Abstract
Background
Autoimmune rheumatic diseases (ARD) have distinct pathogenic mechanisms and are causes of disability and increased mortality worldwide. The goal of this study was to examine annual trends in pain management modalities among patients with ARD.
Methods
We identified newly diagnosed patients in the Merative™ Marketscan® Research Databases (2007–2021) with ankylosing spondylitis (AS), psoriatic arthritis (PsA), rheumatoid arthritis (RA), Sjogren’s syndrome (SjS), systemic sclerosis (SSc), and systemic lupus erythematosus (SLE). We determined the incidence of patients treated with opioids, anticonvulsants, antidepressants, skeletal muscle relaxants, NSAIDs, topical analgesics, and physical therapy in the year following diagnosis annually. Logistic regression was used to estimate the association between calendar time and outcomes.
Findings
The cohort included 141,962 patients (AS [n=10,927], PsA [n=21,438], RA [n=71,393], SjS [n=16,718], SLE [n=18,018], and SSc [n=3,468]). Opioid use increased from 26% in 2008 to 38% in 2014 and reduced to 24% in 2020. The odds of opioid use increased annually by 4% until 2014 [aOR 1.04, 95% CI: (1.03–1.04)] and decreased annually by 15% after 2014 [aOR 0.85, 95% CI: (0.84–0.86)]. Physical therapy use increased annually by 5% until 2014 [aOR 1.05, 95% CI: (1.04–1.06)], decreasing annually by 1% after 2014. Anticonvulsant use increased annually by 8% until 2014 [aOR 1.08, 95% CI: (1.07–1.08)], and subsequently stabilized [aOR 1.00 (0.99–1.00)]. NSAIDS use decreased annually by 5% after 2014 [aOR 0.95, 95% CI: (0.95, 0.96)]. Trends did not differ by gender except for NSAID and topical analgesic use.
Interpretation
Since 2014, the use of nonopioid pain management modalities has increased or stabilized, while opioid and NSAID use has declined. Future studies are needed to evaluate the effectiveness of these changes, and the effects they have had on outcomes such as quality of life, disability, and function.
Introduction
Autoimmune rheumatic diseases (ARD) cause disability, functional limitation, reduced quality of life, and increased mortality worldwide. Chronic pain is common and disproportionately impacts patients with ARD despite the advent of new, highly potent and targeted therapies.1 An estimated 36–62% of patients with ARD have a concomitant diagnosis of a chronic pain syndrome (e.g., chronic low back pain, migraine, and fibromyalgia), with higher prevalence in Black patients and those on public insurance1. The presence of one or more chronic pain syndrome is associated with higher prevalence of depression, anxiety, as well as higher frequencies of emergency department visits, surgeries, and hospitalizations, compared with ARD patients without chronic pain diagnoses1.
There are limited data on the management of pain in ARD, particularly less common diseases such as Sjogren’s syndrome (SjS) and systemic sclerosis (SSc), with much of the recent literature focusing on opioid use and misuse. Clinical guidelines for the management of pain in ARD are dated and do not reflect contemporary trends in pain management, such as the shift away from chronic opioid use2. Indeed, the most recent publication of the Pain Management Task Force of the American College of Rheumatology was published in 20103. Treatment guidelines for the management of inflammatory musculoskeletal pain in SjS recommend the use of hydroxychloroquine as first line treatment, although the strength of evidence was weak4. Treatment considerations also include the use of methotrexate, hydroxychloroquine in combination with methotrexate, short-term corticosteroids, and long-term corticosteroids4. However, the use of more traditional pain treatment modalities such as NSAIDS and opioids were not discussed4. The pain literature for SSc is similarly limited.5,6
The opioid crisis has disproportionately impacted patients with ARD. The incidence of patients with RA who self-reported chronic opioid use has more than doubled in recent years, rising from 7% in 2002 to 17% in 20157. An estimated 10% of patients with ankylosing spondylitis (AS) chronically use opioids, and studies from 2019–2022 show that an estimated 11–35% of patients with SLE are prescribed long-term opioid therapies.8–11 In the general population, utilization of pain-related modalities has evolved in recent years. The incidence of chronic pain patients receiving long term opioid prescriptions decreased from 24% in 2010 to 10% in 2015.12 Relatedly, the utilization of multiple non-opioid treatment modalities has increased, including physical and occupational therapy, anticonvulsants, topical analgesics, and psychosocial treatment.12 However, little is known about recent changes in utilization of alternative chronic pain treatment modalities in ARD. Such knowledge will be foundational for updating chronic pain treatment guidelines in rheumatology. The objective of this study was to investigate the patterns of pain treatment utilization among a nationwide sample of privately insured adults who were newly diagnosed with six different ARD between 2007 and 2021.
We hypothesized that there was a significant change in the utilization of pain management modalities among patients with ARD in the United States from 2007 to 2021. Specifically, we hypothesized that the utilization of non-opioid pain management modalities, such as anticonvulsants and physical therapy, has increased over time in patients with ARD, potentially reflecting a shift towards multidisciplinary and conservative approaches to pain management.
