Barriers to Healthcare and Social Service Utilization Among Rural Older Adults Who Use Drugs

Abstract

The objective of this study was to understand barriers to healthcare and social service utilization among older adults residing in rural areas who use drugs. A cross-sectional survey of persons who use opioids or inject drugs in rural counties with high overdose rates across ten states was conducted. For this analysis, participants were restricted to only the 375 individuals aged 50 and older. They were asked about barriers to utilizing healthcare and social services. Multivariate analyses were conducted. The most common barriers were a lack of transportation and a fear of stigma. The average number of barriers was 2.53. Those who were either uninsured or homeless endorsed 37% more barriers. For every five-year increase in age, the number of barriers reduced by 15%. Efforts to reduce these barriers may include expanding eligibility for transportation and housing services and leveraging trusted community members to broker linkages to providers to overcome stigma.

INTRODUCTION

Substance use and misuse remain a pervasive public health concern, with over 100,000 reported drug overdose deaths in the United States (U.S.) in 2021 (Prevention, 2021). While the reported rate of drug overdose deaths in urban areas is slightly higher than that in rural areas (28.6 per 100,000 people versus 26.2 per 100,000) (Spencer et al., 2022), people who use drugs (PWUD) residing in rural areas face unique barriers to accessing healthcare and social services as a means of addressing their substance use, particularly compared with their counterparts in urban areas. Specifically, PWUD in rural areas have insufficient access to harm reduction and treatment resources (Jenkins, 2021; Jenkins & Hagan, 2020), and may face additional barriers to receiving treatment due to transportation-related challenges (e.g., substantial distance to travel, a lack of sufficient public transit) (Lister et al., 2020), and may receive sub-standard care relative to those in urban areas (Lister et al., 2020). As such, researchers, policymakers, and practitioners have prioritized ensuring that the unique treatment needs of substance users in rural areas are better understood and that interventions can effectively mitigate their unique challenges and barriers.

Rural areas are not monolithic and less focus has been paid to rural older adults who use drugs, which is detrimental to the overall goal of reducing the impact of the opioid epidemic in rural areas (Rigg et al., 2018). In general, rural areas have a greater proportion of older adults than urban areas. In rural areas of the U.S., 17.5% of the population is aged 65 and older, compared to only 13.8% of the population in urban areas (Smith & Trevelyn, 2018). This means there are approximately 10 million older adults living in the rural U.S (Smith & Trevelyn, 2018). Rural older adults have disparate health outcomes compared to older adults in urban areas, such as lower survival rates for Alzheimer’s Disease and related dementias, higher incidence of and mortality from falls, and a higher risk of cardiovascular disease and stroke (Cohen & Greaney, 2022). This is coupled with significantly lower access to healthcare and resources such as screening and treatment services as well as home and community-based services (Cohen & Greaney, 2022).Given these facts, older adults are a unique and important sub-group of rural drug users to understand.

Substance use disorders among all older adults are increasing (U.S. Department of Health and Human Services Substance Abuse and Mental Health Services Administration Center for Behavioral Health Statistics and Quality, 2018) and there are a myriad of reasons why rural older adults may be particularly affected. In the case of opioid misuse, risk factors such as low income, substandard housing, and unemployment are more common in rural areas and may be especially salient for rural older adults (Benson et al., 2019). Additionally, older adults are more likely than younger adults to have other health issues that may contribute to their substance use (e.g., chronic pain that necessitates the use of prescription opioids) (Yarnell et al., 2020) and the physiological changes experienced by older adults as part of the normal aging process can have an impact on the short- and long-term effects of certain substances, which further differentiates them from other (i.e., younger) drug users (Lehmann & Fingerhood, 2018). In a study of rural stimulant users, older adults had worse physical and mental health than their younger rural counterparts, despite having fewer substance use problems, suggesting rural older adult drug users may have significant healthcare needs (Woodhead et al., 2019).

