Association between self-reported pain severity and characteristics of United States adults (age ≥50 years) who used opioids

1 Introduction

Pain is physiologically a tool for protection, but when its adaptive function is lost, it can develop into a pathologic state that negatively influences the quality of life [1]. The revised definition of pain by the International Association for the Study of Pain defines pain as “an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage” [2]. Mechanisms underlying pain have a neural basis; however, being aware of pain is a perception that is subjective and cannot be measured directly and objectively [3].

Findings from the National Health Institute Survey in 2019 estimated that 20.4% of United States (US) adults experienced chronic pain and 7.4% of US adults experienced high-impact pain (pain interfering with life or work activities). This study also noted that pain increases with age and peaks in prevalence among adults age ≥65 years [4]. In 2008, the national cost of pain in the US was estimated to range from $560 to $635 billion and this was more than the annual cost of diabetes ($188 billion), cancer ($243 billion), and heart disease ($309 billion) [5]. Over time, it is conceivable that the financial burden of pain will increase.

Physical and behavioral therapies, opioids, and non-opioid analgesics are some of the treatment options for managing pain [6,7,8,9]. Opioids are important for the management of severe, transient pain during acute painful events and near the end of life [10]. However, the use of opioids in pain management is debatable owing to the potential harm associated with the long-term use of opioids such as high opioid misuse rates, addiction unacceptable fatality rates, and a significant burden on the societies affected [10]. Existing studies have also reported that a high risk of falls or fractures in older adults with osteoarthritis is associated with increased use of opioids [11,12]. Using data from Medicare and Medicaid Services, there has been an estimated greater than 3-fold increase in opioid use disorder among older US adults between 2013 and 2018 and pain has been identified as a risk factor for problematic opioid use [13,14].

Given the challenges associated with opioid use and pain management, it is pertinent to explore and understand the characteristics of the older population with pain who use opioids. Effective pain management requires a comprehensive assessment of pain [15]. In addition to assessing pain characteristics (pain intensity, frequency, severity, and duration) in the older population, individual characteristics may also be considered for a holistic approach. In previous studies, emphasis has been on pain characteristics and their effect on the quality of life and disability in the aging populace [16,17,18]. However, these pain characteristics in older adults are influenced by several factors, and understanding the individual characteristics can point to potential predictors of pain severity, frequency, and duration. It can also provide critical insight into traits that may be harnessed to enhance pain management techniques and inform individualized approaches or strategies for managing pain. The aim of this study is to identify individual characteristics that are linked to pain severity in older US adults ≥50 years who used opioids (Figure 1).

Figure 1

Study subject eligibility flowchart.

2 Methods

The Medical Expenditure Panel Survey (MEPS) data source consists of a set of large-scale surveys of families and individuals, their medical providers (doctors, hospitals, pharmacies, etc.), and employers across the US and it is conducted by the Agency for Healthcare Research and Quality [19]. The 2019 full-year consolidated data file consists of MEPS data obtained from Panel 23 (rounds 3–5) and Panel 24 (rounds 1–3). This file contains variables pertaining to survey administration, demographics, income, person-level conditions, health status, disability days, quality of care, employment, health insurance, and person-level medical care use and expenditures [20]. The 2019 prescribed medicine file provides detailed information on household-reported prescribed medicines for a nationally representative sample of the civilian noninstitutionalized population of the US [21].

Eligible participants for this retrospective cross-sectional study were US adults who were alive for the full calendar year, aged ≥50 years, with self-reported pain within the past 4 weeks, and ≥1 opioid prescription within the calendar year. The dependent variable (pain) was assessed based on the participant’s response (of a little bit/moderately; quite a bit/extremely) to the question, “During the past 4 weeks, pain interfered with normal work outside the home and housework” [22]. Therapeutic class codes 60 (narcotic analgesics) and 191 (narcotic analgesics combination) from the MEPS 2019 prescribed medicines file were used to identify participants who had ≥1 opioid prescription in 2019 [23].

The Behavioral Model of Health Services Use was used to group the independent variables [24]. These variables included:

Predisposing variables: age (50–64, ≥65 years); gender (male, female); race (white, other); ethnicity (Hispanic, non-Hispanic).

Enabling variables: education status (high school or less, more than high school); employment status (employed, unemployed); insurance status (private, public, none); marital status (married, other); and income level (poor/near poor/low income, middle/high income).

