No Place to Rest: Housing Instability and Pain Experiences in the Emergency Department

Abstract

Introduction:

Housing instability increases chronic pain risk and creates care barriers, making emergency departments the primary points of health care for people experiencing housing instability. Although it is clear that housing instability contributes to chronic pain, the patient's experience of pain is less understood. This study examines differences in pain characteristics by housing status.

Methods:

This secondary analysis used data from the Decision-Making Factors for Therapeutic Opioid Use after Emergency Care. In the original study, randomly selected eligible participants self-reported housing status, in addition to validated measures of pain interference, pain self-efficacy, pain anxiety, depression, and sleep disturbances. Regression models explored the relationship between housing status and pain-related outcomes adjusting for demographics.

Results:

Among participants (N = 496), 47 (9.5%) reported unstable housing. Compared with stably housed patients, those with unstable housing had higher adjusted odds of current (adjusted odds ratio, 2.05; 95% CI, 1.11–3.86) and lifetime chronic pain (adjusted odds ratio, 2.40; 95% CI, 1.28–4.58) and higher pain-related anxiety scores (β = 13.43; 95% CI, 6.51–20.35), depressive symptoms (β = 4.62; 95% CI, 2.83–6.41), and sleep disturbance (β = 2.37; 95% CI, 1.07–3.54). No differences were found in pain self-efficacy (β = −0.78; 95% CI, −6.88 to 3.62).

Discussion:

Unstably housed emergency department patients experience higher rates of chronic pain and pain-related anxiety while maintaining similar pain self-efficacy. Systems-level interventions that address social and structural barriers to pain management among unstably housed patients are needed to improve pain management and reduce emergency department burden. The results should be interpreted with caution given the small sample size and large confidence intervals.

Introduction

Housing instability, ranging from literal homelessness to difficulty paying rent or utilities and fear of losing one’s home,¹ is a growing social determinant of health with serious implications for ED care. On any given night, approximately 771,480 individuals in the United States are unhoused,² and more than 1.5 million spend at least 1 night in temporary housing annually.³ Importantly, housing instability is prevalent among renters in communities where housing costs have outpaced incomes. Although there is not a singular definition of housing instability, in 2023, approximately 21 million renter households (50%) spent more than 30% of their income on rent, placing them at risk of housing loss, meeting the US Department of Housing and Urban Development’s definition of being rent burdened.⁴ Moreover, nearly 40% of individuals experiencing housing instability report at least 1 ED visit annually, 3 times the national average.^3,5 Chronic pain is a common but often underaddressed medical concern in this population. Individual studies report chronic pain prevalence of 43% to 63% among unstably housed individuals^6–8 with pain being 1 of the leading reasons for ED visits in this population.^9,10

Housing instability contributes to both the onset and exacerbation of chronic pain through pathways such as trauma, violent victimization, psychosocial stressors, environmental exposure, inadequate sleeping surfaces, and the physical burden of carrying belongings over long distances.^6,11–14 These challenges are further compounded by systemic barriers to care such as cost, limited service availability, transportation challenges, experiences of stigma in the health care setting, and the logistic complexities of managing pain without stable housing.^7,15 As a result, emergency departments frequently serve as the primary source of pain care for unstably housed individuals,^9,10,15 and may lead to increased health care costs, fragmented care, suboptimal treatment, financial burden, stigma, and strain on health care resources.^15–18 Despite this, chronic pain among unstably housed ED patients remains understudied.¹⁹

Emergency nurses play a central role in pain management, conducting assessments, implementing treatment protocols, and coordinating discharge planning. However, current practices rarely include routine assessment of housing status or adapting pain care to account for patients’ living conditions. This limits nurses’ ability to provide patient-centered care for individuals experiencing housing instability. Moreover, research exploring how housing instability influences pain-related outcomes, such as self-efficacy, functional interference, and pain-related anxiety is limited,¹⁹ leaving nurses with limited evidence to guide tailored interventions.

This study addresses a significant gap in the scientific literature by examining pain-related experiences among ED patients stratified by housing status. The goal of this investigation was to inform more equitable and responsive pain care practices in emergency nursing to help shape future patient-centered approaches to pain in patients experiencing housing instability.