Rural older adults are also significantly less likely to have access to substance use disorder (SUD) treatment than older adults in urban areas. In general, there are fewer SUD treatment options available to rural drug users but the options that do exist may be especially challenging for older adults to access due to issues such as a lack of transportation. Furthermore, the diagnostic criteria in the DSM-5 for SUD include contextual considerations, such as the impact of an individual’s drug use on their work, school, or home life, and many older adults do not work, attend school, or live with others; therefore, strict adherence to the diagnostic criteria may mean that many older adults with SUD may not be identified and diagnosed even if SUD treatment is accessed (Lehmann & Fingerhood, 2018).

Due to the lack of designated SUD treatment options, primary care providers are critical in the effort to treat SUDs in rural areas. This is particularly important for older adults who, compared to younger adults, utilize primary care services twice as often and generally have more health issues that require healthcare services (Institute of Medicine (US) Committee on the Future Health Care Workforce for Older Americans, 2008). Primary care providers may be the only way rural older adults misusing drugs may be identified, supported, and treated. Additionally, SUD treated in the primary care setting is associated with improved SUD treatment and chronic disease outcomes (Turi et al., 2024). Linking rural older adult drug users to primary care would not only increase the likelihood of a SUD being addressed and treated effectively but would also provide an opportunity for other health issues to be addressed that, as stated above, are more prevalent in these individuals due to their age, rurality, and drug use. Furthermore, linking rural older adult drug users to primary care could possibly reduce their use of emergency services for SUD, which contributes to healthcare inefficiencies and costs and puts additional strain on already underresourced rural hospitals (Turi et al., 2024).

Understanding rural older adult drug users’ healthcare utilization and the barriers they face to accessing healthcare is essential for providing both healthcare and SUD treatment to this unique and important population. Therefore, the goal of these study activities is to identify the type and frequency of, as well as barriers to, rural older adult drug users’ utilization of healthcare and social services. We also aimed to identify variations in these patterns based on characteristics such as race, education, and health insurance status. Notably, we chose to define older adults as individuals aged 50 and older based on the Life Course Perspective, which emphasizes the value of linking life stages and examining transitions (Crosnoe & Elder, 2002). In other words, individuals ages 50–65 will be aging into the more traditional definition of “older adults” (age 65+) and understanding their experiences during the “50–65 years old” stage will provide valuable perspective and insight into their experiences during the “65+ stage”.

METHODS

Study design, setting, and participants

The goal of this National Institute on Drug Abuse cooperative agreement is to apply the results of a cross-sectional survey of PWUD in rural counties with high overdose rates across 8 project areas (“sites”) in ten states participating in the Rural Opioid Initiative (ROI): Illinois, Kentucky, North Carolina, Massachusetts, New Hampshire, and Vermont, Ohio, Oregon, West Virginia, and Wisconsin. All sites obtained local IRB approval for research activities and data sharing. A full description of the ROI’s structure and operations is described elsewhere (Jenkins et al., 2022).

Individuals were eligible for inclusion if they lived in the study area, reported past 30-day use of any opioid “to get high” (heroin, prescription pain medication) and/or past 30-day injection of any drug, were able to communicate in English, and met site-specific age criteria (age ≥15 at two sites and ≥18 at six sites). Participants were recruited between January 2018 and March 2020, provided informed written consent, and received $40-$60 for completion of study procedures.

Data sources/ measurement

Following recruitment and informed consent, participants completed a standardized, structured questionnaire via audio computer-assisted self-interview at five sites, computer-assisted personal interview at two sites, and computer-assisted self-interview at one site. The questionnaire assessed participants’ self-reported demographic and behavioral characteristics. Data were transferred to the ROI Data Coordinating Center for quality review and collation of a cross-site analytic dataset.

The following analytic results are based upon a restricted sample of adults aged 50 or older, collated across all the ROI sites. In additional to the empirical rationale using the Life Course Perspective, individuals 50 or older were chosen as the population of interest for this study because of the limited number of older individuals, in general, within the larger dataset. Narrowing the data to participants 50 or older was done to accurately capture social determinants of health that older drug users face while utilizing a large enough sample size. Additionally, many individuals in this age range are approaching age 65, the age requirement for public services such as Medicare, which could provide insights for policymakers on how to support this population as they age into these services.

