Investigating Prescription Medication Expenditures and Level of Perceived Health Status among Older Adults with Pain in the United States

Background: the objective of this retrospective, cross sectional study was to assess the relationship between different levels of perceived health status and prescription medication expenditures among United States adults aged ≥50 years with self-reported pain. Methods: using 2019 Medical Expenditure Panel Survey (MEPS) data, four levels of perceived health status were formed (excellent, very good, good, and fair/poor health). Differences between characteristics in the groups were compared using chi square tests. Adjusted linear regression models, using logged positive annual prescription medication expenditures, were constructed to assess differences in prescription medication expenditures between levels of perceived health. Results: a total of 4826 individuals were included in the study. Adjusted linear regression analysis indicated those who reported excellent health had 65.8% lower annual prescription medication expenditures than those with fair/poor health. Those with very good health had 49.7% lower annual prescription medication expenditures than those with fair/poor health, while those with good health had 27.2% lower annual prescription medication expenditures than those with fair/poor health. Conclusions: better perceived health status was generally associated with relatively lower prescription medication expenditures. Further research is needed to investigate if this pattern is also observed for other categories of healthcare expenditures and in other populations.


Introduction
Pain is a complex, subjective condition that is difficult to measure due to variation among individuals [1]. The International Association for the Study of Pain has defined pain as "an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage," [2]. Pain is a common condition that has become increasingly more prevalent in recent years and is associated with older age [3,4]. Pain is associated with many common diseases in later life, including dermatology, gastrointestinal, cardiovascular, pulmonary, rheumatology, endocrine, nephrology, immune, neurology, oncology, and miscellaneous complaints [5]. Pain is also one of the most common reasons for individuals to seek medical care [6,7]. For instance, individuals aged ≥65 years accounted for 30.8% of United States (US) adults with chronic pain in 2019 [3]. Another report from 2011 stated half of older adults had bothersome pain in the past month among 7601 US adults aged ≥65 years [8]. Another report found that over 100 million US adults had chronic pain in 2010, and the national cost of pain ranged from USD 560 billion to USD 635 billion (US dollars in 2010) [9,10].
Not only is pain management burdensome from an economic perspective, but individuals with pain often use an onerous quantity of different medications to help manage pain. For example, one study found that individuals with chronic pain used 13 different management strategies on average, many of which were prescription medications [11]. These prescription medications included the following drug classes: opioids, analgesics, anticonvulsants, antidepressants, antipsychotics, barbiturates, beta blockers, calcium channel blockers, muscle relaxants, non-steroidal anti-inflammatory drugs (NSAIDs), sedatives, steroids, triptans, and others [11]. Meanwhile, a recent literature review identified 16 different categories of prescription medications are used by communitydwelling individuals to help manage pain [12]. These categories included: prescription medications generally, analgesics, acetaminophen, anticonvulsants, antidepressants, anxiolytics/sedatives/antipsychotics, salicylates, beta/calcium channel blockers, diseasemodifying anti-rheumatic drugs (DMARD)/steroids, any type of injections, muscle relaxants, NSAIDs, opioids, any type of topical product, triptans, and others [12]. Prescription medications are therefore one of the most common approaches for managing pain [11][12][13].
Total US spending on prescription medications amounted to USD 507.9 billion in 2019, which was a 5.4% increase from the prior year's spending [14]. This number has been steadily rising over the last several years, for instance, there was a similar percent increase of 5.5% from 2017 to 2018 [15]. Reasons to explain this increase include new drugs, drug pricing changes, increasing numbers of patients, disease patterns, state and federal policies, and prescribing practices [15].
Although many characteristics are associated with pain in older adults, one characteristic of interest is self-reported perception of health [16]. Self-perceived health status is considered an important component to promote population health [17]. There is limited information that describes the association between perceived health status and prescription medication expenditures among older US adults with pain, yet it is possible that prescription medication expenditures may differ based on the perceived health status of older adults with pain. Therefore, the objective of this study was to assess the relationship between different levels of perceived health status and prescription medication expenditures among US adults aged ≥50 years with self-reported pain, hypothesizing that better perceived health status was associated with relatively lower prescription medication expenditures.