Methods

STUDY DESIGN, SETTING, AND PARTICIPANTS

Data for this secondary analysis came from the Decision-Making Factors for Therapeutic Opioid Use after Emergency Care (project number: K08DA049948–05): mixed-methods investigation of ED patient cognitive factors that influence opioid use after discharge.²⁰ Participants were recruited from 2 urban emergency departments within a midwestern medical system. Potential participants were identified using randomly selected ED bed numbers and then approached by study staff to assess eligibility and interest in participation. Individuals were eligible to participate if they were adult ED patients (aged ≥18 years) and were not actively undergoing emergent resuscitation. Individuals who did not speak English or were cognitively impaired, currently incarcerated or with a guard present, pregnant or breastfeeding, or expressing suicidal thoughts were not eligible to participate. Once eligibility was determined, research assistants obtained an informed consent. Study staff administered the surveys following a guide and entered responses directly into a Research Electronic Data Capture²¹ database using a tablet computer. This ensured accessibility for individuals with varying literacy levels or other physical limitations. Participants received a $20 gift card at enrollment and an additional $20 gift card upon completion of the follow-up interview 30 (±7) days after discharge from the emergency department. The institutional review board at The Ohio State University approved the original study. The secondary analysis of the deidentified data was exempt from IRB approval.

MEASURES

Demographic Characteristics

Participants reported their age in years, sex assigned at birth, gender, race, and ethnicity. Race and ethnicity were combined into a single variable with the following categories: non-Hispanic Black, non-Hispanic white, Hispanic, and non-Hispanic other. The latter included those who identified as American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, another racial identity, and multiracial owing to the small sample size. Gender included (1) man, (2) woman, and (3) transgender and gender diverse individuals. The latter included those who identified as nonbinary, transgender male, transgender female, or another gender identity. Sex was categorized as male or female given that no participants identified as intersex.

Housing Status

Housing status was ascertained using a single question: “What is your living situation today? (Where do you stay?)” with response options: (1) “I have a steady place to live (same living and sleeping space each day)”; (2) “I have a place to live today, but I am worried about losing it in the future”; or (3) “I don’t have a steady place to live (staying with others, hotel, shelter, in a car, outside such as street/abandoned building/park, bus or train station, or in a park).”²² For this analysis, housing status was dichotomized as stable or unstable. The unstable category included both individuals experiencing homelessness and those worried about losing their housing. Differences in pain experiences and management needs likely exist between those who are currently experiencing homelessness and those who currently have a home but are worried about losing it. The 2 groups were combined owing to the small sample size.

Pain Characteristics

Participants rated their initial pain using the Mankoski pain scale,²³ which combines a numeric pain scale with descriptors (eg, 0 is “Pain free” and 10 is “Unconscious, Pain makes you pass out”). For this analysis, the scale was dichotomized as little to no pain (score ≤5) and moderate or severe pain (score, ≥6).²⁴ Participants were asked, “Did you have any pain in the 24 hours prior to coming to the emergency department?” and “Is this pain the main reason you came to the emergency room?” Both questions were operationalized as dichotomous variables (yes/no).

History of Chronic Pain

Participants were asked whether they ever had chronic pain, with response options: “never,” “yes, currently,” and “yes, but not anymore.” Two dichotomous variables were created: current chronic pain (coded as 1 for “yes, currently”; 0 otherwise) and lifetime chronic pain (coded as 1 for both “yes, currently” and “yes, but not anymore”; 0 for “never”).

Pain Interference

Pain interference with day-to-day life was assessed using the Patient Reported Outcomes Measurement Information System pain interference 8-item short form 8a,²⁵ an 8-item questionnaire that asks how much pain interferes with day-to-day life. Responses to each question include: not at all (1), a little bit (2), somewhat (3), quite a bit (4), and very much (5). Responses across all 8 items were summed to create a total pain interference score, with higher scores indicating greater interference (range, 8–40). Internal consistency for the pain self-efficacy questionnaire (PSEQ) in the present sample was excellent (α = 0.96).

Pain Anxiety

Pain anxiety was measured using the pain anxiety symptom scale, short form (PASS-SF),²⁶ a validated 20-item measure that evaluates anxiety symptoms associated with pain. The measure includes 4 subscales: cognitive anxiety, fear of pain, escape/avoidance, and physiological anxiety. Participants rated how often they experienced specific pain-related anxiety symptoms on a 6-point Likert scale ranging from never (0) to always (5). Responses were summed to generate a total pain anxiety score, with higher scores indicating greater anxiety related to pain (range, 0–100). Subscale scores were also calculated by summing the specific items associated with each subscale (range, 0–25). Internal consistency for the PASS-SF in the present sample was excellent (α = 0.93). All scores were operationalized as continuous variables in analysis, with higher scores indicating higher pain anxiety symptoms.

Pain Self-Efficacy

Pain self-efficacy was evaluated with the PSEQ, a validated 10-item instrument designed to measure individuals’ confidence in their ability to engage in daily activities despite pain.²⁷ Participants rated their confidence in performing activities such as “cope with pain in most situations” or “do some form of work despite the pain” on a 7-point Likert scale ranging from not at all confident (0) to completely confident (6). In the present sample, internal consistency for the PSEQ was excellent (α = 0.94). Scores from each item were summed to produce a total pain self-efficacy score (range, 0–60), with higher scores reflecting greater confidence in functioning despite pain.

