Impacts of social determinants of health on chronic opioid therapy for chronic non-cancer pain

Abstract

Aim: We aimed to investigate the association between social determinants of health and chronic opioid therapy.

Materials & methods: We conducted a retrospective analysis of electronic health records from five family medicine and internal medicine clinics in Oregon in 2020 and 2021. Our outcome variable was whether a patient was receiving chronic opioid therapy for chronic non-cancer pain. Our variables of interest included financial difficulty, insurance types, transportation barriers, currently married or living with a partner and organizations participation.

Results: Our results showed that patients with financial difficulty were more likely to have chronic opioid therapy (OR: 2.69; 95% CI: 1.14, 6.33).

Conclusion: Addressing patients’ social determinants of health disadvantages is important for optimizing pain management.

Plain Language Summary

What is this article about?

Addressing the opioid crisis is a national priority in the USA. Our objective was to focus on a broad set of social determinants of health (SDOH) and examine whether patients with SDOH disadvantages were more likely to receive chronic opioid therapy for chronic non-cancer pain. Current literature has not assessed some important SDOH characteristics. We aimed to address this limitation by using electronic health records that incorporated SDOH data.

What were the results?

Patients with financial difficulty in this study had approximately two-times higher odds of receiving chronic opioid therapy.

What do the results of the study mean?

Our study has important clinical and policy implications. Clinicians should screen for patient SDOH disadvantages and provide support as an integral part of patient-centered pain management. Payers and policymakers should also consider expanding coverage and reimbursement for multimodal treatments for pain.

Plain language summary

Article highlights.

• Addressing the opioid crisis is a national priority in the USA.

• Literature suggests that opioid therapy is associated with higher risks of opioid overdose and misuse, and therefore a key solution to the opioid crisis is to reduce chronic opioid therapy.

• Early research has not assessed some important social determinants of health (SDOH) that could affect pain management.

• Our current study tested a hypothesis that patients with SDOH disadvantages were more likely to receive chronic opioid therapy for chronic non-cancer pain.

• We used electronic health records incorporated with patient-reported data to examine a broader set of SDOH measures.

• Patients with financial difficulty in our study had significantly higher odds of receiving chronic opioid therapy.

• It is important to increase coverage and reimbursement for nonopioid treatment options to mitigate the risks associated with chronic opioid use.

• Clinicians' time and efforts to screen for patients’ SDOH disadvantages should also be appropriately reimbursed.

1. Background

About 80,000 opioid-involved drug overdose deaths were reported in 2022 in the USA [1]. It is a national priority to beat this evolving opioid crisis and to improve pain treatments [2,3]. Literature suggests that opioid therapy is associated with higher risks of opioid overdose and misuse [4–7]. For example, a study showed that prescribed opioids, especially for chronic use, were associated with significantly higher risks of opioid use disorders among patients with chronic non-cancer pain [5]. The 2022 Centers for Disease Control and Prevention (CDC) clinical practice guideline for prescribing opioids for pain highlighted the importance of assessing risks versus benefits of long-term opioid therapy and prioritizing nonopioid evidence-based treatments for pain [8]. The National Institutes of Health launched the Helping to End Addiction Long-term (HEAL) initiative to support science-based solutions to opioid public health emergencies and to improve prevention and treatment for opioid misuse and addiction [9]. The Pain Management Best Practices Inter-Agency Task Force Report published by the US Department of Health and Human Services (HHS) also recommended using nonopioid therapies to mitigate the risks associated with chronic opioid use [10].

To support this national priority, we aim to examine how patients' social determinants of health (SDOH) may affect their risks of receiving chronic opioid therapy for chronic non-cancer pain, which is an imperative but still understudied area. SDOH are the economic and social conditions which have important influences on health inequities [11,12]. Early research found that patients with SDOH disadvantages were less likely to attend appointments for chronic pain services [13,14]. A recent study by Cheng and colleagues focused on chronic musculoskeletal pain and found that patients with chronic opioid therapy were more likely to live in communities among the most socially disadvantaged national quartile [15]. Another study using survey data in Canada found that food-insecure individuals had higher odds of using prescription opioids in the past year [16]. On the other hand, current literature has not assessed some important SDOH characteristics that could also affect pain management, such as marital status, organization participation and transportation barriers.

In this study, we tested a hypothesis that patients with SDOH disadvantages were more likely to receive chronic opioid therapy for chronic non-cancer pain. A primary strength of our study is that we used electronic health records (EHRs) incorporated with patient-reported data to examine a broader set of SDOH measures. Literature suggests that incorporating patient-reported data in EHRs offers great potential to improve comprehensive research and promote patient-centered care [17]. The Institute of Medicine also recommended capturing patients’ SDOH measures in EHRs, which can enable more informative research and improve patient care and health system design [18].