Methods
Participants
We accessed the Merative™ Marketscan® Research Database (previously IBM Truven Marketscan Research Database) through the Stanford Population Health Services Portal. The database includes the de-identified inpatient and outpatient health encounters (including diagnostic and procedural records) among patients with employment-sponsored health insurance claims in the United States.13 This database has been used extensively in rheumatology research.14 The treatments recorded in the Merative™ Marketscan® Research Database are associated with healthcare claims. Healthcare providers submit claims for services they provide to patients, and these claims include details about the services rendered, associated costs, and reimbursement requests. However, the Merative™ Marketscan® Research Database does not provide data on cost to patients (e.g. copays).
Study design
We identified patients in the outpatient setting with International Classification of Diseases 9th and 10th editions (ICD-9 and ICD-10) diagnosis codes for these ARD: AS, PsA, RA, SjS, SLE, or SSc. Patients were included if they had at least one ICD-9/10 code by a specialist (rheumatologist, dermatologist, or nephrologist). Case definitions for each ARD are included in the Appendix (Appendix pg 4).
Patients were excluded if: 1) they did not have ≥2 ICD 9/10 codes for ARD separated by 7 to 365 days; 2) if drug benefits were not identified or were missing in the patient records; 3) they did not have one year of continuous enrollment prior to the first diagnosis code; 4) they did not have one year of continuous enrollment after the first diagnosis code; and 5) if they were aged <18 years at diagnosis. We found minimal occurrences of multiple overlapping conditions, and included only the initial recorded diagnosis for each patient. As a sensitivity analysis, we repeated the baseline and trends analysis, using records of patients who received only a single ARD diagnosis.
We included osteoarthritis and hypertension as non-ARD comparators due to their prevalence in the general population and distinct etiologies and treatment approaches, providing a valuable framework for identifying disease-specific trends and differences in the utilization of pain management strategies within the broader spectrum of conditions. Patients with osteoarthritis and hypertension were included if: 1) they had ≥2 ICD 9/10 codes (See Appendix pg 5–8) separated by 7 to 365 days; 2) drug benefits were identified; 3) they had one year of continuous enrollment prior to the first diagnosis code15; 4) they had one year of continuous enrollment after the first diagnosis code; 5) they were aged ≥18 years at diagnosis; and 6) they did not have hypertension (in the case of patients with osteoarthritis) or osteoarthritis (in the case of patients with hypertension) or ARD (in both cases).
Outcomes
The primary outcome was opioid use, and the secondary outcomes included use of anticonvulsants, antidepressants, skeletal muscle relaxants, non-steroidal anti-inflammatory drugs (NSAIDs), topical pain medications, and physical/occupational therapy. The case definitions for the pain treatment modalities are included in the Appendix (See Appendix pg 9).
We extracted demographic variables and compared them across cohorts. Additionally, we documented the outpatient visit count and assessed the Charlson comorbidity index at baseline, using ICD 9/10 codes. We presented descriptive statistics as frequencies and percentages for categorical variables, and means with standard deviation (SD) for continuous variables. For each year, we extracted the incidence of patients who met the outcomes. The numerator included patients receiving the medications or non-pharmaceutical therapy within one year of their first ARD diagnosis. The denominator included patients with their first ARD diagnosis. Incident cases of ARD were defined as those who had one year of continuous enrollment prior to the first diagnosis code.
Statistical analysis
We presented both absolute and relative changes in the incidence of pain management modalities. Absolute differences offer a straightforward understanding of the actual numerical changes over time, providing a tangible measure of the shifts in incidence. Relative differences provide context by expressing changes relative to a baseline, allowing for a deeper understanding of the significance of these changes. For absolute differences, we compared the incidence between years. For relative differences, we used logistic regression to estimate the association between the trend, i.e. calendar time and outcomes, adjusted for age, sex, and region. We tested the linearity of the relationship on a logit scale between calendar year and outcomes for each model using %RCS_REG SAS macro.16 We used restricted cubic spline transformations flexibly to illustrate relationships between calendar year and outcomes. To interpret results numerically, we fit the piecewise linear spline by calendar year. We chose a spline knot, which balanced model fit by maximizing model likelihood and interpretation of results. The selection of the spline knot in 2014 was determined through visual examination of the adjusted odds ratio graph (See Appendix pg 3), and was further supported by the Drug Enforcement Agency’s decision to reschedule hydrocodone products in that same year.17 Based on this, we reported results for annual trends up to 2014 and after 2014.
We calculated odds ratios (ORs) with 95% CIs for annual change in patients with ARD. We tested the interaction between trends and sex to assess whether the relationship between trends and odds of outcomes were similar by sex. We presented trends in each cohort (i.e. AS, PsA, RA, SjS, SLE, SSc) as exploratory analyses. All the statistical analyses were done using SAS 9.4 (SAS Institute) and R 4.1.1 software (R Project for Statistical Computing).