Sleep Disturbance

Sleep disturbance was determined using the Pittsburgh sleep quality index (PSQI), a validated 19-item measure describing the participant’s self-reported sleep over the last 30 days.²⁸ The PSQI generates a global sleep quality score ranging from 0 to 21, with higher scores indicating poorer sleep quality. A global score of >5 is commonly used to indicate clinically significant sleep disturbance.^29,30 Internal consistency for the PSQI in the present sample was questionable (α = 0.66).

Depression

Depressive symptoms were assessed using the Patient Health Questionnaire (PHQ-9), a widely used, well-validated 9-item self-report measure assessing the frequency of depressive symptoms over the past 2 weeks.³¹ In the present sample, internal consistency was good (α = 0.83). Participants rated how often they experienced each symptom on a 4-point scale ranging from not at all (0) to nearly every day (3). Item scores were summed to produce a total depression severity score (range, 0–27), with higher scores indicating more severe depressive symptoms.

Medical Record Extraction

Clinical characteristics were captured from the electronic health record after discharge. Documentation of “chief complaint,” representing the primary reason for the ED visit, was reviewed from both the nurse and provider notes. In addition, chart documentation was examined for opioid administration, administration of a nonopioid, and an opioid prescription at discharge.

Pain Management Treatments Used After ED Visit

After discharge, participants were contacted at 30 days (±7 days) after enrollment for a follow-up interview that included pain since discharge and modes of pain management. Participants were asked, “When did you last use the following [pain treatments] since leaving the hospital or emergency department?” Response options included: never, not since discharge, yes (past 24 hours), yes in the past 30 days, yes in the past year, and yes longer than 1 year ago. Each treatment was assessed separately, and treatments included prescription medications, over-the-counter medications, injections and topical treatments, and nonpharmacologic treatments such as a transcutaneous electrical nerve stimulation unit and yoga. A dichotomous variable was created for each treatment, classifying participants who answered yes, regardless of how long ago they used the treatment, as having used that specific treatment.

STATISTICAL ANALYSIS

Participant characteristics, responses to the pain interference assessment, and pain management treatments used after the ED visit were summarized using means, SDs, and proportions, as appropriate. Analyses of post-ED treatments included only participants who completed the follow-up survey. All variables, including completion of follow-up surveys, were stratified by housing status (unstable vs stable). Bivariate group differences were assessed using Pearson chi-square test, Fisher exact test, t test, and Mann-Whitney U test as appropriate.

Outcome-specific multivariable regression models were used to assess associations between housing status and pain-related outcomes. Logistic regression was used to determine whether being unstably housed was associated with increased odds of (1) currently having chronic pain, (2) ever having chronic pain, (3) experiencing pain in the 24 hours before the ED visit, (4) pain being the primary reason for the ED visit, (5) rating pain as moderate to severe before any intervention, (6) receiving an opioid in the emergency department, (7) receiving nonsteroidal anti-inflammatory drugs in the emergency department, and (8) receiving a prescription for an opioid at discharge from the emergency department. Linear regression was used to assess whether PHQ-9, PSQI, PSEQ, and PASS-SF (including subscales) scores differed by housing status. All models were adjusted for age, sex, and race/ethnicity. Gender was not included owing to the small sample size of transgender and gender nonconforming participants. Missing data were handled with listwise deletion. Standard diagnostics were performed, and collinearity was assessed using variance inflation factors. No irregularities were noted. All analyses were conducted using RStudio version 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria).

A descriptive analysis of provider-identified chief complaints was conducted using structured key word matching in R. Categories of chief complaints were defined a priori based on clinical relevance, and included categories such as pain, acute problem, chronic problem, mental health, injury/violence, substance use, physiological systems (ie, cardiac, respiratory, neurologic, gastrointestinal, gynecologic), and others. Categories were assigned to each clinician-identified chief complaint in the electronic health record determined upon chart review based on key word presence, with additional logic to classify key word “abuse” under substance use when preceded by a substance use–related term (eg, opioid, alcohol). The data were then stratified by housing status, and the frequency and percentage of each category of chief complaints within housing status groups were calculated.

Results

Of the 880 patients approached for participation, 496 were enrolled in the study. The mean age was 47.3 years (SD = 15.7), 58.8% identified as women, 39.7% male, 1.4% as transgender and gender diverse, and 50.6% identified as non-Hispanic white (Table 1). A total of 47 individuals (9.5%) reported unstable housing, including 20 experiencing homelessness and 27 who were fearful of losing their housing. Of note, 17 individuals (1.9%) were excluded owing to a lack of reliable phone access for follow-up.