2. Materials & methods

We conducted this retrospective study using EHRs at five family medicine and internal medicine clinics in Oregon from 2020 to 2021. The EHRs of these clinics were optimized by adding a new module for providers to document patients’ SDOH characteristics. This recent update of EHRs provided us with an important opportunity to directly measure patients’ SDOH. Our study population included adults who had diagnosis of chronic non-cancer pain. Our study population was consistent with other studies on chronic opioid therapy [5,6,19]. Our outcome variable was whether a patient was an active chronic user of opioid therapy, and we identified this outcome by two methods. First, we defined this binary dependent variable as “yes” if a patient had an opioid medication ordered within the last 180 days where the medication was supposed to be used for at least 90 days and that medication was prescribed or dispensed. Similar definition has been commonly used in previous studies on chronic opioid use [5,19–21]. In addition, medication management agreements have also been used to document chronic opioid therapy [22–24]. Therefore, we also identified patients on chronic opioid therapy if they had medication management agreements on file within the last year. Other researchers have used similar definitions in literature [22].

As our main interest in this study, we constructed a series of SDOH variables, including financial difficulty (coded “yes” if a patient reported somewhat hard, hard or very hard to pay for the very basics like food, housing, medical care and heating), transportation barriers (coded “yes” if a patient reported lack of transportation had kept them from medical appointments, work, getting things needed for daily living, or getting medications in the past 12 months), currently married or living with a partner (coded “yes” if a patient was married or living with partner), and organizations participation (coded “yes” if a patient belonged to any clubs or organizations). We also constructed a categorical variable of insurance type and analyzed various types of insurance such as commercial insurance (including Medicare Advantage), Medicare, Medicaid or others, as suggested by the literature [13].

Our control variables included comorbidities (COPD, asthma, obesity, diabetes, hypertension, chronic heart failure, coronary artery disease and chronic kidney disease), psychological symptoms (whether a patient reported feeling depressed, hopeless, stress, or little interest in doing things, or unable to sleep at night), and binge drinking problem. We also controlled for demographic variables such as age, sex and race.

We included the most recently entered data for variables in analysis. Because SDOH were variables of interest in our study, individuals who had missing values on all these variables were excluded from our analysis. Our main analysis focused on individuals with complete data, and we used a logit model to analyze the association between chronic opioid therapy and SDOH. In addition, we conducted sensitivity analysis by using multivariate imputations by chained equations (MICE) to account for missing data with 100 MICE imputations. MICE is a widely used imputation method due to its flexibility and robustness [25]. We used SAS version 9.4 (SAS Institute) and Stata version 14 (StataCorp) for statistical analyses. Statistical significance was defined as p < 0.05. Cell size or derived cell size containing a value less than 11 are not reported according to cell size suppression policy to protect data confidentiality [26]. Our study was approved by the Institutional Review Boards at Samaritan Health Services and Oregon State University.

3. Results

Our data included 325 patients with chronic non-cancer pain who had complete data. As shown in Table 1, around 18% of these patients received chronic opioid therapy. Regarding SDOH measures, about one out of five patients reported financial difficulty, over 60% had commercial insurance, about 5% reported transportation barriers, about two-thirds were married or living with partner, and around 35% reported belonging to organizations. Most of our samples were female and had psychological issues and comorbidities. Less than 10% had binge drinking.

Table 1.

Descriptive results of complete cases of adults with chronic non-cancer pain (n = 325).

	Frequency	Percent (%)
Chronic opioid therapy	60	18.46
Financial difficulty	69	21.23
Insurance type
Commercial	205	63.08
Medicare	56	17.23
Medicaid	>54†	>15
Other	<11†	<5
Transportation barrier	15	4.62
Married or lived with partner	213	65.54
Belong to organization	111	34.15
Comorbidities	276	84.92
Psychological issue	196	60.31
Binge drinking	26	8
Age
18–44	65	20.00
45–64	109	33.54
65–74	107	32.92
75 and above	44	13.54
Female	223	68.62
Not-White	<11†	<5†

^†Small cell size (<11) or derived small size in other cells are not reported to protect data confidentiality.

Table 2 shows results of our analysis restricted to individuals who had complete data. We found that patients who reported financial difficulty had significantly higher odds of receiving chronic opioid therapy (OR: 2.69; 95% CI: 1.14, 6.33). Having Medicare insurance was also associated with higher odds of receiving chronic opioid therapy, although the association was marginally significant.

