional household member with diabetes risk factors.
We identified 364 563 coresiding household members (Table 2). The mean (SD) age was 41.5 (16.4) years for adult household members and 10.2 (4.8) years for child household members. Diabetes risk factors were identified in 64.6% of adults and 35.0% of children (aged 10-17 years); overweight or obesity was the most identified risk factor (adults, 54.9%; children, 34.8%). Abnormal blood glucose was present in 32.5% of adult household members, with 20.3% having prediabetes laboratory test results and 12.2% having evidence of diabetes. Prediabetes was most prevalent among Asian American adult household members (26.0%), and diabetes was most prevalent among Hawaiian or Pacific Islander adult household members (15.6%).
Discussion
In a cohort of patients with prediabetes and their household members, 75.9% of multiresident households had at least 1 additional household member with diabetes risk factors. Limitations include generalizability of EHR data to the general population, potentially incomplete risk factor assessment for household members that may underestimate risk factors, and ability to identify only coinsured household members, which may underestimate household size and overestimate the proportion of family members with risk factors.
Our study highlights that EHR data can be used to identify households at risk for diabetes. Health systems could use EHRs to screen family members for risk factors and support care coordination among families for cardiometabolic risk reduction. Family care coordination could reduce health system costs; however, few prevention programs enroll households at risk, reflecting a missed opportunity for population-level diabetes prevention.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Epstein LH, Wilfley DE, Egede LE. Transgenerational clinical care—the case for family-based treatment. JAMA Pediatr. 2025;179(2):120-121. doi:10.1001/jamapediatrics.2024.5904 39714810 · doi ↗ · pubmed ↗
- 2Aasbjerg K, Nørgaard CH, Vestergaard N, . Risk of diabetes among related and unrelated family members. Diabetes Res Clin Pract. 2020;160:107997. doi:10.1016/j.diabres.2019.107997 31901471 · doi ↗ · pubmed ↗
- 3Hales CM, Freedman DS, Akinbami L, Wei R, Ogden CL. Evaluation of alternative body mass index (BMI) metrics to monitor weight status in children and adolescents with extremely high BMI using CDC BMI-for-age growth charts. Vital Health Stat 1. 2022;(197):1-42. 36598420 · pubmed ↗
- 4Moffet HH, Adler N, Schillinger D, . Cohort profile: the Diabetes Study of Northern California (DISTANCE)—objectives and design of a survey follow-up study of social health disparities in a managed care population. Int J Epidemiol. 2009;38(1):38-47. doi:10.1093/ije/dyn 040 18326513 PMC 2635421 · doi ↗ · pubmed ↗