Results
There were 141,962 patients in our study (AS [n=10,927], PsA [n=21,438], RA [n=71,393], SjS [n=16,718], SLE [n=18,018], and SSc [n=3,468]) (Figure 1). The average age at baseline was 48.2 years and varied between 44.2 years in SLE to 50.2 years in SSc. Approximately 76% (107,475/141,962) were female, ranging from 56% (6,099/10,927) female for AS to 93% (15,523/16,718) female for SjS (Table 1). Regarding the absolute changes in pain management modalities, ~26% (2610/10098) were given opioid prescriptions within the initial year of their diagnosis (Figure 2). This incidence increased to 38% (4325/11373) in 2014 and reduced to 24% (1598/6705) in 2020 (Figure 2). The incidence of physical therapy use increased from 12% (1212/10098) in 2008 to 24% (1598/6705) in 2020. The incidence of anticonvulsant use was 11% (1120/10098) in 2008 and increased to 24% (1578/6705) in 2020.
Figure 1.
Consort Diagram
Table 1.
Baseline demographic for autoimmune patients
| Total | AS | PsA | RA | SjS | SLE | SSc | |
|---|---|---|---|---|---|---|---|
| n | 141962 | 10927 | 21438 | 71393 | 16718 | 18018 | 3468 |
| Age, mean (SD) | 48.2 (10.9) | 44.3 (12.1) | 47.9 (10.4) | 49.5 (10.3) | 49.3 (10.1) | 44.2 (12.0) | 50.2 (9.8) |
| Sex (n, %) | |||||||
| Female | 107475 (75.7) | 6099 (55.8) | 12216 (57.0) | 54415 (76.2) | 15523 (92.9) | 16213 (90.0) | 3009 (86.8) |
| Male | 34487 (24.3) | 4828 (44.2) | 9222 (43.0) | 16978 (23.8) | 1195 (7.1) | 1805 (10.0) | 459 (13.2) |
| Charlson comorbidity index, mean (SD)a | 1.3 (1.2) | 0.6 (1.1) | 0.7 (1.1) | 1.6 (1.1) | 1.4 (1.2) | 1.6 (1.2) | 1.7 (1.3) |
| Number of annual outpatient visits, mean (SD)b | 10.0 (6.8) | 10.0 (7.1) | 9.2 (6.3) | 9.7 (6.7) | 11.2 (7.5) | 10.4 (6.9) | 10.4 (6.7) |
| Region (n, %) | |||||||
| Northeast | 26692 (18.8) | 2005 (18.4) | 4345 (20.3) | 12992 (18.2) | 3169 (19.0) | 3414 (19.0) | 767 (22.1) |
| North Central | 26077 (18.4) | 1863 (17.1) | 4170 (19.5) | 13392 (18.8) | 2931 (17.5) | 3048 (16.9) | 673 (19.4) |
| South | 67191 (47.3) | 4871 (44.6) | 9743 (45.5) | 34521 (48.4) | 7957 (47.6) | 8649 (48.0) | 1450 (41.8) |
| West | 20609 (14.5) | 2075 (19.0) | 2978 (13.9) | 9787 (13.7) | 2497 (14.9) | 2732 (15.2) | 540 (15.6) |
| Unknown | 1393 (1.0) | 113 (1.0) | 202 (0.9) | 701 (1.0) | 164 (1.0) | 175 (1.0) | 38 (1.1) |
Data are mean with standard deviation (SD) for continuous variables and n (%) for categorical variables.
AS: Ankylosing spondylitis; PsA: Psoriatic arthritis; RA: Rheumatoid arthritis; SjS: Sjogrens syndrome; SLE: Systemic lupus erythematosus; OA: Osteoarthritis; HTN: Hypertension; Autoimmune rheumatic diseases
.Charlson Comorbidity index was defined using outpatient claims 1 year prior to the diagnosis of ARD.
. Number of annual outpatient visits were calculated using outpatient claims 1 year before and 1 year after the diagnosis of ARD.
Figure 2.
Trends of pain-related treatments among overall autoimmune patients, and stratified by sex.
We also observed significant relative differences in the incidence of pain management modalities during the period under study. Overall, the odds of opioid use increased annually until 2014 by 4% [aOR, 1.04, 95% CI: (1.03, 1.04)] and decreased annually by 15% after 2014 [aOR, 0.85, 95% CI: (0.84, 0.86)] (Table 2). The odds of anticonvulsant use increased annually by 7% until 2014 [aOR, 1.07, 95% CI: (1.07, 1.08)] and did not significantly change after 2014 [aOR, 1.00, 95% CI: (0.99, 1.00)]. The odds of physical therapy use increased annually by 5% until 2014 [aOR, 1.05, 95% CI: (1.04, 1.06)], with a slight decrease annually by 1% after 2014 [aOR, 0.99, 95% CI: (0.98, 1.00)]. The odds of topical analgesics had a yearly increase of 9% until 2014 [aOR, 1.09, 95% CI: (1.07, 1.10)]. Prior to 2014, the odds of NSAIDs use increased by 2% annually [aOR, 1.02, 95% CI: (1.02, 1.03)], however, after 2014, the odds of NSAIDs use decreased annually by 4% [aOR: 0.96, 95% CI: (0.95, 0.96)]. These trends did not differ by sex except for NSAID use before 2014 (interaction p=0.02) and topical analgesic use after 2014 (interaction p<0.01) (Figure 2).
