Prevalence and Correlates of Social Risks and Receipt of Resources Among Rural and Urban Veterans

ABSTRACT

Purpose

Rural Veterans face barriers to accessing clinical care and community‐based services, often exacerbated by social needs. Assessing Circumstances and Offering Resources for Needs (ACORN) is a Veterans Affairs (VA) healthcare system, Veteran‐tailored intervention to screen for social risks (e.g., food insecurity, social isolation) and address social needs. We examined rural–urban differences in social risks and receipt of resources/referrals.

Methods

This was a retrospective, cross‐sectional evaluation using VA administrative data for Veterans screened with ACORN between January 2024 and September 2025. Veterans were screened for social risks across nine domains and offered resources/referrals to address needs. Multivariable logistic regression models were estimated to identify associations between rurality, social risks, and receipt of resources/referrals.

Findings

During the evaluation period, 92,017 Veterans across 87 VA sites were screened with ACORN; 29.37% were rural (n = 28,346). Overall, 60.30% of Veterans reported one or more social risks (55.65% rural vs. 62.37% urban, p < 0.001). Rural Veterans were less likely to endorse all social risks except for digital needs (21.16% vs. 20.27%, p = 0.002). In adjusted models, rural Veterans had lower odds of reporting one or more social risks (aOR, 0.85; 95% CI, 0.76–0.94). Among those screening positive, rural Veterans were less likely to receive resources/referrals (aOR, 0.89; 95% CI, 0.85–0.94) or report already receiving services/assistance (aOR, 0.85; 95% CI, 0.76–0.95), and more likely to decline assistance (aOR, 1.27; 95% CI, 1.17–1.37).

Conclusions

Rural Veterans were less likely to report social risks than urban Veterans. However, among those screening positive, rural Veterans were less likely to receive assistance. Future work is needed to understand rural Veterans’ experiences with social risks and differences in receipt of resources/referrals.

1. Introduction

Nearly one‐quarter of US Veterans, approximately 4.4 million individuals, reside in rural areas. Among these, 2.7 million are enrolled in the Department of Veterans Affairs (VA) healthcare system, the nation's largest integrated healthcare system [1]. Veterans in rural areas often face significant barriers to accessing healthcare services and community‐based services, which are often more readily available in urban areas. These barriers include geographic isolation, long travel distances to services, provider shortages, and reduced access to primary care, specialty care, and mental health services [2, 3, 4, 5, 6]. Such obstacles contribute to well‐documented health disparities between rural and urban Veterans. Specifically, rural Veterans report lower health‐related quality of life and experience higher morbidity, poorer chronic disease management, and increased rates of mortality, including significantly higher suicide rates, compared to their urban counterparts [7, 8, 9, 10, 11, 12, 13].

VA has dedicated substantial staffing, infrastructure, and resources to address these disparities [10, 14]. Key initiatives include the development of Clinical Resource Hubs, which utilize a hub‐and‐spoke model to provide primary care, mental health, and other specialty care services both in person and via telehealth to enhance access to care and bridge service gaps for rural Veterans [15, 16]. The VA has launched several other innovative telehealth initiatives to improve care access for rural Veterans, such as Accessing Telehealth through Local Area Stations (ATLAS) [17], Clinical Video Telehealth (CVT) [18], and Geriatric Research Education and Clinical Center (GRECC) Connect [19]—but barriers persist.

Access barriers and resulting health disparities for rural Veterans are compounded by a high burden of social risks and health‐related social needs (herein “social needs”) [20, 21, 22, 23]. Social risks are defined as social, physical, and economic conditions (e.g., food insecurity, social isolation, lack of access to transportation) that influence health and well‐being, whereas social needs refer to an individual's perception and prioritization of their needs, as well as interest in offered assistance [24]. Social risks and social needs are associated with a range of adverse health outcomes and reduced access to care among rural Veterans, including poorer physical and mental health, as well as higher odds of missed appointments and medication nonadherence [25, 26].

While prior work has examined survey‐based prevalence of social risks and social needs among rural Veterans [20, 26], there remains a knowledge gap related to the identification of social risks among rural Veterans as part of VA clinical care, receipt of resources/referrals for reported social needs, and potential opportunities for better tailoring screening and intervention strategies to the unique needs and experiences of rural Veterans. Assessing Circumstances and Offering Resources for Needs (ACORN) [27] is a national VA, Veteran‐tailored social risk screening and follow‐up intervention that, as of September 2025, had been implemented at nearly 63% (n = 87) of VA medical centers and their associated community‐based outpatient clinics (herein collectively referred to as “sites”) across a variety of clinical settings (Table 1). We leveraged ACORN administrative data to examine rural–urban differences in the prevalence and correlates of social risks and receipt of resources/referrals to address identified needs.

TABLE 1.

ACORN sites included in analytic sample, by ACORN national EHR template implementation date.