Need variables: activities of daily living (ADL) (ADL, no ADL); instrumental activity of daily living (IADL) (IADL, no IADL); chronic conditions (<2 vs ≥2); health status (excellent/very good/good vs fair/poor perceived health); and mental health status (excellent/very good/good vs fair/poor perceived mental health).

Personal health practice variables: regular exercise (yes, no) and smoker (yes, no).

External environmental variables: region (Northeast; Midwest; South; and West).

Chi-squared tests were used to compare the demographic characteristics of subjects stratified by their self-reported pain severity (little/moderate vs quite a bit/extreme). Multivariable logistic regression models were used to assess statistically significant associations between various characteristics and quite a bit/extreme self-reported pain severity, with little/moderate self-reported pain severity as the reference group. The odds ratio and 95% confidence limits for each variable were reported. The a priori alpha level was 0.05. The Taylor series linearization method was used to compute the variance of MEPS estimates and analyses were weighted to obtain nationally representative estimates. The cluster and strata analyses were used to maintain the effect of the sampling design on the MEPS estimates of variance and sampling errors while a domain analysis was done to differentiate the eligible population from the ineligible population. Analyses were conducted using SAS (v9.4, SAS institute Inc., Cary, NC, USA).

3 Results

The 2019 MEPS dataset contained data for 28,512 persons. There were 1,077 subjects who met the eligibility criteria of the study (low/moderate pain = 556, quite a bit/extreme pain = 521). The weighted population was 12,109,702 (low/moderate pain = 54.8% [95% confidence Interval (CI) = 51.2–58.4%], quite a bit/extreme pain = 45.2% [95% CI = 41.6–48.8%]).

Participants for both stratified levels of pain (low/moderate pain, quite a bit/extreme pain) most commonly had the following characteristics: ≥65 years of age, female, white, non-Hispanic, more than high school education completed, unemployed, private insurance, married, middle/high income, had no limitations (ADL or IADL), ≥2 chronic conditions, excellent/very good/good perceived health status, excellent/very good/good perceived mental health status, no regular exercise, and non-smokers. The most common region to reside in was the South region. The following variables had statistically significant differences (p < 0.05) between the groups: sex, race, education completed, employment status, insurance status, marital status, income level, ADL limitation, IADL limitation, perceived health status, perceived mental health status, and exercise status. For further details see Table 1.

In the fully adjusted logistic regression model, higher odds of reporting quite a bit/extreme pain were associated with the following: age (years) 50–64 vs ≥ 65 (adjusted odds ratio [AOR] = 1.76; 95% CI = 1.22–2.54) and IADL limitation, yes vs no (AOR = 2.27; 95% CI = 1.30–3.96). Lower odds of reporting quite a bit/extreme pain were associated with the following: Hispanic vs non-Hispanic (AOR = 0.50; 95% CI = 0.30–0.85), employed vs unemployed (AOR = 0.50; 95% CI = 0.33–0.75); private insurance vs no health insurance (AOR = 0.15; 95% CI = 0.03–0.70), excellent/very good/good health vs fair/poor (AOR = 0.32; 95% CI = 0.23–0.46), excellent/very good/good mental health vs fair/poor (AOR = 0.46; 95% CI = 0.30–0.72), and smoker vs non-smoker (AOR = 0.56; 95% CI = 0.37–0.84). For further details see Table 2.

4 Discussion

This study identified several characteristics associated with self-reported pain severity in US adults ≥50 years who have used an opioid. Age (50–64 years vs >65 years) and ethnicity (Hispanic vs non-Hispanic) were the predisposing variables that had the greatest association with quite a bit/extreme pain severity. Previous work by Dahlhamer et al. found that compared to other age groups, individuals aged 45–64 had a greater prevalence of both chronic pain and high-impact pain [25]. A possible explanation for this finding was that older people (over 65 years) are more likely to be retired and may be able to avoid activities that could cause pain at work or may be able to handle their health and pain better because they have more time [26]. Additionally, it has been reported that non-Hispanic white adults have a substantially greater age-adjusted prevalence of persistent pain than any other racial and ethnic subgroups [25]. This may explain why the proportion of Hispanic adults in the current study was lower than the proportion in the general population. It may also be the case that Hispanic adults use the healthcare system less (either due to lack of choice, or lack of access or health insurance).