Table 2.
Time trends of pain-related treatments among autoimmune patients
| Outcomes | Unadjusted OR (95% CI)* | Adjusted OR(95% CI)** |
|---|---|---|
| Opioid use up to 2014 | 1.03 (1.03, 1.04) | 1.04 (1.03, 1.04) |
| Opioid use after 2014 | 0.86 (0.85, 0.86) | 0.85 (0.84, 0.86) |
| Anticonvulsant up to 2014 | 1.07 (1.06, 1.08) | 1.07 (1.07, 1.08) |
| Anticonvulsant after 2014 | 1.00 (0.99, 1.01) | 1.00 (0.99, 1.00) |
| Antidepressant up to 2014 | 1.05 (1.04, 1.05) | 1.05 (1.04, 1.06) |
| Antidepressant after 2014 | 0.98 (0.97, 0.99) | 0.98 (0.97, 0.99) |
| Skeletal muscle relaxants up to 2014 | 1.04 (1.03, 1.05) | 1.04 (1.03, 1.05) |
| Skeletal muscle relaxants after 2014 | 0.97 (0.96, 0.98) | 0.96 (0.96, 0.97) |
| NSAIDS up to 2014 | 1.02 (1.01, 1.03) | 1.02 (1.02, 1.03) |
| NSAIDS after 2014 | 0.96 (0.95, 0.97) | 0.95 (0.95, 0.96) |
| Topical analgesics up to 2014 | 1.09 (1.08, 1.10) | 1.08 (1.07, 1.10) |
| Topical analgesics after 2014 | 0.98 (0.97, 0.99) | 0.99 (0.97, 1.00) |
| Physical therapy up to 2014 | 1.05 (1.05, 1.06) | 1.05 (1.04, 1.06) |
| Physical therapy after 2014 | 0.98 (0.98, 0.99) | 0.99 (0.98, 1.00) |
Per one-year increase
Per one-year increase. adjusted for age, gender, region, and Charlson comorbidity index
There were some differences in the use of pain management modalities among individual ARD. In AS, PsA, and RA, the incidence of NSAIDs prescribing was highest, followed by opioids, while the incidence of antidepressants, anticonvulsants and topical analgesics prescribing were lowest (Figure 3). The incidence of opioid and NSAID prescribing was lower for patients with SjS and SSc than for patients with AS, PsA, and RA, however in patients with SSc, opioid use had the highest incidence compared with other pain management modalities, ranging between 20–40%.
Figure 3.
Trends of pain-related treatments among autoimmune patients, stratified by disease status
Patients with AS experienced the greatest decline in opioid use (17% reduced odds annually) after 2014, compared with patients with SSc, who experienced a 10% annual decrease in odds after 2014 (Supplemental Table 5). Patients with RA experienced the highest increase in the odds of anticonvulsant use before 2014 (10% annually), while there was no change in anticonvulsant use in patients with SSc during that period. Although the odds of antidepressant use decreased annually for most conditions after 2014, these odds increased annually by 5% for patients with SSc, although it was not statistically significant. There were no differences in the use of skeletal muscle relaxants, NSAIDs, topical analgesics, or physical therapy when stratified by condition.
We identified 639,850 patients with osteoarthritis and 3,277,292 patients with hypertension. A higher proportion of patients with ARD and osteoarthritis were prescribed pain management modalities compared with patients with hypertension (See Appendix pg 2). In addition, the incidence of opioid prescription decreased since 2014 across all diseases under study.
As a sensitivity analysis, we identified 132,702 patients (93.5%) with only one ARD diagnosis. After the exclusion of patients with multiple ARD diagnoses, the findings did not meaningfully change, demonstrating the robustness of the results (See Appendix pg 12–13).
Discussion
In this study of annual trends in pain management modalities in patients with newly diagnosed ARD in the United States, we found that the use of various non-opioid pain management modalities increased or plateaued in recent years, while opioid and NSAID use has declined since 2014. The inflection point that occurred in 2014 may reflect changes in policy by the Drug Enforcement Administration, in which hydrocodone combination products were rescheduled from schedule III to schedule I, and the 2016 guidelines by the Centers for Disease Control and Prevention which were updated to recommend nonopioid therapy as the preferred modality for the treatment of chronic pain.18,19
We found that the odds of opioid use have decreased annually by 15% in recent years. This reduction is congruent with trends found in patients with chronic pain not specific to ARD. For example, in a study of >2 million veterans with incident chronic pain, Frank et al found that long-term opioid therapy reduced annually by 14% between 2010–201612. In another study, Chen et al evaluated the use of long-term prescription opioids among patients with ARD (i.e, RA, SLE, PsA and AS) compared with patients with hypertension from 2003–201420. They found that 41–52% of patients with ARD received long-term opioid prescriptions. They also found that AS was associated with the highest relative risk of receiving long-term opioid prescriptions compared with patients with hypertension. Similar to the current study, Chen et al found an increase in opioid prescriptions in all ARD cohorts between 2003 and 201420. In contrast, using nationally representative data from physician offices and hospital outpatient departments, Taylor et al found that opioid prescribing during visits among patients with PsA in the United States significantly increased when comparing 2006–2011 (12%) to 2012–2016 (41%)21. There are several reasons why these studies may have reached different conclusions, including differences in case definitions of the population under study (e.g., incidence, prevalence, etc.), differing case mix (e.g., veterans, ambulatory visits, insured patients), statistical methods, and period of study.