Site number a	Rurality	US census region	ACORN national EHR template screening start date	Unique Veterans screened with ACORN
1	Rural	Midwest	January 2024	4099
2	Rural	Midwest	January 2024	1066
3	Rural	Midwest	January 2024	962
4	Urban	Midwest	January 2024	3163
5	Rural	Northeast	January 2024	1105
6	Urban	Northeast	January 2024	3053
7	Urban	Northeast	January 2024	1321
8	Rural	South	January 2024	1214
9	Urban	South	January 2024	5507
10	Urban	South	January 2024	401
11	Urban	South	January 2024	327
12	Urban	South	January 2024	91
13	Urban	Midwest	February 2024	3517
14	Urban	Midwest	February 2024	1888
15	Urban	Midwest	February 2024	808
16	Urban	Midwest	February 2024	166
17	Rural	Northeast	February 2024	339
18	Urban	Northeast	February 2024	3652
19	Urban	Northeast	February 2024	1897
20	Urban	Northeast	February 2024	747
21	Urban	Northeast	February 2024	714
22	Urban	Northeast	February 2024	275
23	Urban	Northeast	February 2024	110
24	Urban	Northeast	February 2024	3
25	Rural	South	February 2024	1954
26	Rural	South	February 2024	1145
27	Urban	South	February 2024	6622
28	Urban	South	February 2024	1610
29	Urban	South	February 2024	1492
30	Urban	South	February 2024	675
31	Urban	South	February 2024	530
32	Urban	South	February 2024	107
33	Urban	West	February 2024	1376
34	Urban	West	February 2024	1166
35	Urban	West	February 2024	765
36	Urban	West	February 2024	48
37	Urban	West	February 2024	21
38	Urban	Northeast	March 2024	1452
39	Urban	Northeast	March 2024	229
40	Urban	Northeast	March 2024	95
41	Rural	South	March 2024	1547
42	Urban	South	March 2024	9282
43	Urban	West	March 2024	14
44	Urban	Northeast	April 2024	1015
45	Urban	Northeast	April 2024	380
46	Urban	West	April 2024	2627
47	Urban	West	April 2024	1746
48	Rural	Midwest	May 2024	622
49	Urban	Midwest	May 2024	1189
50	Urban	South	May 2024	106
51	Urban	West	May 2025	724
52	Urban	West	May 2024	148
53	Urban	South	June 2024	749
54	Urban	South	June 2024	685
55	Rural	Midwest	July 2024	2464
56	Urban	Midwest	July 2024	127
57	Rural	Northeast	July 2024	634
58	Urban	Northeast	July 2024	149
59	Rural	South	July 2024	2184
60	Urban	South	July 2024	4263
61	Urban	South	July 2024	2808
62	Urban	South	July 2024	1717
63	Urban	South	July 2024	1030
64	Rural	West	July 2024	328
65	Rural	Midwest	August 2024	352
66	Urban	South	August 2024	248
67	Urban	West	August 2024	587
68	Rural	Northeast	September 2024	112
69	Urban	South	September 2024	434
70	Urban	West	September 2024	32
71	Rural	Midwest	November 2024	334
72	Urban	South	November 2024	539
73	Urban	Midwest	December 2024	74
74	Urban	South	December 2024	654
75	Urban	South	December 2024	298
76	Rural	Midwest	January 2025	119
77	Rural	South	February 2025	147
78	Urban	Outlying Area	February 2025	124
79	Urban	South	February 2025	94
80	Urban	Midwest	March 2025	18
81	Rural	Northeast	April 2025	25
82	Urban	Midwest	April 2025	295
83	Urban	Midwest	June 2025	5
84	Urban	South	July 2025	19
85	Urban	South	July 2025	2
86	Rural	South	September 2025	1
87	Rural	West	September 2025	1

^a Sites are numbered for anonymity.

2. Methods

As part of ongoing quality improvement for ACORN, this work was determined by the VA Providence Healthcare System institutional review board to not require regulatory review.

2.1. Program Description

ACORN is a national VA social risk screening and follow‐up intervention to systematically screen for, assess, and address social risks and social needs among Veterans; it is conducted in partnership with the VA Office of Health Equity and the VA National Social Work Program, Care Management and Social Work Services [27]. ACORN's core components consist of (1) administration of a standardized screening tool to identify social risks across nine domains (housing instability, food insecurity, utility insecurity, transportation insecurity, legal needs, social isolation/loneliness, employment needs, educational assistance, and digital needs) and (2) the provision of resources and referrals to address identified needs, including a mechanism to address urgent needs at the time of screening [27]. When a Veteran screens positive for one or more social risks, staff offer a spectrum of low‐touch to high‐touch interventions, including referrals or warm handoffs to social work, mental health, nutrition, and/or other relevant VA clinical service lines; support with resource navigation including referrals for ongoing case management, if indicated; and/or provision of geographically tailored resource guides with VA and community services [27, 28, 29, 30]. Staff document resources/referrals provided. Veterans are presumed to have social needs within domains in which they accept assistance, though the distinction between social risks and needs is not formally captured in the screening tool.

As described in detail elsewhere [27, 29, 31], ACORN was first developed in 2018 by an interprofessional team in the VA New England Healthcare System. Candidate social risk domains and corresponding measures were selected based on expert body recommendations and established screening tools [27, 28]. For social risk domains without existing measures, or where existing measures were not applicable to the Veteran population or VA clinical settings, questions were adapted or newly developed [27, 28]. The screening tool was then iteratively refined based on cognitive interviews with Veterans and field testing [27, 29, 31]. In 2020, the ACORN team collaborated with the VA Office of Connected Care to develop and pilot measures for a new digital needs domain, which was subsequently incorporated in the ACORN screening tool [32]. An ACORN electronic health record (EHR) template was first developed in May 2020 for use by participating sites. A revised version of this template was subsequently approved and made available for VA‐wide utilization in January 2024 [29]. Although most of the social risk measures remained unchanged from the original version of the EHR template (May 2020) to the ACORN national EHR template (January 2024), some measures included minor wording refinements based on feedback from staff and subject matter experts. The digital needs domain also underwent substantial revision, including the removal of two measures. Other notable modifications to the template included the addition of an “ACORN Prior Screening Results” button where staff could review dates, positive responses, and resources/referrals provided for prior ACORN screenings; the addition of a question to capture whether the screening tool was staff‐administered or Veteran self‐administered; and the option for staff to document relevant International Statistical Classification of Diseases, Tenth Revision (ICD‐10) Z‐codes for inclusion.