TABLE 1.

Demographic and pain characteristics of a random sample of ED patients stratified by housing status

Characteristic	Housed (n = 449)	Unstably housed (n = 47)	P value
Age (y), mean (SD)	47.3 (15.7)	47.3 (13.3)	.989
Gender, n (%)			.771
Man	176 (39.2)	21 (44.7)
Woman	266 (59.2)	26 (55.3)
TGD*	7 (1.6)	0 (0)
Sex, n (%)			.586
Male	177 (39.4)	21 (44.7)
Female	272 (60.6)	26 (55.3)
Race/Ethnicity, n (%)			.004
Non-Hispanic white	238 (53.0)	13 (27.7)
Non-Hispanic Black	164 (36.5)	30 (63.8)
Non-Hispanic other†	30 (6.7)	2 (4.3)
Hispanic	17 (3.8)	2 (4.3)
Current chronic pain, n (%)			.0436
Yes	184 (41.0)	27 (57.4)
No	265 (59.0)	20 (42.6)
Lifetime chronic pain, n (%)			.0154
Yes	189 (42.1)	29 (61.7)
No	260 (57.9)	18 (38.3)
Given an opioid while in ED, n (%)			.224
Yes	142 (31.6)	20 (42.6)
No	294 (65.5)	27 (57.4)
Missing	13 (2.9)	0 (0)
Given a nonopioid while in ED, n (%)			.599
Yes	260 (57.9)	30 (63.8)
No	183 (40.8)	17 (36.2)
Missing	6 (1.3)	0 (0)
Given an RX for opioid at discharge, n (%)			.683
Yes	55 (12.2)	7 (14.9)
No	385 (85.7)	37 (78.7)
Missing	9 (2.0)	3 (6.4)
Had pain 24 h before ED visit, n (%)			.614
Yes	383 (85.3)	42 (89.4)
No	65 (14.5)	5 (10.6)
Missing	1 (0.2)	0 (0)
Pain main reason for ED visit, n (%)			>.9999
Yes	289 (64.4)	32 (68.1)
No	93 (21.0)	10 (22.2)
Missing	66 (14.7)	5 (10.6)
Initial pain rated moderate to severe, n (%)			.3351
Yes	309 (68.8)	37 (78.8)
No	74 (16.5)	5 (10.6)
Missing	66 (14.7)	5 (10.6)
Pain self-efficacy score, mean (SD)	36.5 (16.5)	35.8 (15.8)	.763
Total pain anxiety score, mean (SD)	50.4 (23.0)	64.6 (21.3)	<.001
Pain anxiety cognitive subscore, mean (SD)	15.4 (6.57)	18.1 (6.35)	.00702
Pain anxiety avoidance subscore, mean (SD)	13.4 (6.21)	16.1 (6.44)	.00638
Pain anxiety fear subscore, mean (SD)	11.1 (7.27)	15.6 (7.57)	<.001
Pain anxiety physiological anxiety subscore, mean (SD)	10.6 (6.82)	15.6 (7.57)	<.001
PROMIS pain interference 8-item short form score, mean (SD)	27.9 (10.1)	28.7 (9.76)	.604
PHQ-9 score. mean (SD)	7.19 (5.63)	11.6 (7.67)	<.001
Pittsburgh sleep quality index score, mean (SD)	8.68 (4.05)	11.0 (3.92)	<.001

PHQ-9, Patient Health Questionnaire; PROMIS, Patient Reported Outcomes Measurement Information System; RX, prescription; TGD, transgender and gender diverse.

^*Includes those who identified as nonbinary, transgender male, transgender female, or another gender identity.

^†Includes those who identified as American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, another racial identity, and multiracial.

PAIN HISTORY AND TREATMENT

In fully adjusted models, participants with unstable housing had higher odds of reporting current chronic pain (adjusted odds ratio [aOR], 2.05; 95% CI, 1.11–3.86; P = .02) and lifetime chronic pain (aOR, 2.40; 95% CI, 1.28–4.58; P = .007) than stably housed participants (Table 2). Unstably housed individuals were also more likely to receive an opioid in the emergency department (aOR, 1.90; 95% CI, 1.00–3.59; P = .04) than their housed counterparts. Although unstably housed individuals were 1.8 times more likely to report moderate to severe pain, this result was not statistically significant (aOR, 1.8; 95% CI, 0.73–5.40; P = .28). No significant differences were found in the odds of receiving nonopioid medication for pain, receiving an opioid prescription at discharge, experiencing pain 24 hours before the ED visit, or reporting pain as the primary reason for the ED visit (Table 2). Among the 337 participants (stably housed, 311; unstably housed, 26) who completed the follow-up interview, there were no significant differences in pain management treatments used after the ED visit (Table 3). No significant differences were observed in completion of follow-up interviews by housing status (69% of stably housed participants completed follow-up interviews vs 55% of unstably housed; P = .074).