Among the enabling variables, employment and health insurance status were the characteristics significantly associated with self-reported pain severity. Individuals who were employed had lower odds of reporting quite a bit/extreme pain, compared to those who were unemployed. The presence of chronic pain may result in functional limitations and disabilities that can adversely affect an individual’s ability to secure employment. This suggestion is supported by previous findings that the risk of developing limitations in activities or participatory restrictions, such as being unable to work for a livelihood, is highly correlated with pain [27]. Individuals with private health insurance had lower odds of reporting quite a bit/extreme pain, than those who had no insurance. It is not surprising that unemployment and lack of health insurance coverage are associated with increased pain severity. Low earnings and lack of insurance further restrict the ability to get medical care and may relegate people to less costly treatments that may be less suitable [28].

Among the need variables, IADL limitation, perceived health status, and perceived mental health status showed significant association with reporting quite a bit/extreme pain. These findings are similar to those of previous studies, which found that greater levels of self-reported pain severity were associated with greater odds of reporting a functional limitation [29]. Another study also found that functional limitations that are typically linked with aging appear in subjects with severe pain much earlier in life [30]. In another study, pain was negatively associated with health-related quality of life, and physical and mental health was found to have an inverse relationship with pain intensity [31]. Several studies have also found that mental health disorders such as depression are associated with pain. For example, a study on older adults in Germany found that the factors that best predicted depression in later life were multisite pain, frequency, and severity of pain [32,33,34,35].

Smoking status was the only personal health practice variable to be associated with self-reported pain severity. Smokers had a lower likelihood of reporting quite a bit/extreme pain compared to non-smokers. The association between pain and smoking status can be influenced by various factors. For example, in a previous study, after adjusting for depression in the multivariate linear regression analyses, smoking status was not associated with baseline pain severity [36]. In contrast to our findings, when pain intensity between smokers and non-smokers was compared, it was discovered that smokers reported considerably higher pain levels than non-smokers at the time of consultation [37]. This difference in findings may be due to the complex and bi-directional relationship between smoking and pain [37,38].

Previous research using adjusted logistic regression analysis in a similar cohort of US adults aged ≥50 years with pain who used opioids has recently been conducted. In one study, those aged 60–69 years (vs ≥80 years), those who reported they were in excellent/very good/good (vs fair/poor) health, those who reported they were normal/underweight and overweight (vs obese), and those who reported they had little (vs extreme) pain had higher odds of reporting doing frequent exercise [39]. In another study, those aged 50–64 (vs ≥65 years), Hispanic (vs non-Hispanic) ethnicity, employed (vs unemployed), and frequent exercise (vs none) were associated with lower odds of multimorbidity [40]. In a further study, those with extreme (vs little) and quite a bit of (vs little) pain, males (vs females), white race (vs not white race), high school or less (vs more than high school) education, and current smoker (vs not a current smoker) were associated with lower odds of reporting good health. Being employed (vs unemployed), having <2 chronic conditions (vs ≥2), and doing regular physical activity (vs not doing so) were associated with higher odds of reporting good health [41]. In one final study, having extreme, quite a bit, and moderate (vs little) pain, being unemployed (vs employed), unmarried (vs married), having poor (vs good) overall health, and residing in the Midwest (vs West) were all associated with greater odds of having any limitation [42]. Several of the characteristics observed in these studies were also observed in the current study. These findings demonstrate that there are some characteristics that are frequently associated with outcomes of interest among the older US population with pain who used opioids while others appear to depend on the outcome being studied.

The strengths of this study include the use of a nationally representative database of the US population which consists of different racial and ethnic minorities, and this makes the results more generalizable. That said, the proportion of white people in this study cohort was higher than the general population. Further research with a sample consisting of more diverse race and ethnicities is warranted to further investigate the effect of race. This study involved adults who were opioid users and MEPS provides data on prescribed medicine use and this facilitates research on prescribed medicine utilization. Despite these strengths of using MEPS, there are also potential biases and limitations; bias may arise from the MEPS response rate, recall bias may exist because MEPS data are self-reported and since the study design is cross-sectional, causal inference cannot be made.

The various characteristics associated with self-reported pain severity in this study provide critical insight that may be useful in developing and implementing targeted and innovative individualized pain management strategies in older adults. Future research may be geared toward integrating characteristics peculiar to an individual in the process of pain evaluation and in selecting interventions to manage pain, as it may impact response to treatment positively. Additionally, future studies can explore the challenges associated with integrated and comprehensive pain management programs in older adults.