The ACORN intervention has been implemented nationally in a variety of outpatient, emergency department, and inpatient settings. It is administered by interprofessional care team members across a range of staff roles and disciplines, including social workers, nurses, Whole Health coaches [33], and peer specialists. To accommodate differences in local context, leadership priorities, staffing capacity, and resource availability, sites have the flexibility in determining which staff roles and disciplines administer ACORN, the clinical setting(s) in which screening occurs, and the populations screened. Screening approaches range from universal screening of all Veterans presenting for care in a given clinical setting to targeted screening of specific populations, such as new patients or older Veterans [31, 32]. Consequently, uptake and screening volume vary substantially across sites due to differences in implementation strategies, site complexity, patient volume, and clinical setting.

2.2. Data Sources and Study Population

Sociodemographic data, clinical characteristics, and structured data from the ACORN screening tool (i.e., social risks and resources/referrals provided) were obtained from the VA Corporate Data Warehouse (CDW), a repository of VA EHR data from clinical and administrative databases. Our cohort consisted of all Veterans screened with the ACORN national EHR template between the date of the template's release, January 16, 2024, and September 30, 2025. Analyses were limited to screening data from the national template to ensure the consistency of measures across screens. We excluded observations for Veterans missing rurality information (n = 4503). We also excluded an additional 37 Veterans who had duplicate ACORN screens listed at the exact same date/time but with varying responses. Most Veterans (91.24%) were screened only once during the study period. For Veterans screened more than once, we created an index screen, defined as their first positive ACORN screen (i.e., screened positive for one or more social risk domains), or, if they never screened positive, then their first screen; thus, each individual appeared in the sample once [32, 34, 35]. Our final analytic sample included 92,017 Veterans.

By the end of our evaluation period, ACORN had been implemented in a geographically diverse group of 87 sites across 38 states and Puerto Rico. All sites that used the ACORN national EHR template between January 16, 2024, and September 30, 2025, were included in our analysis, regardless of their implementation start date. Of sites that administered ACORN during the study period, 22 (25.29%) were considered rural based on the VA Office of Rural Health's “Rurality Calculator,” which classifies sites as rural if they serve at least 50% rural‐residing Veterans [36]. These 22 rural sites administering ACORN represent half of all rural sites in VA [36, 37].

2.3. Measures

2.3.1. Social Risks and Resources/Referrals Provided

The ACORN screening tool includes 13 core questions (with up to five response‐dependent follow‐up questions) across nine domains: housing instability, food insecurity, utility insecurity, transportation insecurity, legal needs, social isolation/loneliness, employment needs, educational assistance, and digital needs (including device access, internet access, and digital health literacy) [27, 28]. The ACORN national EHR template also includes a “disposition” section in which staff can document a range of resources/referrals provided following screening. Staff may also document reasons resources were not provided (if applicable), including nonmutually exclusive response options of “Veteran already receiving services or assistance,” “Veteran declines assistance at this time,” and “screened negative for all needs” [28, 29, 30, 31]. Among Veterans screening positive for one or more social risks, we assessed whether any resources/referrals were provided (dichotomized as yes/no), and if Veterans reported already receiving services/assistance and/or if they declined assistance.

Since the EHR template requires completion of at least one disposition field regardless of screening result, the ACORN team created a “screened negative for all needs” option for staff to indicate that no resources/referrals were provided because the Veteran screened negative across all nine ACORN domains. However, some Veterans screening positive for one or more social risk domains were also documented as “screened negative for all needs.” To capture these instances, we created a discordant response category, defined as a positive ACORN screen paired with documentation of “screened negative for all needs.” Based on structured EHR data alone, it is not possible to determine whether these discordant responses reflected Veterans screening positive for a social risk but not perceiving it as a need or wanting resources, or staff misunderstanding which screening responses (including “sometimes” responses on often/sometimes/never scales) constituted a positive screen. Disposition data were coded as missing if no disposition field was completed.

2.3.2. Sociodemographic Characteristics

Sociodemographic characteristics included age at the time of screening, sex (male/female), race (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, and White), Hispanic or Latino ethnicity (yes/no), relationship status (married/partnered; divorced/widowed/separated; and single/never married), enrollment priority group (1‐8), rurality (urban vs rural/highly rural), and if a Veteran served in combat (yes/no). Veteran rurality was identified using a standardized VA definition based on rural‐urban commuting area (RUCA) codes [37, 38]. Veteran rurality is grouped into three categories: urban (i.e., RUCA code of 1.0 or 1.1), rural (i.e., RUCA code of 2–9 or 10.1–10.3), and highly rural (i.e., RUCA code of 10.0) [37], and was based on the Veteran's geocoded address at the time of screening. Given relatively low numbers of Veterans living in highly rural areas in our sample, which is consistent with the VA population overall, we collapsed rural and highly rural into a single group, hereafter “rural.” In reporting race and ethnicity data, we used the “alone or in combination” approach described in the US Office of Management and Budget reporting standards [39], such that categories are not mutually exclusive. We additionally created a separate “multiple race” category to identify Veterans reporting more than one race while still preserving all individual race responses, such that responses sum to more than 100%. Although race and ethnicity in the CDW are primarily based on Veteran self‐reported data collected during VA enrollment or registration for healthcare encounters and entered by health care staff, these data are occasionally based on staff “observation” [40]. Enrollment priority group refers to Veterans’ service‐connected disability rating and service‐connected disability compensation, as well as eligibility for and cost share associated with VA health benefits and services [41]. We collapsed the eight enrollment priority groups into four categories: >0% service‐connected disability (groups 1–3); aid and attendance or housebound benefits recipient, or catastrophically disabled (group 4); no service‐connected disability or non‐compensable service‐connected disability (0%) with an annual income below VA's adjusted income limit, VA pension benefits recipient, or Medicaid‐eligible (group 5); and Veterans in all other priority groups, which generally include higher‐income Veterans (groups 6–8) [42, 43].