Variables of interest

Demographics and other key sample characteristics were assessed through the following variables: age, gender, race, ethnicity, sexual orientation, education level, income sources, marital status, insurance status, internet use, cell phone use, and a history of homelessness in the previous six months. Due to the limited variation in the sample for some of these variables, they were re-coded to allow for more meaningful comparisons. The distributions of the original variables are presented in Appendix Table 1.

Utilization of healthcare and social services was measured with a series of variables (Appendix Table 2). Healthcare was measured with a single item in which participants were asked, “What is the main place where you received medical care in the past six months?” and response options included: private doctor; community health center; health department; urgent care center; emergency department; mobile health clinic or van; other; and “I have not received medical care in the past six months.” Utilization of social services was measured with multiple items using the same question stem – “Have you received…in the past six months?” and services included: SNAP/food stamps; home visits from a nurse or other provider; well-baby care; food from a food pantry; case management from a health provider, clinic, or other agency; and job counseling or training.

Barriers to healthcare utilization were measured with a series of items using the question stem – “In the last six months I did not go for needed medical care because…” with the following response options: unable to pay; unsure where to go; did not have transportation; clinic’s hours of operation were not convenient; treated poorly at a clinic in past; did not want to be seen at a clinic; distrusts doctors; do not care about taking care of self; no child care; too drunk or high; afraid providers would be disrespectful because of individual’s drug use; and either received treatment from someone else (e.g., a friend, family member, or other non-healthcare-provider community members) or the individual treated themselves.

In addition to examining the barriers individually, a count variable was created for each participant with the number of barriers they endorsed.

Analysis

Univariate analyses were conducted, including calculating counts, percentages, means, ranges, and standard deviations (SD). A number of models were conducted and likelihood ratio tests were conducted to determine the best fit; the outcome was the count of barriers endorsed. These models included null (no predictors), Poisson, and negative binomial models. The potential predictors were the variables of interest listed above. One key characteristic, sexual orientation, was excluded from the model due to one of the sites not including this question in their data collection survey. However, a sensitivity analysis of the impact of this variable on the model was conducted and is presented in Appendix Table 3. All analyses were conducted in R. Any rows with missing values were dropped.

RESULTS

Sample Characteristics

Of the 3,038 total participants aged 18 and older, 375 (12.3%) were aged 50 or older and thus were included in the sample. Of these 375 participants, the average age was 55 (SD: 4.03, range: 5070). The sample predominately identified as male (64%), white, non-Hispanic (84%), straight (93%), and not currently partnered (i.e., not currently married or living with a partner; 83%). A majority had at least a high school diploma (77%), had health insurance (85%), and their primary income source was a form of public assistance (i.e., Social Security Insurance, military, TANF, etc.; 53%). A majority also reported having access to the Internet (68%) and a cell phone (64%). Notably, 44% reported experiencing homelessness in the previous six months. These results are summarized in Table 1.

Table 1.

Characteristics of Sample of Rural Older Adult Drug Users

Characteristic (n=375)
	Mean (SD) Range
Age	55 (4.03) 50–70
Gender (n=374)	% (n)
Male	64 (240)
Female	36 (134)
White race
White	84 (316)
Non-White	16 (59)
Education (n=374)
High School Diploma or More	77 (288)
Less than High School	23 (86)
Sexual orientation (n=278)
Straight	93 (258)
Not Straight	7 (20)
Income Sources (n=374)
Working	28 (103)
Government assistance	53 (199)
Other	14 (52)
Unknown	5 (20)
Marital status (n=362)
Married or Living with Partner	17 (60)
Not Currently Partnered	83 (302)
Health Insurance Status (n=369)
Insured	85 (312)
Uninsured	15 (57)
Internet use
Yes	68 (256)
No	30 (113)
Missing	2 (6)
Cell use
Yes	64 (240)
No	35 (132)
Missing	1 (3)
Homeless in the past six months	44 (165)

Utilization of and Barriers to Service

A negative binomial model was fitted as a means of estimating the association between participant demographics and other key characteristics (with the exception of sexual orientation due to some sites not collecting these data) and the number of barriers reported.