TABLE 2.

Results of logistic regression models assessing the associations between housing status and pain history and pain management while in the emergency department among a sample of ED patients

Question	Unadjusted OR (95% CI)	Adjusted* OR (95% CI)
Does patient currently have chronic pain	1.9 (1.06–3.61)†,§	2.05 (1.11–3.86)†,§
Has the patient had chronic pain in their lifetime	2.2 (1.21–4.18)‡,§	2.4 (1.28–4.58)‡,§
Receipt of an opioid in the ED	1.5 (0.82–2.81)	1.9 (1.00–3.59)†,§
Receipt of a nonopioid in the ED	1.2 (0.67–2.36)	1.4 (0.77–2.80)
Receipt of opioid prescription at discharge	1.3 (0.52–2.95)	1.6 (0.58–3.46)
Pain in 24 h before ED visit	1.9 (1.03–3.61)	1.4 (0.58–4.26)
Pain primary reason for ED visit	1.04 (0.51–2.30)	0.98 (0.47–2.12)
Initial pain rated moderate to severe	1.8 (0.73–5.29)	1.8 (0.73–5.40)

OR, odds ratio.

^*Model adjusted for age, race/ethnicity, and sex.

^†P < .05.

^‡P ≤.01.

^§Significant results.

TABLE 3.

Pain management strategies used by a sample of ED patients after discharge from the hospital or ED stratified by housing status

Have you taken/used the following since leaving the emergency department	Housed (n = 311) n (%)	Unstably housed (n = 26) n (%)	P value
Acupuncture			>.999
No	306 (68.2)	26 (100)
Yes	4 (0.9)	0 (0)
Missing	1 (0.3)	0 (0)
Physical therapy			.329
No	280 (62.4)	22 (84.6)
Yes	31 (6.9)	4 (15.4)
NSAIDs/Tylenol			.484
No	86 (19.2)	5 (19.2)
Yes	225 (50.1)	21 (80.8)
Muscle relaxer			.438
No	251 (55.9)	23 (88.5)
Yes	60 (13.4)	3 (11.5)
Gabapentin			.697
No	244 (54.3)	19 (73.1)
Yes	67 (14.9)	7 (26.9)
Topical treatments			.84
No	250 (55.7)	20 (76.9)
Yes	60 (13.4)	6 (23.1)
Missing	1 (0.3)	0 (0)
Steroids			.0558
No	272 (87.5)	26 (100)
Yes	39 (12.5)	0 (0)
Injections			>.999
No	290 (93.2)	25 (96.2)
Yes	21 (6.8)	1 (3.8)
Chiropractor			>.999
No	300 (96.5)	26 (100)
Yes	9 (2.9)	0 (0)
Missing	2 (0.6)	0 (0)
Herbal supplements			.521
No	301 (96.8)	25 (96.2)
Yes	8 (2.6)	1 (3.8)
Missing	2 (0.6)	0 (0)
Meditation			.673
No	244 (78.5)	19 (73.1)
Yes	66 (21.2)	7 (26.9)
Missing	1 (0.3)	0 (0)
Heat/ice			>.999
No	193 (62.1)	16 (61.5)
Yes	118 (37.9)	10 (38.5)
Massage			>.999
No	280 (90.0)	24 (92.3)
Yes	29 (9.3)	2 (7.7)
Missing	2 (0.6)	0 (0)
Sauna			>.999
No	306 (98.4)	26 (100)
Yes	3 (1.0)	0 (0)
Missing	2 (0.6)	0 (0)
Yoga			.974
No	222 (71.4)	18 (69.2)
Yes	88 (28.3)	8 (30.8)
Missing	1 (0.3)	0 (0)
Electrical			>.999
No	304 (97.7)	26 (100)
Yes	6 (1.9)	0 (0)
Missing	1 (0.3)	0 (0)
Completed follow-up survey			.7428
No	138 (30.7)	21 (44.7)
Yes	311 (69.2)	26 (55.3)

NSAID, nonsteroidal anti-inflammatory drug.

REASONS FOR ED VISITS

Pain-related complaints were the predominant primary reasons for ED visits in both the nurse (n = 263; 53.0%) and provider (n = 286; 57.7%) documented chief complaints (Figure). Among persons with stable housing, the most frequent thematic reasons for ED visits were pain (n = 258; 64.3%), acute issues (n = 61; 15.2%), and mental health concerns (n = 26; 6.5%). For persons with unstable housing, the most frequent themes for ED visits were pain (n = 28; 63.6%), acute issues (n = 5; 11.4%), injury/violence (n = 4; 9.1%), and respiratory issues (n = 4; 9.1%).