2.3.3. Clinical Characteristics

Clinical characteristics included Veterans’ most recent Care Needs Assessment (CAN) score (using CAN 2.5), a VA‐validated risk prediction model based on sociodemographic and medical characteristics, as well as healthcare utilization. Scores in the 95^th percentile or greater indicate a high risk of hospitalization or mortality in the next 90 days [44, 45, 46, 47]. We also included a chronic conditions count comprised of 48 chronic conditions (37 physical and 11 mental health and substance use conditions) identified based on the presence of ICD‐10 codes in VA outpatient or inpatient encounters in the prior year (see Supporting Information Appendix Table 1 for a list of included conditions) [48, 49]. We grouped the chronic condition count based on percentile spread into 0, 1–3, 4–6, and 7+.

2.4. Statistical Analysis

All analyses were conducted using R Studio with R Version 4.4.3 64‐bit (Vienna, Austria).

Means and standard deviations were calculated for continuous variables, and bivariate differences were examined using the Kruskal–Wallis rank sum test. Frequencies and percentages of categorical variables were compared using Pearson's Chi‐squared tests. Veteran characteristics were summarized overall and stratified by urban vs. rural status. Statistical significance was set at p < 0.05. Using the lme4 package in R, we fit separate multivariable logistic regression models with site‐level random effects to estimate the association between rurality and (1) screening positive for one or more social risks, (2) each individual social risk domain, and (3) four disposition outcomes among Veterans screening positive. Disposition outcomes included (1) receipt of any resource/referral, (2) Veteran reporting already receiving services/assistance, (3) Veteran declining assistance, and (4) documentation that the Veteran “screened negative for all needs” despite a positive ACORN screen (i.e., discordant response documented).

Site‐level random effects were included to account for unobserved site‐level variability, while preserving rurality effects. Urban Veterans served as the reference group, and regression coefficients are represented as odds ratios (ORs) with 95% confidence intervals (CIs). All models were adjusted for age (reference:18–34 years), sex (male), race (modeled as separate binary indicators because Veterans could select multiple races), ethnicity (not Hispanic or Latino), relationship status (married/partnered), enrollment priority group (groups 6–8), combat status (no), CAN score >95 (no), and chronic conditions count (0). Complete case analyses were used for all models. We additionally conducted a sensitivity analysis for disposition outcome models that further adjusted for the number of social risks endorsed to assess whether associations between rurality and receipt of resources/referrals varied after accounting for the overall social risks burden. Only main effects are presented.

3. Results

Among 92,017 Veterans screened, the mean [SD] age was 65 [16.49] years, 13,980 (15.19%) were female, 22,199 (24.12%) were Black or African American, 5681 (6.17%) were Hispanic or Latino, and 60,634 (65.89%) were White. Overall, 40,677 (44.21%) Veterans were married/partnered, 59,956 (65.16%) had a service‐connected disability, and 20,892 (22.70%) served in combat. Mean [SD] number of chronic conditions was 4.23 [3.58], and 4623 (5.02%) Veterans had a CAN score ≥95 (Table 2).

TABLE 2.

Sociodemographic characteristics of Veterans screened with ACORN between January 2024 and September 2025, overall and by rurality.

	Overall (N = 92,017)		Urban (n = 63,671)		Rural (n = 28,346)		p‐value
	#	%	#	%	#	%
Age at encounter (mean, SD)	65.59 (16.49)		65.23 (16.67)		66.41 (16.05)		<0.001
Age groups a							<0.001
18–34	5582	6.07	4155	6.53	1427	5.03
35–49	11,927	12.96	8336	13.09	3591	12.67
50–64	20,348	22.11	14,370	22.57	5978	21.09
65–79	38,645	42.00	26,143	41.06	12,502	44.10
80+	15,515	16.86	10,667	16.75	4848	17.10
Sex a							<0.001
Female	13,980	15.19	10,047	15.78	3933	13.87
Male	78,037	84.81	53,624	84.22	24,413	86.13
Race a , b
American Indian or Alaska Native	1311	1.42	643	1.01	668	2.36	<0.001
Asian	713	0.77	615	0.97	98	0.35	<0.001
Black or African American	22,199	24.12	18,526	29.10	3673	12.96	<0.001
More than one race c	917	1.00	665	1.04	252	0.89	0.028
Native Hawaiian or Pacific Islander	862	0.94	645	1.01	217	0.77	<0.001
White	60,634	65.89	38,707	60.79	21,927	77.35	<0.001
Race unknown/missing	7264	7.89	5245	8.24	2019	7.12	<0.001
Ethnicity a
Hispanic or Latino	5681	6.17	4686	7.36	995	3.51	<0.001
Not Hispanic or Latino	80,769	87.78	55,053	86.46	25,716	90.72	<0.001
Ethnicity unknown/missing	5566	6.05	3931	6.17	1635	5.77	0.017
Relationship status a							<0.001
Divorced/separated/widowed	32,751	35.59	23,243	36.50	9508	33.54
Married/partnered	40,677	44.21	26,075	40.95	14,602	51.51
Single/never married	16,027	17.42	12,499	19.63	3528	12.45
Missing	2562	2.78	1854	2.91	708	2.50
Enrollment priority groups a , d							<0.001
≥10% service connected (1–3)	59,956	65.16	41,030	64.44	18,926	66.77
Catastrophically disabled (4)	1748	1.90	1274	2.00	474	1.67
Non‐service connected and low income (5)	18,671	20.29	13,220	20.76	5451	19.23
Non‐service connected and not low income (6–8)	11,504	12.50	8045	12.64	3459	12.20
Missing	138	0.15	102	0.16	36	0.13
Combat status a							<0.001
Yes	20,892	22.70	14,245	22.37	6647	23.45
No	71,125	77.30	49,426	77.63	21,699	76.55
Chronic conditions count e (mean, SD)	4.23 (3.58)		4.27 (3.60)		4.14 (3.52)
Chronic conditions count a , e
0	20,095	21.84	13,694	21.51	6401	22.58
1–3	22,490	24.44	15,578	24.47	6912	24.38
4–6	26,495	28.79	18,305	28.75	8190	28.89
7+	22,937	24.93	16,094	25.28	6843	24.14
CAN score a , f
≥95	4623	5.02	3293	5.17	1330	4.69
≤95	87,394	94.98	60,378	94.83	27,016	95.31