The primary source of medical services was reported as private doctor (39%), community health center (17%), and emergency department (15%). Twelve percent of participants had not received medical care in the past six months. The remaining participants utilized the following as their primary medical source: urgent care (8%), other (5%), health department (2%), and mobile health clinic or van (1%). Services related to securing food were the most commonly used social services among participants (SNAP, 69%; food pantries, 64%). Less than 20% of participants used case management services (19%) while less than 10% used job counseling services (9%), home visits from a healthcare provider (2%), and WIC services (2%).

The most common barriers to utilizing medical care when needed were a lack of transportation (36%) and a fear of being treated with disrespect due to reported drug use (34%). Other commonly reported barriers included an inability to pay (26%), being too high or drunk (25%), a distrust of doctors (20%), and having been treated poorly at a clinic in the past (20%). Less than one-fifth of participants reported not using medical care when they needed it because they did not care about taking care of themselves (17%), they did not want to be seen at a medical clinic (16%), they were not sure where to go (16%), the clinic’s hours were inconvenient (11%), or they did not have childcare (3%). Importantly, more than a quarter of participants did not seek needed medical care because someone (described to participants in the questionnaire as “not a doctor or nurse”) treated them or they treated themselves (28%). The average number of barriers endorsed by participants was 2.53 (SD: 2.43, Range: 0–10, with 28% [n=106] of participants reporting zero barriers). These results are summarized in Tables 2 and 3.

Table 2.

Healthcare and Social Service Utilization among Rural Older Adult Drug Users

Type of service utilized in past six months (n=375)	% (n)
Main source of medical care
Private doctor	39 (147)
Community health center	17 (64)
Health department	2 (6)
Urgent care center	8 (30)
Emergency department	15 (57)
Mobile health clinic or van	1 (2)
Other	5 (18)
I have not received medical care in the past six months	12 (46)
SNAP, food stamps	69 (259)
Home visits from nurse or other provider	2 (9)
Food from food pantry	64 (240)
WIC	2 (7)
Case management from healthcare provider, clinic, or other agency	19 (73)
Job counseling or training	9 (33)

Table 3.

Barriers to medical care access and utilization among Rural Older Adult Drug Users

“In the last six months, I didn’t go for needed medical care because…” (check all that apply)	% (n)
I could not pay	26 (98)
I was not sure where to go	16 (61)
I did not have transportation	36 (134)
The clinic’s hours of operation were not convenient	11 (42)
I was treated poorly at a clinic in the past	20 (75)
I did not want to be seen at a medical clinic	16 (61)
I do not trust doctors	20 (76)
I don’t really care about taking care of myself at this time	17 (63)
I didn’t have childcare	3 (13)
I was too drunk or high	25 (93)
I was afraid they would treat me with disrespect since I use drugs	34 (127)
Someone else, not a doctor or nurse, treated me, or I treated myself	28 (105)
Average number of barriers endorsed (SD)	2.53 (2.43)

Modeling the Number of Barriers Experienced

Dropping cases with any missing data resulted in 348 observations for the model comparisons. Likelihood ratio tests found that the Poisson model performed better than the null (X2 = 51.61, p<.001), and the negative binomial outperformed the Poisson (X2 = 163.71, p<.0001). Holding constant participants’ sex, race, employment, education, marital status, internet use, and cell phone use, barrier counts were increased if a participant was uninsured, homeless, and/or of lower age. Full model results are available in Table 4.

Table 4.