There is sparse literature examining trends in opioid prescribing and non-opioid pain management modalities in SjS and SSc, despite reports of a disproportionately high burden of concomitant pain conditions in patients with these diseases. In a recent study of ~6000 patients with ARD attending rheumatology clinics, the prevalence of chronic overlapping pain conditions such as migraine, chronic low back pain, and fibromyalgia was highest in women with SjS (63%), compared with 54% in PsA, 57% in RA, and 38% in SSc1. Low back pain was found to be the most frequent pain manifestation in SSc and was associated with higher levels of pain chronification and psychological problems, underscoring the importance of managing non-disease specific symptoms such as low back pain to improve the quality of life in patients with ARD22. A recent review of publications evaluating the use of opioids in patients with inflammatory rheumatic disease found multiple studies examining the efficacy, tolerability, and side effects of oral opioids in patients with AS, PsA, RA, and SLE, but it did not find relevant studies of patients with SSc23. The review identified only two studies evaluating the efficacy of oxycodone and its derivatives in patients with SSc digital ulcers, and it did not include SjS. However, pain manifestations in SSc are myriad, including consequences of Raynaud’s phenomenon, tendon sheath involvement, arthritis/arthralgias, and fibromyalgia, highlighting the need for effective pain management strategies in this condition.24
We observed a preponderance of nonopioid pain management modalities in ARD included during the study period--a finding consistent with the extant literature. While evidence for the efficacy of some drug therapies (e.g., anti-convulsants) for pain management in ARD may not be established, we chose these six modalities due to evidence for their use in pain management in the general population.12 Using data from the National Ambulatory Medical Care Survey from 2005–2016, Soprano et al found that office visits with prescriptions of muscle relaxants more than doubled from 15 million visits in 2005 to 31 million in 201625. While they found that the use of newly prescribed skeletal muscle relaxants remained stable, the use of muscle relaxants as continued therapy tripled during the study period. In the current study, we found that the odds of muscle relaxant prescriptions increased by 4% annually until 2014, when the odds decreased annually by 4%. In addition, recent trends indicate that self-reported use of gabapentin, an anticonvulsant, tripled between 2002 to 2015 among participants in the Medical Expenditure Panel, a finding that is in agreement with claims data and the current study.26 Our study includes contemporary trends (2015 to 2021) indicating that the prescribing pattern for anticonvulsants may have stabilized in recent years. These findings are reassuring due to recent concerns about the prevalence of gabapentin in drug overdose postmortem toxicology testing results in the United States.26 However, neuropathic pain remains a challenging and underestimated manifestation in ARD27. Additionally, in the general population, trends indicate a rise in the use of physical therapy and topical medications in earlier years, however, we found in the current study that the use of these pain management modalities have stabilized in recent years.12
This study has several limitations. First, we did not include the duration of the pain treatment modalities, and as such, we were unable to differentiate between acute and chronic use. Second, we did not examine individual drugs within each pain treatment modality class to determine whether the trends are consistent across all drugs; in particular, those with increasing notoriety (e.g., gabapentin among anticonvulsants) were not individually examined.28 Third, this study was limited to prescriptions and claims from patients with newly diagnosed ARD, and thus these trends may not generalize to prevalent cases. Certainly, opioids and other pain management modalities could serve as temporary relief medications during the initial diagnosis phase and before disease-modifying antirheumatic drugs take full effect.29 However, limiting the analysis to the first year after diagnosis minimized changes in the composition of the patient population over time that can occur due to shifts in insurance coverage and changes in healthcare policies, and it provided a standardized time frame for assessment, minimizing bias associated with variations in disease duration and progression. Future studies could consider conducting a comparative analysis at two time points: initially at diagnosis and again after six months. This would allow for an assessment of changes over time. Fourth, administrative claims data primarily include individuals who are employed and insured, which is likely to under-represent disabled, unemployed, and uninsured individuals – characteristics that are particularly salient in pain. Fifth, definitions based on diagnosis codes are not 100% sensitive or specific. It is possible that patients may have been diagnosed previously but not had a diagnosis code in the one year look back period, which would result in misclassification. There have been no changes to the ICD-10 codes corresponding to ARD throughout the duration of our study. Therefore, we anticipate minimal issues regarding misclassification. Sixth, administrative claims datasets do not capture over-the-counter medications such as NSAIDs and topical analgesics, which may underestimate the overall utilization of these pain management modalities. Seventh, we lacked detailed information on the numbers of prescriptions, dosing variations, and the continuation of treatment for pain management modalities. This limitation restricts the ability to assess the intensity, frequency, and long-term adherence to prescribed pain medications, potentially providing an incomplete picture of the overall treatment landscape. Eighth, we did not examine trends in cannabis prescriptions, which have been hypothesized as opioid-sparing alternatives. This may hinder a comprehensive understanding of the nuanced choices individuals make in seeking alternative therapies, particularly in the context of the opioid epidemic and the growing interest in non-traditional pain management approaches. However, the evidence is mixed as some research indicates the proportion of patients receiving opioid