^a Variables are included as covariates in adjusted analyses.

^b Race is reported as “alone or in combination” (39) except for “more than one race,” so responses sum to more than 100%.

^c Created by the study team to denote if Veterans selected more than one race.

^d Enrollment priority group refers to Veterans’ service‐connected disability rating and service‐connected disability compensation (e.g., service‐connected disability rated as 50% disabling), as well as eligibility for and cost share associated with VA health benefits and services (41). Enrollment priority groups represent: 0% service‐connected disability (groups 1–3); aid and attendance or housebound benefits recipient, or catastrophically disabled (group 4); no service‐connected disability or non‐compensable service‐connected disability (0%) with an annual income below VA's adjusted income limit, VA pension benefits recipient, or Medicaid‐eligible (group 5); and Veterans in all other priority groups, which generally include higher‐income Veterans (groups 6–8).

^e Forty‐eight chronic conditions (37 physical and 11 mental health and substance use conditions) identified based on the presence of International Classification of Diseases, Tenth Revision, Clinical Modification (ICD‐10) codes in VA outpatient or inpatient encounters in the prior year.

^f Clinical characteristics included Veterans’ most recent Care Needs Assessment (CAN) score (using CAN 2.5), a VA‐validated risk prediction model based on sociodemographic and medical characteristics, as well as healthcare utilization. Scores in the 95th percentile or greater indicate a high risk of hospitalization or mortality in the next 90 days (44, 45, 46, 47).

Nearly one‐third of Veterans (28,346, 30.81%) resided in rural areas. Rural Veterans were slightly older (66 [SD 16.05] years vs. 65 [16.67] years), more likely to be male (86.13% vs. 84.22%), service‐connected (66.77% vs. 64.44%), and married/partnered (51.51% vs. 40.95%), as well as more likely to have served in combat compared to urban Veterans (23.44% vs. 22.43%) (all p < 0.001). Rural Veterans were also likely to be Black or African American (12.96% vs. 29.10%; p < 0.001) or Hispanic or Latino (3.51% vs. 7.36%; p = 0.028). Rural Veterans also had a slightly lower chronic conditions count (4.14 vs. 4.27) (p < 0.001) and were less likely to have a CAN score ≥95 (4.69% vs 5.17%; p = 0.002) (Table 2).

3.1. Reported Social Risks

Overall, 55,486 Veterans (60.30%) screened positive for one or more social risks. The prevalence of social risks ranged from 28.90% for social isolation/loneliness to 6.71% for employment needs. Rural Veterans were more likely to report having no social risks (44.19% vs. 37.16%) or only one social risk (29.14% vs. 27.60%) compared to their urban counterparts (all p < 0.001). Rural Veterans were significantly more likely to endorse digital needs (21.24% vs. 20.58%; p = 0.002), but they were less likely to report social risks across all other domains: social isolation/loneliness (26.74% vs. 30.26%), food insecurity (13.18% vs. 18.27%), transportation insecurity (11.43% vs. 17.47%), utility insecurity (10.77% vs. 13.35%), educational assistance (7.08% vs. 9.48%), housing instability (7.05% vs. 12.92%), legal needs (6.10% vs. 8.79%), and employment needs (4.35% vs. 7.87%) (all p < 0.001) (Table 3).

TABLE 3.

Positive social risk domains among Veterans screened with ACORN between January 2024 and September 2025, overall and by rurality.

	Overall (N = 92,017)		Urban (n = 63,671)		Rural (n = 28,346)
	#	%	#	%	#	%	p‐value
Positive ACORN screen	55,486	60.30	39,712	62.37	15,774	55.65	<0.001
Digital Needs	18,904	20.54	12,906	20.27	5998	21.16	0.002
Educational Assistance	7828	8.51	5866	9.21	1962	6.92	<0.001
Employment Needs	6178	6.71	4957	7.79	1221	4.31	<0.001
Food Insecurity	15,091	16.40	11,402	17.91	3689	13.01	<0.001
Acute Food Needs	3906	4.24	3006	4.72	900	3.18	<0.001
Housing Instability	10,141	11.02	8161	12.82	1980	6.99	<0.001
Acute Housing Needs	1966	2.14	1622	2.55	344	1.21	<0.001
Legal Needs	7291	7.92	5557	8.73	1734	6.12	<0.001
Social Isolation/Loneliness	26,596	28.90	19,009	29.86	7587	26.77	<0.001
Transportation Insecurity	14,186	15.42	10,962	17.22	3224	11.37	<0.001
Utility Insecurity	11,409	12.40	8359	13.13	3050	10.76	<0.001
Acute Utility Needs	2872	3.12	2127	3.34	745	2.63	<0.001
Number of domains endorsed							<0.001
Zero (0)	36,531	39.70	23,959	37.63	12,572	44.35
One (1)	25,912	28.16	17,648	27.72	8264	29.15
Two (2)	12,022	13.06	8651	13.59	3371	11.89
Three or more (3+)	17,552	19.07	13,413	21.07	4139	14.60