FIGURE.

Provider documented themes for the reason for the ED visit by housing status.

PAIN INTERFERENCE

The mean pain interference score was 27.9 (SD = 10.1) among stably housed individuals and 28.7 (SD = 9.76) among housed individuals. There was no significant difference in total pain interference scores by housing status (β = 1.73; 95% CI, −1.47 to 4.94; P = .29) (Table 4).

TABLE 4.

Results of linear regression models testing differences in pain interference, depressive symptoms, sleep disturbances, pain self-efficacy, and pain anxiety scores by housing status among a random sample of ED patients

Scale	Unadjusted model		Adjusted model*
	β (SE)	95% CI	β (SE)	95% CI
Total score for pain interference†	0.83 (1.64)	−2.40 to 4.06	1.73 (1.63)	−1.47 to 4.94
PhQ-9 score	4.42 (0.90)‡,††	2.64–6.20††	4.62 (0.91)‡,††	2.83–6.41††
Pittsburgh sleep quality index score	2.37 (0.62)‡,††	1.15–3.58††	2.30 (0.63)‡,††	1.07–3.54††
Pain self-efficacy§	−0.78 (2.67)	−6.02 to 4.46	−1.63 (2.67)	−6.88 to 3.62
Pain anxiety¶	14.2 (3.50)‡,††	7.32–21.08††	13.43 (3.52)‡,††	6.51–20.35††
Pain anxiety cognitive subscore	2.73 (1.00)∥,††	0.76–4.70††	2.83 (1.02)∥,††	0.83–4.82††
Pain anxiety avoidance subscale#	2.79 (0.96)∥,††	0.91–4.67††	2.41 (0.96)**,††	0.52–4.28††
Pain anxiety fear subscale#	4.47 (1.12)‡,††	2.29–6.69††	3.96 (1.12)‡,††	1.75–6.16††
Pain anxiety physical anxiety subscale#	4.12 (1.05)‡,††	2.06–6.18††	4.20 (1.05)‡,††	2.14–6.26††

PHQ-9, Patient Health Questionnaire.

^*Model adjusted for age, race/ethnicity, and sex.

^†Assessed using the Patient Reported Outcomes Measurement Information System pain interference 8-item short form.

^‡P ≤ .001.

^§Assessed using the pain self-efficacy questionnaire.

^¶Assessed using the pain anxiety symptom scale, short form (PASS-SF).

^∥P < .01.

^#Subscale of the PASS-SF.

^**P < .05.

^††Significant results.

DEPRESSION AND SLEEP SCORES

The mean PHQ-9 score among unstably housed individuals was 11.6 (SD = 7.7), indicating mild depressive symptoms. The mean PHQ-9 score for stably housed individuals was 7.19 (SD = 5.63) (Table 1). Compared with stably housed participants, those with unstable housing had significantly higher PHQ-9 scores (β = 4.62; 95% CI, 2.83–6.41; P < .001).

The mean PSQI score for the full sample was 8.9 (SD = 4.1). Unstably housed participants had a mean score of 11.04 (SD = 3.9) compared with 8.7 (SD = 4.0) among stably housed participants. Unstable housing was associated with significantly higher PSQI scores (β = 2.30; 95% CI, 1.07–3.54; P < .001) (Table 4).

PAIN ANXIETY AND PAIN SELF-EFFICACY

The mean PSEQ score for the sample was 36.47 (SD = 16.40), indicating a moderate degree of confidence in managing pain among participants. The mean score for unstably housed individuals was 35.8 (SD = 15.8) and 36.5 (SD = 16.5) among stably housed individuals (Table 1). There was no statistically significant difference in PSEQ scores by housing status (β = −1.63; 95% CI, −6.88 to 3.62; P = .54) (Table 4).

The mean PASS-SF score for the sample was 51.8 (SD = 23.2), indicating a moderate level of pain anxiety. The mean score for unstably housed participants was 64.6 (SD = 21.3) compared with 50.4 (SD = 23.0) for stably housed participants (Table 1). In fully adjusted models, unstably housed participants had higher total PASS-SF scores (β = 13.43; 95% CI, 6.51–20.35; P < .001) and higher scores on all 4 subscales: cognitive (β = 2.83; 95% CI, 0.83–4.82; P = .006), avoidance (β = 2.41; 95% CI, 0.52–4.28; P = .01), fear (β = 3.96; 95% CI, 1.75–6.16; P < .001), and physiological anxiety (β = 4.20; 95% CI, 2.14–6.26; P < .001) (Table 4).