Methods
This was a retrospective, cross-sectional study using the 2019 Medical Expenditure Panel Survey (MEPS) data. The MEPS sampling framework is adapted from the National Health Interview Survey sampling framework, and weighting variables are provided by MEPS to produce nationally representative estimates of the non-institutionalized civilian US population during analysis [18]. MEPS has three main components, including the MEPS household component (MEPS-HC), MEPS insurance component (MEPS-IC), and MEPS medical provider component (MEPS-MPC) [18]. MEPS-HC data are collected by surveying eligible households five times over a two-year period, and include demographic and personal characteristics, health conditions, and health status, access to healthcare, healthcare service use, and healthcare expenditure data. Data from MEPS-IC and MEPS-MPC supplement MEPS-HC data to improve the validity and reliability of the data [18]. This study used the 2019 full-year consolidated data file (MEPS-HC-216), which contained data from MEPS panel 22 (interview rounds 3, 4, 5) and panel 3 (interview rounds 1, 2, 3) [19,20] and was the most up-to-date data available at the time of the study. All MEPS subjects provided verbal informed consent before data collection.
The study inclusion criteria were: MEPS subjects alive for the full 2019 calendar year; aged ≥50 years; reported pain that interfered with normal work (including both work outside the home and housework) in the past four weeks; and positive annual prescription medication healthcare expenditures.
The dependent variable was annual positive prescription medication expenditures. The independent variable was self-perceived health status, categorized as excellent, very good, good, and fair/poor. Table 1 outlines the potential confounders that served as control variables. These included age (50-64, ≥65 years), sex (male, female), ethnicity (Hispanic, non-Hispanic), race (white, other), marital status (married, other), education status (up to high school, high school, more than high school), employment status (employed, unemployed), health insurance coverage (private, public, uninsured), poverty status (based on income: poor/near poor/low, middle/high), help with activities of daily living (yes, no), help with instrumental activities of daily living (yes, no), frequent exercise (yes, no), current smoker (yes, no), number of chronic conditions (≥5, <5), pain intensity (quite a bit/extreme, little/moderate), mental health status (excellent, very good, good, fair/poor), and US census region (Northeast, Midwest, South, West) [19,20].
Chi-square tests were used to identify statistical differences between groups. Because of the non-linear nature of expenditure data, logarithmically transformed data were used in the adjusted linear regression model to assess differences in prescription medication expenditures between health status categories, with fair/poor acting as the reference group. The adjusted model included the independent variable (perceived health status) and all the control variables. Analyses accounted for the complex MEPS design. Nationally representative estimates were obtained using the relevant weighting variable, and variance estimates were calculated using the Taylor-series linearization method. An alpha level of 0.05 was chosen a priori, and all analyses were conducted using the SAS Studio statistical software (SAS Institute Inc., Cary, NC, USA).

Results
There were 28,512 available subjects in the 2019 MEPS set of data, with 4826 meeting the eligibility criteria to be included in the study. The prevalence of excellent perceived health status was 8.1% 95% confidence interval (CI) 7.1, 9.0). Very good health status had a prevalence of 28.9% (95% CI 27.2, 30.6), good health status was 36.3% (95% CI 34.6, 38.0), and fair/poor health status was 26.8% (95% CI 25.2, 28.3).
The findings of the adjusted linear regression models are reported as a percent difference relative to those individuals who reported having fair/poor health. As shown in Figure 1, older US adults with pain who reported excellent health had 65.8% lower annual prescription medication expenditures than those who reported having fair/poor health. Those who reported having very good health had 49.7% lower annual prescription medication expenditures than those with fair/poor health, while those with good health had 27.2% lower annual prescription medication expenditures than those with fair/poor health.
Medicines 2022, 9, x FOR PEER REVIEW 6 of 10 The findings of the adjusted linear regression models are reported as a percent difference relative to those individuals who reported having fair/poor health. As shown in Figure 1, older US adults with pain who reported excellent health had 65.8% lower annual prescription medication expenditures than those who reported having fair/poor health. Those who reported having very good health had 49.7% lower annual prescription medication expenditures than those with fair/poor health, while those with good health had 27.2% lower annual prescription medication expenditures than those with fair/poor health.