Characteristics Associated with Number of Barriers to Healthcare Utilization among Rural Older Adult Drug Users

	Incidence Rate Ratio (95% Confidence Interval)a*
Age	0.97 (0.94 – 1.00)*
Sex
Not Male	Ref
Male	0.88 (0.69 – 1.12)
Education
Less than high school	Ref
At least high school	1.16 (0.88 – 1.52)
Employment
Not Working	Ref
Working	0.89 (0.68 – 1.17)
Insurance status
Insured	Ref
Not insured	1.37 (1.01 – 1.88)*
Marital status
Not currently married or living with a partner	Ref
Married or living with a partner	0.96 (0.71 – 1.31)
White Race status
Non-White	Ref
White	1.17 (0.85 – 1.60)
Homeless in previous six months
No	Ref
Yes	1.37 (1.09 – 1.71)*
Current Internet access
No	Ref
Yes	0.79 (0.62 – 1.02)
Current cell phone access
No	Ref
Yes	1.04 (0.82 – 1.32)

^aIncidence rate ratio for number of healthcare barriers.

^*p-value <0.05

Outcome is a count variable of the number of barriers endorsed out of 12 total. Model variables included sex, age, insurance status (uninsured vs. insured), employment (working full or part-time vs. not working), education (at least high school diploma vs. less than high school diploma), marital status (married or living with partner vs. not married or living with partner), race (white vs. non-white), history of homelessness in previous six months, current access to the Internet, and current access to a cell phone.

For a one-unit change in each predictor variable, the difference in counts of barriers changes accordingly. Specifically, there was a 37% increase in the number of barriers endorsed by uninsured participants compared to insured participants (IRR:1.37; 95% CI = 1.01, 1.87). Likewise, experiencing homelessness in the past six months increased the number of barriers endorsed also by 37% (IRR:1.37; 95% CI 1.09, 1.71). For every five-year increase in age, there was a 15% decrease in the number of barriers a person faced (IRR for one year: 0.97; 95% CI = 0.94, 1.00).

DISCUSSION

Older adult drug users in rural areas represent a unique population due to their many potential vulnerabilities. Despite this, we know little about this important subgroup. Our focus of this study was to analyze the patterns of healthcare and social service utilization and barriers to utilization and identify any demographics or other key characteristics associated with the barriers individuals experienced.

In general, rural older adult drug users were well connected to healthcare, with over half reporting primary care providers or community health clinics as their primary sources of healthcare. This may be due to the large percentage (85%) who reported having health insurance, which could be a result of the large number of individuals covered under Medicaid (62%) or Medicare (23%). Transportation was the most commonly reported barrier to healthcare utilization, which is a commonly reported barrier among all rural residents. However, older adults may be particularly vulnerable to this barrier given that many rural older adults stop driving, may have mobility impairments that require specialized transportation (i.e., chair lifts), or may lack caregiver support for rides (Morken & Warner, 2012). There was also a 37% increase in the number of healthcare barriers reported by individuals who were uninsured compared to insured individuals, which suggests that even though a majority of rural older adult drug users have health insurance, those who do not still face more barriers to accessing and utilizing healthcare.

It is also important to note that despite a majority of the sample having health insurance and receiving healthcare at a primary care or community health clinic, 23% still reported their primary source of healthcare to be either the emergency department (15%) or urgent care (8%) and 12% reported not receiving healthcare at all in the previous six months. This suggests that although a majority of individuals have established healthcare at primary care or community clinics, many are still using emergency/urgent care for non-urgent care needs or not receiving care at all. Shifting this care from the emergency/urgent care setting to primary care would be beneficial not only for healthcare system costs but also for improved patient outcomes. Recent findings suggest that supporting rural primary care nurse practitioners in the treatment of SUD among rural older adults can reduce emergency department utilization for SUD among this population (Turi et al., 2024).

Stigma was also a prevalent theme in the barriers that participants reported. Thirty-four percent of individuals reported a fear of being treated with disrespect, 20% a distrust of doctors, 20% being treated poorly at a clinic in the past, and 16% that they did not want to be seen at a medical clinic. Taken with the fact that many individuals were receiving care from emergency/urgent clinics and given the inconsistency of providers working at any given time at these locations, it is understandable that many individuals have not been able to establish trusting, respectful relationships with healthcare providers. A lack of established relationships with providers is also evident with the fact that 28% of individuals reported not seeking healthcare because someone else (not a doctor or nurse) treated them or they treated themselves.