Discussion

This study measured the relationship between housing stability and pain experiences among a random sample of ED patients, revealing associations between unstable housing and multiple pain-related outcomes. This study is among the first to examine pain characteristics in depth, including pain anxiety and self-efficacy, in this population. Consistent with previous research,^7,8,19,32 unstably housed participants were twice as likely as stably housed participants to report current and lifetime chronic pain. In addition, unstably housed individuals were nearly twice as likely to report moderate to severe pain compared with their stably housed peers, a difference that, although not statistically significant, is clinically meaningful, particularly in the context of a small sample that may make it difficult to detect statistical differences.

These findings support the well-documented bidirectional relationship between housing instability and chronic pain.^6,7,33,34 Chronic pain can precipitate housing instability through reduced work capacity, medical debt, and economic precarity. In turn, housing instability exacerbates pain through environmental exposures, stress, violence, inadequate sleep, physical strain from carrying belongings, and limited health care access. This mutually reinforcing cycle suggests that addressing housing instability is potentially an important component of comprehensive pain management.

Unstably housed individuals were almost twice as likely to receive an opioid in the emergency department. This may reflect appropriate treatment of more acute injuries and severe pain presentations in this population. This is consistent with the finding that pain, acute issues, and injuries/violence were among the top 3 reasons unstably housed participants sought care in the emergency department. Importantly, no significant differences were observed in opioid prescriptions at discharge or in the use of nonsteroidal anti-inflammatory drugs, suggesting that the observed difference was specific to acute ED treatment rather than broader prescribing patterns.

Contrary to expectations and previous literature,^32,35 no significant differences were identified in postdischarge use of nonopioid pain management treatments (eg, physical therapy, yoga, and acetaminophen) by housing status. Several factors may explain these results: participants may have accessed treatments before the onset of housing instability, insurance coverage may have temporarily facilitated access to services, and overall utilization of some therapies was low across the entire sample. Importantly, self-reported use of a treatment does not indicate adequate duration, adherence, or efficacy, factors that may differ substantially by housing status even when initial access appears similar.

Participants experiencing housing instability had significantly higher PHQ-9 and PSQI scores, indicating increased depressive symptoms and sleep disturbance, findings that are consistent with previous research.^8,36–38 Poor sleep quality is known to exacerbate pain intensity, and both pain intensity and depressive symptoms may contribute to increased health care utilization and ED visits.^30,39,40 These comorbidities highlight the importance of comprehensive screening for mental health and sleep disturbances in ED settings, particularly among unstably housed patients who may face barriers to accessing routine care for these conditions.

Pain-related anxiety was also elevated among unstably housed individuals across all 4 PASS-SF subscales (cognitive anxiety, avoidance behaviors, fear of pain, and physiological responses) compared with stably housed participants. In contrast, pain self-efficacy scores did not differ by housing status. This pattern may suggest that individuals experiencing housing instability retain confidence in their ability to deal with their pain but face social and structural barriers that inhibit implementation of pain management strategies, which could contribute to increased pain anxiety. These barriers may include inconsistent access to safe sleeping environments; nonpharmacologic treatments and medications, including over-the-counter medications; and mental health care and fear of stigmatization in health care settings.^11,41–43 Previous research indicates that unstably housed individuals experience judgment and dismissal in health care settings,^41,43 which may deter them from seeking medical treatment for pain.

These findings suggest pain-related anxiety among unstably housed individuals may be driven, in part, by external, modifiable factors such as a lack of access to stable and safe environments and equitable care, rather than individual factors. Interventions to address pain in this population should extend beyond the individual-level approaches and include low-barrier pain management strategies, access to housing, and medical respite. However, these results should be interpreted with caution owing to the small sample size of unstably housed participants in the sample. Future research with larger samples is needed.

Limitations

Several limitations should be considered when interpreting our results. First, the original study’s cross-sectional design limits the ability to establish temporality and assess causation. Second, our use of self-report measures may have introduced social desirability and recall bias. Third, generalizability is limited because participants were recruited from 2 urban emergency departments within the same hospital system and city, which may not reflect experiences in other geographic or clinical settings. In addition, our study only included unstably housed individuals who accessed care in the emergency department. Those who are unstably housed and do not access care may have different pain experiences and management needs. The small sample size further limits generalizability and reduces the likelihood that our participants are representative of the broader population.