3.2. Receipt of Resources and Referrals

Among Veterans endorsing one or more social risks on the ACORN screening tool, 42,461 (76.53%) were provided resources/referrals, 4143 (7.47%) declined assistance, 2490 (4.49%) reported already receiving services/assistance, and 9489 (17.10%) had a discordant response documented (i.e., positive ACORN screen paired with documentation of “screened negative for all needs”). Rural Veterans screening positive for social risks were less likely to receive resources/referrals (71.55% vs. 78.50%; p < 0.001) or report already receiving services/assistance (4.19% vs. 4.61%; p = 0.033), and more likely to decline assistance compared to their urban counterparts (10.63% vs. 6.21%; p < 0.001) or to have a discordant response documented (20.00% vs. 15.95%; p < 0.001) (Table 4).

TABLE 4.

Reported follow‐up for Veterans with positive ACORN screens between January 2024 and September 2025, overall and by rurality.

	Overall (N = 55,486)a		Urban (n = 39,712)		Rural (n = 15,774)
	#	%	#	%	#	%	p‐value
Resource and/or referral provided	42,461	76.53	31,174	78.50	11,287	71.55	<0.001
Already receiving services/assistance	2490	4.49	1829	4.61	661	4.19	0.033
Declined assistance	4143	7.47	2466	6.21	1677	10.63	<0.001
Screened negative for all needs (discordant response documented)	9489	17.10	6334	15.95	3155	20.00	<0.001
Missing	59	0.11	41	0.10	18	0.11	0.72

Note: Responses in the disposition section are not mutually exclusive, so staff may select multiple response options in the disposition section, including providing resources or referrals to the Veteran, the Veteran reporting already receiving services/assistance, the Veteran declining assistance, and a discordant response, defined as documentation of “screened negative for all needs” despite a positive ACORN screen.

^a The overall N for this table is based on the number of positive ACORN screens.

3.3. Association Between Rurality and Social Risks and Receipt of Resources and Referrals

In adjusted models, rural Veterans had lower odds of screening positive for one or more social risks (aOR, 0.92; 95% CI, 0.89–0.96). Rural Veterans also had lower odds of endorsing all social risk domains except for digital needs (aOR, 1.06; 95% CI, 1.02–1.11) (Table 5).

TABLE 5.

Association between rurality and social risk endorsement, overall and by each social risk domain.

	Unadjusted odds ratios (ORs)			Adjusted odds ratios (aORs)
	OR	95% CI	p‐value	aOR	95% CI	p‐value
Positive ACORN screens	0.76	0.74–0.78	<0.001	0.92	0.89–0.96	<0.001
Digital Needs	1.06	1.02–1.09	0.002	1.06	1.02–1.11	0.004
Educational Assistance	0.73	0.69–0.77	<0.001	0.88	0.83–0.94	<0.001
Employment Needs	0.53	0.50–0.57	<0.001	0.73	0.67–0.78	<0.001
Food Insecurity	0.69	0.66–0.71	<0.001	0.81	0.78–0.86	<0.001
Acute Food Needs	0.66	0.61–0.71	<0.001	0.76	0.69–0.83	<0.001
Housing Instability	0.51	0.49–0.54	<0.001	0.79	0.74–0.84	<0.001
Acute Housing Needs	0.47	0.42–0.53	<0.001	0.82	0.72–0.94	0.004
Legal Needs	0.68	0.64–0.72	<0.001	0.89	0.83–0.95	<0.001
Social Isolation/Loneliness	0.86	0.83–0.89	<0.001	0.89	0.85–0.92	<0.001
Transportation Insecurity	0.62	0.59–0.64	<0.001	0.80	0.76–0.84	<0.001
Utility Insecurity	0.80	0.76–0.83	<0.001	0.92	0.88–0.97	0.003
Acute Utility Needs	0.78	0.72–0.85	<0.001	0.86	0.78–0.95	0.003

Note: Multiple logistic regression models (n = 89,311) with site‐level random effects were adjusted for age (reference: 18–34 years), sex (male), race (modeled as separate binary indicators because Veterans could select multiple races), ethnicity (not Hispanic or Latino), relationship status (married/partnered), enrollment priority group (Groups 6–8), combat status (no), CAN score ≥95 (no), and chronic conditions count (0). Reference groups are in parentheses following each variable. Bolded values indicate a p‐value <0.05.

Among Veterans screening positive for one or more social risks, rural Veterans had lower odds of receiving resources/referrals (aOR, 0.89; 95% CI, 0.85–0.94) and reporting already receiving services/assistance (aOR, 0.85; 95% CI, 0.76–0.95) than their urban counterparts. Rural Veterans also had higher odds of declining assistance than urban Veterans (aOR, 1.27; 95% CI, 1.17–1.37) and having a discordant response documented (aOR, 1.09; 95% CI, 1.02–1.15) (Table 6). In sensitivity analyses additionally adjusting for the number of social risks endorsed, findings were somewhat attenuated but substantively unchanged, except for discordant responses, which were no longer statistically significant (Supporting Information Appendix Table 2).