Ageism and stereotypes may influence this stigma as well. Many providers are not properly trained to identify and treat patients with SUD. This, compounded with the misconception that drug users are young, may prevent older adults from achieving equitable treatment within the healthcare system (Rao & Arora, 2015). Furthermore, many older adults feel ashamed or embarrassed about their substance use (Rao & Arora, 2015), which means that even with access to healthcare barriers to receiving support for a SUD among this population remain. Thus, while mechanisms to link individuals to primary care or community clinics are needed, it is clear that without addressing the stigma and distrust between healthcare providers and people who use drugs, it will be difficult for these linkages to be formed or sustained. Identifying those in the community whom these individuals already trust and leveraging those trusted individuals to broker linkages to healthcare providers could prove useful in overcoming this barrier.

In addition to the relationship between insurance and healthcare barriers, our study found that individuals who had experienced homelessness in the past six months had more barriers to healthcare than those who had not experienced homelessness. This is not surprising given that people experiencing homelessness have repeatedly and routinely reported difficulty accessing care and services and have reported experiencing stigma from service providers (Ramsay et al., 2019). Rural areas also have well-documented housing shortages. Rural housing tends to be older, meaning that necessary renovations or construction puts barriers on those looking to buy a home. Additionally, the number of older adults living in rural areas is expected to increase by 30% within the next decade, totaling 13 million older adults living in a rural community. Current infrastructure in rural areas, especially those most desirable to older individuals, will struggle to meet this demand (Pendall et al., 2016). This evolving housing crisis is evident given that nearly half (45%) of the sample had experienced homelessness in the past six months. In addition to reducing the barrier of stigma mentioned above, efforts should be made to increase affordable housing availability in rural areas to help mitigate its effect on healthcare utilization.

Lastly, with the negative binomial model and its marginal effects, we found an age-specific reduction in the number of barriers endorsed as age increases, meaning older individuals reported fewer barriers than younger individuals. This may seem contradictory at first; however, it might be due to individuals aging into Medicare coverage, which would help to establish more routine care and become eligible for more services such as transportation services. If this is the case, expanding eligibility for programs such as Medicaid may further reduce this barrier.

To fully contextualize the results of this study, its limitations must be considered. First, this is a cross-sectional study and, although the sample comes from a diverse group of sites across the U.S., it is not representative of all rural older adult drug users. Second, the lack of variation or availability on key variables, such as sexual orientation and type of drugs used in the past or present, hindered our ability to identify disparities among sub-groups. There may be relevant distinctions between, for example, a rural older adult injecting drugs compared to a rural older adult misusing prescription opioids, and further research is needed to understand these distinct subgroups. Importantly, the results of this study can serve to generate hypotheses and research questions for future researchers.

Despite these limitations, to our knowledge, this is the first study of healthcare and social service utilization and barriers among rural older adult drug users in the U.S. These individuals face a number of challenges and, as with all people who use drugs, their needs and supports should not begin and end with their drug use. Attention and resources should be paid to all of their needs, including their healthcare and social service needs.

What this paper adds

• Contributes to the evidence on rural older adult drug users, a significantly understudied subgroup.

• Participants cited a lack of access to transportation, a fear of stigma, and a fear of being treated with disrespect due to their drug use as the most common barriers to utilizing healthcare and social services.

• Participants who were homeless, uninsured, and younger reported more barriers than their counterparts.

Applications of study findings

• Expanding access and/or eligibility to services such as transportation and housing in rural areas may help overcome some of the barriers to utilizing healthcare and social services for rural older adult drug users.

• Identifying who in a given rural community older adult drug users trust and leveraging those individuals to broker relationships with healthcare and social service providers may help overcome the barriers related to stigma and fear.