Fourth, standardized pain assessment measures may not be relevant to the lived experience of unstably housed individuals.³⁹ For example, the pain interference scale and PSEQ include items assessing interference with activities that assume stable housing circumstances (eg, household chores, home maintenance). Participants experiencing housing instability may have reported minimal interference not caused by the absence of pain impact, but because these activities were not applicable to their current living situation, potentially under-representing the true impact of pain on daily function. Fifth, although the percentage of unstably housed participants in the sample was consistent with estimates of housing instability in health care settings,³ the overall sample size was small, which may have been underpowered, leading to potential type II errors. These results should be interpreted with caution. In addition, the limited sample size precluded assessment of differences in pain characteristics and outcomes by gender. This is an important area for future research, given that existing evidence demonstrates that women experiencing housing instability face unique challenges and consistently experience worse health outcomes.^44–47 Developing tailored interventions to address the specific needs of women may be needed. Finally, combining unstably housed and unhoused individuals as a single group may have obscured important differences in pain experiences and pain management utilization among housed, unhoused, and unstably housed individuals. Future research should differentiate the pain experience and management needs of unstably housed and unhoused individuals.

Finally, retention and follow-up among participants experiencing housing instability are a challenge. Of the individuals approached for participation, 1.9% were excluded owing to a lack of reliable cell phone access for follow-up. Although studies indicate many participants experiencing housing instability retain reliable cell phone access for follow-up,⁴⁸ barriers including data plans, difficulty charging, and loss or theft may introduce additional bias.

Implications for Emergency Nurses

The impact of social determinants on health must be considered by policy makers, hospital administrators, and clinical prescribers, but especially the emergency nurse. At the bedside, emergency nurses are positioned to recognize patient needs and support navigation toward solutions. They are uniquely positioned to implement trauma-informed, patient-centered strategies that address pain. Examples include offering comfort measures, avoiding stigmatizing language, and conducting pain assessments using trauma-informed, unbiased approaches. These practices reflect the core values of emergency nursing and are especially important when caring for unstably housed individuals. Emergency nurses are also well positioned to advocate for equitable, multimodal pain management that addresses the complex needs of unstably housed individuals. However, ED discharge practices often fail to consider housing status when making discharge plans, potentially setting patients up for treatment failure and repeat visits.⁴⁹ These findings suggest that, before discharge from the emergency department, clinicians should assess housing status and consider whether prescribed pain management treatments are feasible given the patient’s circumstances.

Through recognition of the health-housing intersections, emergency nurses play a crucial role in mitigating these disparities and promoting patient-centered outcomes. Evidence-based nurse-led interventions to support pain management among unstably housed patients in the emergency department include systematic screening for housing stability as part of pain assessment protocols, providing documentation that daytime access to shelters is medically necessary for those with acute or chronic pain, and facilitating warm handoffs to community resources such as housing first programs, medical respite care, palliative care, community health centers, and case management services.

Systems-level interventions are needed to break the cycle of unstable housing and chronic pain. Discharging an individual to unsafe conditions such as the street is a driver of ED recidivism and can be life limiting. ED protocols should incorporate screening for housing instability into standard practice, during pain assessments, and during discharge planning.^50–52 Embedding clinical supports such as community health workers or health promotion advocate models in the emergency department can facilitate warm handoffs to housing programs, referrals to palliative care, frequent touch base after discharge, and communication among the clinical team. Policy efforts should consider prioritizing housing first interventions, respite care funding, and continuity of care through community connections and case management. Housing should be recognized as a part of effective pain management care, and emergency departments should be equipped to address these needs, which have potential for improved long-term outcomes.^53,54

Conclusion

This study provides important insight into the relationship between housing instability and pain experiences among ED patients. Unstably housed individuals demonstrated higher odds of chronic pain, increased pain-related anxiety, and greater depressive symptoms and sleep disturbances. Notably, although pain-related anxiety was elevated, there was no difference in pain self-efficacy by housing status, lending support to the notion that structural barriers to pain management and care contribute to pain anxiety rather than individual capabilities.

These findings highlight the need for health care systems to recognize housing as a social determinant of health that affects pain outcomes. Emergency departments are well positioned to implement housing screening protocols and multidisciplinary interventions that address both immediate pain management and housing stability needs. Comprehensive systems-level interventions that address social and structural barriers to pain management among unstably housed patients are needed to improve pain management and reduce ED burden. The emergency department is more than a site for acute care; it provides a safety net and a critical access point for a range of health care services, particularly for individuals experiencing housing instability.

Contribution to Emergency Nursing Practice.

• Unstably housed individuals are disproportionately affected by chronic pain and often rely on emergency departments for pain management owing to barriers to accessing care.

• Unstably housed individuals had similar pain self-efficacy and higher pain anxiety scores compared with stably housed ED patients, suggesting comfort with managing pain but limited ability to implement pain management, potentially owing to social and structural barriers.

• Housing status should inform ED pain assessments and discharge planning. Emergency nurses are uniquely positioned to implement nurse-led interventions across care settings to support pain management in unstably housed individuals.

Data, Code, and Research Materials Availability

This activity was approved by The Ohio State University institutional review board. Participation was voluntary, and an informed consent was obtained before completing interviews. All participants were compensated with a $20 gift card at each interview.