prescriptions did not change after the implementation of state cannabis laws, while other research indicates reduction in opioid dosages among patients receiving medical cannabis.30 Ninth, this study focuses on individual pain management modalities and not combinations of these treatments. Combinations of medications are often employed in clinical settings to optimize therapeutic outcomes while minimizing side effects associated with high doses of a single medication. The restriction to individual modalities potentially limits the generalizability of findings to real-world patient scenarios. Future research that includes an exploration of combination therapies would provide a more comprehensive understanding of the evolving landscape of pain management and better inform clinical decision-making. Tenth, almost all large healthcare databases in the United States are administrative claims data and are subject to biases. Further investigations using complementary databases such as Medicaid, Medicare, and the Veterans Health Administration are necessary to validate these results. Eleventh, the analysis did not include race or ethnicity due to limitations in the dataset. This is an important consideration due to known heterogeneity in the burden of pain and pain outcomes across different sociodemographic profiles. Lastly, administrative claims data do not include information on disease severity and activity. ARD often exhibit varying degrees of severity, and the effectiveness of pain management modalities can be influenced by the ARD disease activity. This potentially overlooks important factors that could influence the choice and outcomes of pain management strategies in ARD.
Researchers and clinicians should interpret the findings with caution, recognizing that the scope of this study was confined to the available administrative data, which may not comprehensively reflect the nuances of managing ARD.
This study has several strengths. To our knowledge, this is the first evaluation of multiple pain treatment modalities using records of >140,000 newly diagnosed ARD patients. In addition, this study investigated pain management modalities other than just opioid and NSAID prescriptions, which is what prior work has primarily focused on in patients with ARD. We also included 15-year trends, including recent years, which allowed us to evaluate contemporary trends in pain management modalities. Additionally, we included osteoarthritis and hypertension as non-ARD comparators and observed analogous trends in opioid prescription patterns with ARD. However, patients with hypertension received opioid medications at lower rates—an observation that aligns with expectations as we hypothesized a lower prevalence of chronic pain in patients with hypertension compared with osteoarthritis or ARD. Lastly, administrative claims datasets, such as the one leveraged in this study, provide a solution to the challenge of studying rare diseases such as SLE and SSc by aggregating clinical data from demographically diverse settings.
Considering the findings from this study, we present a few recommendations. First, updated clinical guidelines for the management of chronic pain in rheumatology are greatly needed, with particular emphasis on less prevalent conditions such as SjS and SSc. These guidelines should include evidence-based approaches for the diagnosis, screening, and treatment of chronic pain and best practices for classifying pain conditions for individualized and personalized medicine, including non-pharmacological treatments and psychological support. Second, future studies are needed to evaluate the effectiveness of the changes we observed over time in pain management modalities, and to understand the possible effects they have had on outcomes such as quality of life, disability, health status, and function. Third, future research is needed to determine the best ways to manage and treat pain, considering the evolving landscape of immunosuppressive agents for treating ARD, including risks as well as potential short- and long-term benefits.
In conclusion, our administrative claims study highlights a shift towards increased utilization of non-opioid pain management modalities and a decline in opioid and NSAID prescriptions among patients with newly diagnosed ARD in the United States, reflecting changes possibly influenced by policy shifts and updated clinical guidelines. However, further research is warranted to assess the long-term impact of these trends on patient outcomes and to address gaps in pain management strategies for less prevalent conditions like SjS and SSc.
Supplementary Material
Research in context.
Evidence before this study
On August 4, 2023, we searched Pubmed and Google Scholar for articles published in English from January 1, 2010 to July 31, 2023. We used different combinations of the following search terms: pain, opioids, anticonvulsants, physical therapy, muscle relaxants, NSAIDS, antidepressants, topical analgesics, ankylosing spondylitis (AS), psoriatic arthritis (PsA), rheumatoid arthritis (RA), and systemic lupus erythematosus (SLE). Much of the recent literature focused on opioid use and misuse, given the disproportionate impact this has had on patients with ARD. We identified multiple studies examining the efficacy, tolerability, and side effects of oral opioids in patients with AS, PsA, RA, and SLE. However, few relevant studies included patients with less prevalent ARD such as systemic sclerosis (SSc) and Sjogren’s syndrome (SjS). The incidence of patients with RA who self-reported chronic opioid use more than doubled in recent years, rising from 7% in 2002 to 17% in 2015. An estimated 10% of patients with AS chronically used opioids, and studies from 2019–2022 showed an estimated 11–35% of patients with SLE were prescribed long-term opioid therapies. Factors associated with chronic opioid use included longer disease duration, higher disease activity, increased functional impairment, and higher level of disability. Our search yielded sparse literature examining trends in nonopioid pain management modalities in ARD. Clinical guidelines for the management of pain in ARD were dated and did not reflect contemporary trends in pain management, such as the shift away from chronic opioid use and changes in utilization of nonopioid pain management modalities in ARD, such as physical and occupational therapy, anticonvulsants, topical analgesics, and psychosocial treatment.