Table 2.

Odds ratios of chronic opioid therapy among adults with chronic non-cancer pain in analysis of complete cases (n = 325).

	Odds ratio (SE†)	p-value	95% Confidence interval
Financial difficulty (ref = no)	2.69 (1.17)	0.024‡	1.14	6.33
Insurance type (ref = Commercial)
Medicare	2.46 (1.11)	0.048‡	1.01	5.98
Medicaid	0.68 (0.39)	0.496	0.22	2.06
Other	0.58 (0.42)	0.450	0.14	2.36
Transportation barrier (ref = no)	1.34 (1.03)	0.706	0.29	6.08
Married or lived with partner (ref = no)	1.75 (0.71)	0.173	0.78	3.89
Belong to organization (ref = no)	0.69 (0.27)	0.335	0.33	1.47
Comorbidities (ref = no)	1.13 (0.60)	0.812	0.40	3.19
Psychological issue (ref = no)	1.34 (0.57)	0.489	0.58	3.08
Binge drinking (ref = no)	0.64 (0.42)	0.503	0.18	2.35
Age (ref = 18–44)
45–64	0.85 (0.42)	0.741	0.32	2.23
65–74	0.53 (0.30)	0.266	0.17	1.62
75 and above	0.90 (0.64)	0.883	0.23	3.59
Female (ref = male)	1.19 (0.44)	0.640	0.57	2.48
Not-White (ref = white)	1.01 (0.76)	0.988	0.23	4.40

^†SE were estimated based on bootstrap of 10,000 repetitions.

^‡p < 0.05.

SE: Standard errors.

Table 3 shows our sensitivity analysis results. We completed imputations for 8555 adults with chronic non-cancer pain. Our main finding remains robust that having financial difficulty was associated with significantly higher odds of chronic opioid therapy (OR: 3.16; 95% CI: 1.84, 5.42).

Table 3.

Odds ratios of chronic opioid therapy among adults with chronic non-cancer pain in imputation analysis model (n = 8555).

	Odds ratio (SE)	p-value	95% Confidence interval
Financial difficulty (ref = no)	3.16 (0.86)	<0.001†	1.84	5.42
Insurance type (ref = Commercial)
Medicare	1.35 (0.14)	0.004‡	1.10	1.65
Medicaid	0.89 (0.12)	0.381	0.69	1.15
Other	0.87 (0.22)	0.571	0.53	1.42
Transportation barrier (ref = no)	1.08 (0.52)	0.880	0.42	2.78
Married or lived with partner (ref = no)	1.32 (0.37)	0.325	0.76	2.29
Belong to organization (ref = no)	0.64 (0.18)	0.118	0.37	1.12
Comorbidities (ref = no)	1.96 (0.26)	<0.001§	1.51	2.54
Psychological issue (ref = no)	1.09 (0.14)	0.482	0.85	1.40
Binge drinking (ref = no)	0.52 (0.21)	0.111	0.24	1.16
Age (ref = 18–44)
45–64	2.36 (0.29)	<0.001§	1.86	3.00
65–74	2.69 (0.43)	<0.001§	1.96	3.70
75 and above	2.90 (0.52)	<0.001§	2.04	4.12
Female (ref = male)	1.02 (0.10)	0.840	0.83	1.25
Not-White (ref = white)	0.81 (0.19)	0.380	0.52	1.29

^†p < 0.05.

^‡p < 0.01.

^§p < 0.001.

SE: Standard error.

4. Discussion

Our goal was to address the knowledge gap of how SDOH disadvantages may affect opioid therapy among patients with chronic non-cancer pain. Patients with financial difficulty in our study had approximately two-times higher odds of receiving chronic opioid therapy. It is possible that these individuals had financial barriers to nonopioid options of pain treatments. The HHS Task Force report pointed out the lack of reimbursement and coverage for diversified, multimodal interventions for pain [10]. For example, acupuncture and massage are recognized as complementary and integrative therapies for pain. However, a study found these nonopioid treatments were not covered in most states [27]. Another study analyzed commercial and public insurance coverages and found the median in-network co-payments for occupational and physical therapies were around US$30–40, while the co-insurance for out-of-network providers were around 30% [28]. These high out-of-pocket costs, plus the general lack of insurance coverage for alternative pain therapies imposed substantial burden on patients, especially those who already experienced financial difficulty to pay for other basic needs in life, keeping them from access to nonopioid but more expensive treatment options.