• Rural areas are aging quicker than urban areas and to ensure this important and growing subgroup is supported, it is imperative that we deepen our understanding of their experiences.

Appendix Table 1: Unedited variable distributions

(n=375)
	Mean (SD) Range
Age	55 (4.03) 50–70
Gender	% (n)
Male	64 (240)
Female	36 (l34)
Transgender	0 (1)
Race/ethnicity
American Indian	6 (21)
Asian	0 (1)
Black	6 (21)
Latinx	3 (12)
Mixed race	2 (6)
Other	1 (4)
White	83 (310)
Education
Less than high school	23 (86)
High school diploma or GED	47 (175)
Some college	20 (74)
Associate’s degree, trade, or technical school	7 (28)
Bachelor’s degree, other 4 year college degree, or more	3 (11)
Missing	0 (1)
Sexual orientation
Straight	69 (258)
Lesbian or gay	1 (2)
Bisexual	4 (14)
Other	1 (4)
Missing	26 (97)
Income sources (check all that apply)
Full-time work	9 (34)
Part-time work	19 (70)
Retirement check	7 (28)
Public assistance check (TANF, AFDC, etc.)	5 (19)
Disability check (SSI, military, etc.)	46 (173)
Selling drugs	9 (35)
Selling sex	2 (9)
Theft, shoplifting, or stealing	5 (19)
Someone supports me	14 (52)
Marital status (n=362)
Married	10 (38)
Divorced or separated	53 (202)
Widowed	9 (35)
Never married	17 (65)
Living with partner	6 (22)
Missing	4 (13)
Homeless in the past six months	44 (165)
Health insurance status
Uninsured	15 (57)
Private or commercial insurance	3 (11)
Medicaid	54 (202)
Medicaid Expansion	8 (30)
Medicare	23 (85)
Insurance from the VA	3 (12)
Other	3 (12)
Internet use
Yes	68 (256)
No	30 (113)
Missing	2 (6)
Cell use
Yes	64 (240)
No	35 (132)
Missing	1 (3)