TABLE 6.

Association between rurality and reported follow‐up among Veterans with a positive ACORN screen.

	Unadjusted odds ratios (ORs)			Adjusted odds ratios (aORs)
	OR	95% CI	p‐value	aOR	95% CI	p‐value
Any resource or referral provided	0.69	0.66–0.72	<0.001	0.89	0.85–0.94	<0.001
Already receiving services/assistance	0.91	0.83–0.99	0.033	0.85	0.76–0.95	0.003
Declined assistance	1.80	1.68–1.92	<0.001	1.27	1.17–1.37	<0.001
Screened negative for all needs (discordant response documented)	1.32	1.26–1.38	<0.001	1.09	1.02–1.15	0.006

Note: Multiple logistic regression models (n = 53,912) with site‐level random effects were adjusted for age (reference: 18–34 years), sex (male), race (modeled as separate binary indicators because Veterans could select multiple races), ethnicity (not Hispanic or Latino), relationship status (married/partnered), enrollment priority group (Groups 6–8), combat status (no), CAN score ≥95 (no), and chronic conditions count (0). Reference groups are in parentheses following each variable. Responses in the disposition section are not mutually exclusive, so staff may select multiple response options in the disposition section, including providing resources or referrals to the Veteran, the Veteran reporting already receiving services/assistance, the Veteran declining assistance, and a discordant response, defined as documentation of “screened negative for all needs” despite a positive ACORN screen. Bolded values indicate a p‐value <0.05.

4. Discussion

Among the 92,017 Veterans screened with ACORN between January 2024 and September 2025, over 60% reported at least one social risk. While rural Veterans were less likely than urban Veterans to screen positive across most social risk domains, they still faced a substantial social risk burden. More than half of rural Veterans endorsed at least one social risk, including over one‐quarter endorsing social isolation/loneliness and more than 13% reporting food insecurity. Additionally, over one‐fifth of rural Veterans reported at least one digital need (i.e., lack of access to a digital device, lack of affordable and reliable internet access, and/or needing assistance learning how to use a digital device to access their VA health care online), a significantly higher rate than among urban Veterans. Notably, although rural Veterans endorsed fewer social risks overall, those screening positive were less likely to receive resources/referrals or report already receiving services/assistance, and more likely to decline assistance. These differences persisted after adjustment for sociodemographic and clinical characteristics, as well as the number of social risks endorsed.

Our findings that rural Veterans had a high social risk burden but overall endorsed fewer risks than their urban counterparts is consistent with prior literature using VA EHR and administrative data, in which studies have found Veterans residing in rural areas had lower odds of food insecurity [42], housing inability/homelessness [42, 50, 51], employment/financial problems, legal problems, family/social support problems, access to care/transportation, and nonspecific psychosocial needs [42]. Our findings are also overall similar to those from a national survey of social needs among Veterans with or at risk for cardiovascular disease receiving VA care, in which overall need burden for both rural and urban Veterans was high, and rural Veterans were significantly less likely than urban Veterans to endorse loneliness, homelessness or housing instability, and transportation needs [20]. Contrary to our study, however, Gurewich et al. found no significant differences between rural and urban Veterans in the rates of food insecurity, utility insecurtity or needs, employment needs, or legal needs [20]. Our finding that rural Veterans reported higher rates of digital needs, including lower access to a digital device or affordable and reliable internet, is also consistent with prior work [32, 52, 53, 54, 55, 56]. Although the VA's Digital Divide Consult can help bridge digital needs by assisting Veterans without a video‐capable device or reliable internet access in obtaining a VA‐loaned internet‐connected device or applying for internet or device subsidies, Veterans living in certain rural and highly rural areas may still face challenges due to limited broadband options and connectivity issues [57].

To our knowledge, this study is the first to examine the provision of resources/referrals to rural Veterans endorsing social risks. Our finding that among rural Veterans screening positive for one or more risks, more than one‐quarter did not receive any resource/referral, and that they had significantly lower odds of receiving resources/referrals than urban Veterans, is striking. Potential reasons for lower rates of resource/referral provision among rural Veterans may include actual and/or perceived lack of available resources [58], as well as more limited staffing capacity and bandwidth in rural settings [59, 60, 61, 62, 63, 64, 65]. For example, in some rural sites, staffing constraints, cross‐site coverage demands, and space navigation often lead to social workers balancing multiple clinical duties across various areas and programs and spreading their work across multiple physical locations. These factors, coupled with the geography of the rural setting, may influence the types of resources/referrals offered to Veterans who report social needs. Additionally, certain virtual tools or services that could help bridge resource gaps—such as services provided through telehealth, online resources, or apps designed to address loneliness or social isolation—may be more difficult for rural Veterans to utilize given disparities in digital access. Future qualitative work is needed with staff administering ACORN in rural settings to better understand how these factors may influence the types of resources/referrals offered to rural Veterans.