Added value of this study
The main objective of this study was to assess the use of pain management modalities among recently diagnosed patients with ARD over time and to provide updated trends using current data. We found that since 2014, the use of nonopioid pain management modalities has increased or stabilized, while opioid and NSAID use has declined. To our knowledge, this is the first evaluation of multiple pain treatment modalities using records of >140,000 newly diagnosed patients with ARD and included 15-year trends through the year 2021. This study also investigated pain management modalities other than opioids and NSAIDs, which have been the focus of prior work in patients with ARD. Additionally, the use of a large administrative claims dataset enabled us to study rare diseases, such as SLE and SSc, by aggregating clinical data from demographically diverse settings. This work sets the stage for future studies, which are needed to evaluate the effectiveness of the changes in pain treatment modalities we observed. The inclusion of osteoarthritis and hypertension as comparators offers a broader perspective on pain management trends, facilitating a better understanding of disease-specific patterns.
Implications of all the available evidence
Updated clinical guidelines for the management of chronic pain in rheumatology are greatly needed, with particular emphasis on less prevalent conditions such as SjS and SSc. These guidelines should include evidence-based approaches for the diagnosis, screening, and treatment of chronic pain. They should also include methods for classifying pain conditions to allow for individualized and personalized medicine, including non-pharmacological treatments and psychological support. Our work, along with the published literature, highlights the need for future studies to evaluate the effectiveness of pain management modality changes over time, and to understand the possible effects that changes have had on outcomes such as quality of life, disability, health status, and function.
Funding
National Institute of Arthritis and Musculoskeletal and Skin Diseases
Footnotes
All authors had full access to all the data in the study. All authors had final responsibility for the decision to submit for publication. At least two identified authors directly accessed and verified the underlying data reported in the manuscript.
Declaration of interest
All authors declare no conflict of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Data sharing statement
The data used in this study is derived from the Merative MarketScan database. Access to the dataset was obtained under appropriate licensing agreements, and the data sharing statement adheres to the terms and conditions outlined by the data provider. Due to confidentiality and privacy concerns, the raw data cannot be shared, but aggregated results and statistical summaries derived from the analysis will be made available in publications or upon request, ensuring compliance with data usage policies and regulations. Researchers interested in accessing specific details or collaborating on related projects are encouraged to contact the corresponding author for further discussion and potential collaboration within the bounds of data sharing agreements.
References
- 1.Falasinnu T, Nguyen T, Jiang TE, et al. The Problem of Pain in Rheumatology: Clinical Profiles Associated With Concomitant Diagnoses With Chronic Overlapping Pain Conditions. ACR Open Rheumatol 2022; 4: 890–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Borenstein DG, Hassett AL, Pisetsky D. Pain management in rheumatology research, training, and practice. Clin Exp Rheumatol 2017; 35 Suppl 107: 2–7. [PubMed] [Google Scholar]
- 3.American College of Rheumatology Pain Management Task Force. Report of the American College of Rheumatology Pain Management Task Force. Arthritis Care Res 2010; 62: 590–9. [DOI] [PubMed] [Google Scholar]
- 4.Carsons SE, Vivino FB, Parke A, et al. Treatment Guidelines for Rheumatologic Manifestations of Sjögren’s Syndrome: Use of Biologic Agents, Management of Fatigue, and Inflammatory Musculoskeletal Pain. Arthritis Care Res 2017; 69: 517–27. [DOI] [PubMed] [Google Scholar]
- 5.Guseva NG, Alekperov RT, Anikina NV. [Tramadol in the vascular pains of patients with systemic scleroderma and other rheumatic diseases]. Ter Arkh 1994; 66: 20–4. [PubMed] [Google Scholar]
- 6.Giuggioli D, Manfredi A, Colaci M, Ferri C. Oxycodone in the long-term treatment of chronic pain related to scleroderma skin ulcers. Pain Med 2010; 11: 1500–3. [DOI] [PubMed] [Google Scholar]
- 7.Lee YC, Kremer J, Guan H, Greenberg J, Solomon DH. Chronic Opioid Use in Rheumatoid Arthritis: Prevalence and Predictors. Arthritis Rheumatol 2019; 71: 670–7. [DOI] [PubMed] [Google Scholar]