Appedix Table 2: Survey Items

Age	How old are you?
Gender	What is your gender? Male Female Transgender Other
Education	How much school did you finish? Less than high school High school diploma or GED Some college Associates degree, trade, or technical school Bachelors degree, other 4 year college degree, or more
Race	What race are you? Choose one. White African American or Black American Indian Alaskan Native Asian, Pacific Islander, or Native Hawaiian African Mixed race Other
Ethnicity	Are you Hispanic or Latino? Yes No Don’t know
Sexual Orientation	What is your sexual orientation? Choose one. Straight Lesbian or gay Bisexual Other Don’t know or not sure
Medical Care Utilization	What is the main place where you received medical care in the past 6 months? Private doctor Community health center Health department Urgent care center Emergency department Mobile health clinic or van Other I have not received medical care in the past 6 months
Barrier_1	In the last 6 months I didn’t go for needed medical care because I could not pay for medical care. Yes No Don’t know
Barrier_2	In the last 6 months I didn’t go for needed medical care because I was not sure where to go to get medical care. Yes No Don’t know
Barrier_3	In the last 6 months I didn’t go for needed medical care because I did not have transportation to medical care. Yes No Don’t know
Barrier_4	In the last 6 months I didn’t go for needed medical care because the clinic’s hours of operation were not convenient. Yes No Don’t know
Barrier_5	In the last 6 months I didn’t go for needed medical care because I was treated poorly at a clinic in the past. Yes No Don’t know
Barrier_6	In the last 6 months I didn’t go for needed medical care because I did not want to be seen at a medical clinic. Yes No Don’t know
Barrier_7	In the last 6 months I didn’t go for needed medical care because I don’t trust doctors. Yes No Don’t know
Barrier_8	In the last 6 months I didn’t go for needed medical care because I don’t really care about taking care of myself at this time. Yes No Don’t know
Barrier_9	In the last 6 months I didn’t go for needed medical care because I didn’t have child care. Yes No Don’t know
Barrier_10	In the last 6 months I didn’t go for needed medical care because I was too drunk or high. Yes No Don’t know
Barrier_11	In the last 6 months I didn’t go for needed medical care because I was afraid they’d treat me with disrespect since I use drugs. Yes No Don’t know
Barrier_12	In the last 6 months I didn’t go for needed medical care because someone else, not a doctor or nurse, treated me, or I treated myself. Yes No Don’t know
Internet	Do you use to the internet? Yes No Don’t know
Cell	Do you have a cell phone with active service now? Yes No Don’t know
Income	What are your main sources of income over the last 6 months? Choose all that apply Full-time work (40 hrs/week) Part-time work Retirement check Public assistance check - like TANF, AFDC, etc. Disability check, like SSI, military, or other Selling drugs Selling sex Theft, shoplifting, or stealing Someone supports me
SNAP	Have you received SNAP, like food stamps, in the past 6 months? Yes No Don’t know
Home_Visits	Have you received home visits from a nurse or other provider in the past 6 months? Yes No Don’t know
Well_Baby	Have you received well baby care in the past 6 months? Yes No Don’t know
Food	Have you received food from a food pantry in the past 6 months? Yes No Don’t know
WIC	Have you received WIC, the Women, Infant, and Children food and nutrition service, in the past 6 months? Yes No Don’t know
Case_mgmt	Have you received case management from a health provider, clinic, or other agency in the past 6 months? Yes No Don’t know
Job_Train	Have you received job counseling or training in the past 6 months? Yes No Don’t know
Marital	What is your current marital status? Married Widowed Divorced Separated Never married Living with partner Refused Don’t know
Homeless	Have you been homeless in the past 6 months? “Homeless” means you were living from place-to-place, “couch-surfing”, on the street, in a car, park, abandoned building, squat, or shelter. Yes No Don’t know
Transportation	Do you have a way to get to medical appointments? Yes, I have a car, access to a car, or I can walk No Maybe, if I can get a ride from a friend or relative Maybe, if public transportation is available, like a medicab
Insurance	Do you currently have health insurance or health care coverage? Yes No Don’t know

Appendix Table 3. Sensitivity Analysis - Characteristics associated with number of barriers to healthcare utilization among rural older adult drug users, with the study site that did not collect sexual orientation removed from the sample thus allowing sexual orientation to be included in the model

	Incidence Rate Ratio (95% Confidence Interval)a*
Age	0.97 (0.94 – 1.00)
Sex
Not Male	Ref
Male	0.94 (0.70 – 1.23)
Education
Less than high school	Ref
At least high school	1.13 (0.82 – 1.57)
Employment
Not Working	Ref
Working	0.75 (0.52 – 1.08)
Insurance status
Insured	Ref
Not insured	1.61 (1.08 – 2.44)*
Marital status
Not currently married or living with a partner	Ref
Married or living with a partner	1.06 (0.73 – 1.55)
White Race status
Non-White	Ref
White	1.03 (0.68 – 1.53)
Sexual Orientation status
Straight	Ref
Lesbian or Gay	0.25 (0.01 – 2.14)
Bisexual	1.14 (0.64 – 2.11)
Other	0.73 (0.23 – 2.76)
Homeless in previous six months
No	Ref
Yes	1.46 (1.10 – 1.95)*
Current Internet access
No	Ref
Yes	0.83 (0.62 – 1.11)
Current cell phone access
No	Ref
Yes	1.08 (0.81 – 1.42)

^aIncidence rate ratio for number of healthcare barriers.

^*p-value <0.05

Outcome is a count variable of the number of barriers endorsed out of 12 total. Model variables included sex, age, insurance status (uninsured vs. insured), employment (working full or part-time vs. not working), education (at least high school diploma vs. less than high school diploma), marital status (married or living with partner vs. not married or living with partner), race (white vs. non-white), sexual orientation (straight vs. lesbian or gay vs. bisexual vs. other), history of homelessness in previous six months, current access to the Internet, and current access to a cell phone.