Rural Veterans were also significantly more likely to decline assistance than urban Veterans, although these higher rates of declining assistance account for only a portion of the disparity in the receipt of resources/referrals. Reasons for lower rates of reported social risks and higher rates of declining assistance among rural Veterans are likely multifactorial. There is some literature to suggest that rural individuals may perceive social risks and social needs differently than their urban counterparts [20, 66]. Individuals in rural areas often possess a strong sense of community [67] and may rely more heavily on services found through networks composed of family, friends, and church or community members [68]. Rural Veterans may feel their needs are adequately addressed through these supports and/or prefer to utilize familiar networks instead of accepting resources/referrals provided through healthcare services like those offered through or associated with the VA. Further, rural Veterans may also underreport social risks and social needs or decline services due to the high value placed on independence and self‐reliance; emotional stoicism; concerns about taking resources away from other Veterans, especially those who they may perceive to be in greater need; mistrust toward individuals and institutions (e.g., healthcare systems); and/or stigma associated with reporting needs and/or seeking or accepting assistance [20, 66, 69, 70, 71]. Given that rural areas often have fewer community‐based organizations and supportive social care services, rural Veterans may also be less likely to report social risks on the ACORN screening tool if they believe no resources are available or that they are not accessible. Future research is needed to better understand rural Veterans’ perceptions about the acceptability and appropriateness of social risk screening and referral interventions such as ACORN, as well as why rural Veterans may be more likely to decline assistance than urban Veterans.

We also observed relatively high rates of discordant responses, defined as documentation of “screened negative for all needs” despite a positive ACORN screen. Although rural Veterans initially appeared more likely than urban Veterans to have a discordant response documented, this association was no longer significant after adjusting for the number of social risks endorsed. These findings highlight important challenges in interpreting structured EHR documentation related to social risk screening and referral workflows. Discordant responses may reflect differences between screening positive for a social risk vs. perceiving a need for resources or assistance, misunderstanding among staff regarding which screening responses constitute a positive screen, or limitations of the current EHR template design. Because the EHR template requires documentation of at least one disposition field to complete the note regardless of screening result, the ACORN team created a “screened negative for all needs” option to allow staff to document that no resources/referrals were provided in instances where Veterans screened negative across all nine ACORN domains. Thus, discordant responses may reflect variation in the interpretation and documentation of disposition options, as well as EHR workflow limitations, more than a distinct clinical phenomenon. These findings underscore the importance of continued refinement of social risk screening and referral documentation workflows, as well as consideration of revised ACORN EHR template wording to reduce ambiguity. To support staff in selecting the most appropriate disposition option, our team developed training materials reinforcing what responses are considered positive and the meaning of each discrete disposition option, and also updated the EHR template to indicate with an “*” in the body of the screening tool which screening responses are considered positive. Future qualitative work should seek to better understand how Veterans and staff interpret positive social risk screens and resource/referral needs.

Another possible reason for lower rates of reported social risks among rural Veterans is the VA's significant investment in responding to the needs of rural Veterans and reducing health and access disparities among this population. This includes the development of Clinical Resource Hubs [15, 16] and ATLAS sites, which offer community‐based locations (e.g., at a Veteran Service Organization, American Legion) where Veterans living in rural areas can conduct video telehealth appointments with their VA clinicians [17]. Additionally, there are collaborative efforts between the VA Office of Rural Health and the VA National Social Work Program to ensure adequate social work staffing for rural primary care teams [65]. These initiatives may have effectively helped mitigate social needs among rural Veterans [65]. Given the heterogeneity of rural populations, however, future research is needed to more fully understand social risks and social needs within various sociodemographic, geographic, and clinical subgroups to better tailor interventions for specific rural Veteran populations.

4.1. Limitations

Findings from this study should be interpreted within the context of several limitations. First, although our cohort included all Veterans with rurality data in the EHR screened using the ACORN national EHR template between January 2024 and September 2025 across 87 sites, this was not a nationally representative sample of all Veterans receiving VA health care. Thus, our findings, including prevalence estimates, may not be generalizable to the overall VA population or to Veterans not receiving VA care. Second, although some clinical settings use Veteran self‐administered ACORN screening (on paper or the Veteran's own electronic device) and several pilots are underway to expand options for self‐administered screening, most ACORN screenings are currently staff‐administered. This may have led to underestimates of reported social risks, as prior work suggests some patients prefer self‐administered screening for sensitive topics [72, 73, 74, 75] and demonstrate higher disclosure rates when questions are self‐administered [72, 74, 75, 76]. However, ACORN screening is conducted by trained clinical staff and contextualized as a means of connecting Veterans with relevant resources/referrals, which may help mitigate underreporting. Third, although ACORN includes a Veteran‐tailored screening tool iteratively refined based on Veteran input and cognitive testing [27], the screening tool was not specifically developed for rural Veterans, and it may underdetect social risks within certain sociodemographic subgroups. Fourth, data on resources/referrals were limited to information documented in structured EHR disposition fields and may not fully capture all aspects of these interactions. These limitations notwithstanding, our findings have important implications for understanding rural–urban differences in social risks and receipt of resources/referrals among Veterans receiving VA care.

5. Conclusion

This study examined rural–urban differences in social risks reported by Veterans receiving care in VA, as well as receipt of resources/referrals for social needs. Although rural Veterans were less likely to report social risks overall, those screening positive were less likely to receive resources/referrals and more likely to decline assistance than urban Veterans. These findings can help guide tailored implementation of social risk screening and follow‐up interventions for rural Veterans and inform future research examining reasons for higher rates of declining assistance among rural Veterans and differences in receipt of resources/referrals to both VA and community services.

Funding

This work was funded by the Veterans Health Administration (VHA) Office of Health Equity and supported by the VHA National Social Work Program, Care Management and Social Work Services. A.J. Cohen was additionally supported by CDA 20‐037 from the US Department of Veterans Affairs (VA) Health Systems Research.

Disclosure

The contents do not represent the position or policy of the US Department of Veterans Affairs, the US government, or any other organizations.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting file 1: jrh70186‐sup‐0001‐Appendix.docx