Financial difficulty can also negatively impact patient adherence to pain management. For example, the CDC guideline recommended regular follow-up and re-evaluation of benefits and risks of continuing opioid therapy with patients [8]. The follow-up and re-evaluation could prevent unintentional initiation of long-term opioid therapy, but the CDC guideline further pointed out that patients’ ability for frequent follow-up visits could be affected by their financial constraints [8].

Our analysis also showed that patients with Medicare insurance had higher odds of receiving chronic opioid therapy. As Medicare is health insurance for people aged 65 or older and for those with disabilities, it is possible that older age and having comorbidities could increase the risk of chronic opioid therapy. This is consistent with our imputation analysis findings.

Our findings not only inform clinical practice but also aid policy making. Recently, the Centers for Medicare & Medicaid Services (CMS) identified a set of streamlined quality measures to promote high-quality, safe and equitable care [29]. Among CMS prioritized measures is screening for SDOH [29]. Our current study highlights the importance of understanding and addressing patients’ SDOH disadvantages as an integral part of improving patient-centered care and optimizing treatments for chronic pain. For high-risk patients like those who had financial difficulty, patient education and consultation should start early. More importantly, clinicians should have intervention plans in place when acquiring patients’ SDOH information [18]. A comprehensive pain management plan should include assistance and referrals to connect patients with financial and social support.

Meanwhile, we urge payers to develop appropriate reimbursement and incentives for clinicians’ time and efforts to screen for patients’ SDOH disadvantages and help patients with their challenges. This is particularly critical for clinicians who provide chronic pain therapies disproportionately for patients with financial difficulty. These clinicians would need extra resources to promote equity as well as to achieve high-quality care [30]. The new CMS approach of Rewarding Excellence for Underserved Populations, which has been put into practice in several quality and value programs, sets up a successful example to provide substantive incentives to care for underserved populations [30]. We therefore call for policymakers to adopt and expand similar reward programs to improve chronic pain treatments for patients with SDOH disadvantages.

Finally, payers should increase coverage and reimbursement for nonopioid treatment options. As discussed, a multimodal approach is essential to achieve optimal treatment for chronic pain. Such an approach must be supported by improving patients’ affordability of a variety of alternative, nonopioid treatments [10]. This is particularly important to help patients with financial difficulty to reduce chronic opioid therapy. Future study should assess whether increasing coverage for nonopioid treatment options will eventually lead to reducing medical and social costs incurred by opioid overdose and misuse.

Our current study has limitations. As a retrospective study using EHR data, our findings suggest associations between financial difficulty and chronic opioid therapy but cannot ascertain causal relationship. This is a limitation common to observational research. Randomized controlled trials may provide stronger evidence of causal effects, but such trials will be difficult and expensive to implement [5]. Our current study is an imperative first step toward addressing the knowledge gap in this understudied area. In addition, our data contained missing values, which as literature points out, are unavoidable in clinical research [31]. In our study, missing values can be mainly due to clinicians’ low adoption or lack of experience of using the new SDOH collection modules in EHRs. Literature shows that, despite the great potential of using EHRs data to improve healthcare and research, there are multilevel, complex barriers to physician adoption of adding any data to EHRs [18,32]. More resources are needed to improve SDOH data collection in EHRs to support more comprehensive research in future.

5. Conclusion

Our study identified financial difficulty as a risk factor of chronic opioid therapy. Our findings highlight the importance of screening for SDOH disadvantages to reduce the risk associated with chronic opioid use and to improve patient-centered pain management. Accordingly, payers and policymakers should improve coverage to mitigate patients’ financial barriers to nonopioid treatment options. Clinicians’ efforts to support patients with SDOH disadvantages also warrant appropriate reimbursement.

Funding Statement

This work was supported by the Good Samaritan Hospital Foundation John C. Erkkila, M.D. Endowment for Health and Human Performance. The content in this article is solely the responsibility of the authors.

Author contributions

Each author contributed substantive research, writing, editing and approved the manuscript.

Financial disclosure

This work was supported by the Good Samaritan Hospital Foundation John C. Erkkila, M.D. Endowment for Health and Human Performance. The content in this article is solely the responsibility of the authors. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

We requested a waiver of consent because our study meet criteria below: The research involves no more than minimal risk; and the waiver or alteration will not adversely affect the rights and welfare of the subjects; and the research could not practicably be carried out without the waiver or alteration; and whenever appropriate, the subjects will be provided with additional pertinent information after participation. Our study was approved by the Institutional Review Boards at Samaritan Health Services (IRB21-067) and Oregon State University (HE-2023-329).

Previous presentation

Preliminary findings from this study have been presented at: American Society of Anesthesiologists Annual Meeting, October 2023, San Francisco, California, USA (poster abstract not published).