- 8.Dau JD, Lee M, Ward MM, et al. Opioid Analgesic Use in Patients with Ankylosing Spondylitis: An Analysis of the Prospective Study of Outcomes in an Ankylosing Spondylitis Cohort. J Rheumatol 2018; 45: 188–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Birt JA, Wu J, Griffing K, et al. Corticosteroid dosing and opioid use are high in patients with SLE and remain elevated after belimumab initiation: a retrospective claims database analysis. Lupus Sci Med 2020; 7. DOI: 10.1136/lupus-2020-000435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Somers EC, Lee J, Hassett AL, et al. Prescription Opioid Use in Patients With and Without Systemic Lupus Erythematosus - Michigan Lupus Epidemiology and Surveillance Program, 2014–2015. MMWR Morb Mortal Wkly Rep 2019; 68: 819–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Figueroa-Parra G, Jeffery MM, Dabit JY, et al. Long-Term Opioid Therapy Among Patients With Systemic Lupus Erythematosus in the Community: A Lupus Midwest Network (LUMEN) Study. J Rheumatol 2023; 50: 504–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Frank JW, Carey E, Nolan C, et al. Increased Nonopioid Chronic Pain Treatment in the Veterans Health Administration, 2010–2016. Pain Med 2019; 20: 869–77. [DOI] [PubMed] [Google Scholar]
- 13.Ludwig CA, Vail D, Al-Moujahed A, et al. Epidemiology of rhegmatogenous retinal detachment in commercially insured myopes in the United States. Sci Rep 2023; 13: 9430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.England BR, Mikuls TR, Xie F, Yang S, Chen L, Curtis JR. Herpes Zoster as a Risk Factor for Incident Giant Cell Arteritis. Arthritis Rheumatol 2017; 69: 2351–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Baser O, Samayoa G, Yapar N, Baser E, Mete F. Use of Open Claims vs Closed Claims in Health Outcomes Research. J Health Econ Outcomes Res 2023; 10: 44–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med 2010; 29: 1037–57. [DOI] [PubMed] [Google Scholar]
- 17.Seago S, Hayek A, Pruszynski J, Newman MG. Change in prescription habits after federal rescheduling of hydrocodone combination products. Proc 2016; 29: 268–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Drug Enforcement Administration, Department of Justice. Schedules of controlled substances: rescheduling of hydrocodone combination products from schedule III to schedule II. Final rule. Fed Regist 2014; 79: 49661–82. [PubMed] [Google Scholar]
- 19.Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain--United States, 2016. JAMA 2016; 315: 1624–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chen SK, Feldman CH, Brill G, Lee YC, Desai RJ, Kim SC. Use of prescription opioids among patients with rheumatic diseases compared to patients with hypertension in the USA: a retrospective cohort study. BMJ Open 2019; 9: e027495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Taylor MT, Noe MH, Horton DB, Barbieri JS. Trends in opioid prescribing at outpatient visits for psoriasis and psoriatic arthritis. Br J Dermatol 2021; 185: 200–1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Evers C, Jordan S, Maurer B, et al. Pain chronification and the important role of non-disease-specific symptoms in patients with systemic sclerosis. Arthritis Res Ther 2021; 23: 34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Anastasiou C, Yazdany J. Review of publications evaluating opioid use in patients with inflammatory rheumatic disease. Curr Opin Rheumatol 2022; 34: 95–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Di Franco M, Bazzichi L, Casale R, Sarzi-Puttini P, Atzeni F. Pain in systemic connective tissue diseases. Best Pract Res Clin Rheumatol 2015; 29: 53–62. [DOI] [PubMed] [Google Scholar]
- 25.Soprano SE, Hennessy S, Bilker WB, Leonard CE. Assessment of Physician Prescribing of Muscle Relaxants in the United States, 2005–2016. JAMA Netw Open 2020; 3: e207664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Pauly NJ, Delcher C, Slavova S, Lindahl E, Talbert J, Freeman PR. Trends in Gabapentin Prescribing in a Commercially Insured U.S. Adult Population, 2009–2016. J Manag Care Spec Pharm 2020; 26: 246–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Cuzdan N, Turk I, Sarpel T, Erken E, Alparslan ZN. Neuropathic pain: is it an underestimated symptom in systemic sclerosis? Clin Rheumatol 2018; 37: 1845–51. [DOI] [PubMed] [Google Scholar]
- 28.Tharp AM, Hobron K, Wright T. Gabapentin-related Deaths: Patterns of Abuse and Postmortem Levels. J Forensic Sci 2019; 64: 1105–11. [DOI] [PubMed] [Google Scholar]
- 29.Sood A, Kuo Y-F, Westra J, Raji MA. Disease-Modifying Antirheumatic Drug Use and Its Effect on Long-term Opioid Use in Patients With Rheumatoid Arthritis. J Clin Rheumatol 2023; 29: 262–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.McGinty EE, Tormohlen KN, Seewald NJ, et al. Effects of U.S. State Medical Cannabis Laws on Treatment of Chronic Noncancer Pain. Ann Intern Med 2023; 176: 904–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data used in this study is derived from the Merative MarketScan database. Access to the dataset was obtained under appropriate licensing agreements, and the data sharing statement adheres to the terms and conditions outlined by the data provider. Due to confidentiality and privacy concerns, the raw data cannot be shared, but aggregated results and statistical summaries derived from the analysis will be made available in publications or upon request, ensuring compliance with data usage policies and regulations. Researchers interested in accessing specific details or collaborating on related projects are encouraged to contact the corresponding author for further discussion and potential collaboration within the bounds of data sharing agreements.