Discussion
This study examined the association between self-perceived health status in older US adults with pain and prescription medication expenditures. The primary finding from this study was that the better older US adults with pain perceived their health status, the lower their relative prescription medications costs were compared to those with poorer perceived health status. Although these findings are perhaps unsurprising, there are no contemporary studies that assess the relationship between perceived health status and prescription medication costs, thus this study adds new information to the literature. However, previous studies have explored similar topics and reported findings that correlate with those in the current study. For instance, a 2003 study found that prescription medication expenditure was higher in older adults with fair (USD 1023) or poor (USD 1302) health status, compared to those who reported excellent/very good health status (USD 509) [21].
Other variables that were controlled for in the analysis may also explain some of these results, such as self-reported pain severity and number of comorbid conditions. A previous study using MEPS data found that over half (54.5%) of US older adults who reported having extreme or quite a bit of pain had fair or poor perceived health, whereas only 17.8% of those with moderate or little pain had fair or poor perceived health [22]. Compared to those with little pain, older US adults (≥50 years of age) had 32% greater

Discussion
This study examined the association between self-perceived health status in older US adults with pain and prescription medication expenditures. The primary finding from this study was that the better older US adults with pain perceived their health status, the lower their relative prescription medications costs were compared to those with poorer perceived health status. Although these findings are perhaps unsurprising, there are no contemporary studies that assess the relationship between perceived health status and prescription medication costs, thus this study adds new information to the literature. However, previous studies have explored similar topics and reported findings that correlate with those in the current study. For instance, a 2003 study found that prescription medication expenditure was higher in older adults with fair (USD 1023) or poor (USD 1302) health status, compared to those who reported excellent/very good health status (USD 509) [21].
Other variables that were controlled for in the analysis may also explain some of these results, such as self-reported pain severity and number of comorbid conditions. A previous study using MEPS data found that over half (54.5%) of US older adults who reported having extreme or quite a bit of pain had fair or poor perceived health, whereas only 17.8% of those with moderate or little pain had fair or poor perceived health [22]. Compared to those with little pain, older US adults (≥50 years of age) had 32% greater prescription medication costs if they reported having extreme pain, 35% greater costs if they reported quite a bit of pain, and 29% greater costs if they reported having moderate pain [23].
The presence of multiple comorbid conditions may also have an impact on prescription medications costs, considering that the current study found that of those with fair/poor health status, 40.8% had ≥5 chronic conditions. Several other factors in the MEPS dataset are known to be associated with having multiple (≥5) chronic conditions including age, gender, ethnicity, race, employment status, functional limitations, work limitations, pain severity, and perceived health status [24]. Another study that also used MEPS data demonstrated that prescription medication expenditures were over 100% greater among older US adults with pain and multimorbidity (≥2 chronic conditions), versus those with no multimorbidity (<2 chronic conditions) [25]. The same study also identified that prescription medication expenditures had the greatest costs compared to several other categories of costs examined (including hospital inpatient, office-based, home health care, and total healthcare costs) [25], which emphasizes the importance of considering prescription medication costs relative to other costs.
The findings of this study may have implications for the management of pain among older US adults. However, it is important to note that these findings should not discourage appropriate prescription medication use. Prescription medications are typically prescribed to improve the health of an individual, either to help prevent a condition from developing or worsening, or to treat an existing condition. Traditionally, pain has been managed through pharmacological treatment, i.e., analgesics [13]. For example, the World Health Organization's analgesic ladder is a standard model in pain management therapy consisting of three steps, starting with non-opioid analgesics for mild pain, up to potent opioids for severe pain [26].
However, opioids are sometimes not a desirable option given that they may be inappropriately prescribed or withdrawn abruptly as healthcare providers try to address the opioid epidemic in the US [27]. There is also a cost associated with opioid use; a recent MEPS study reported that older US adults (≥50 years) with pain who used at least one opioid in 2015 had 63% greater prescription medication costs that those who did not use opioids [28].
There may be opportunities to optimize prescribing that could reduce prescription medication costs or consider non-pharmacological options, as appropriate. Literature reviews have identified several non-pharmacological therapies that exist for pain management. For example, one systematic review found that strategies such as multidisciplinary rehabilitation, massage, acupuncture, and yoga have been reported among adults with low back pain [29].
Another systematic review of 18 studies among community dwelling adults with pain reported the following non-pharmacological pain management strategies: consulted medical practitioner, chiropractor, surgery, activity modification or restriction, acupuncture, altering the body potion or posture, using an assistive device, exercise, hot/cold modalities, massage, physical therapy, transcutaneous electrical nerve stimulation (TENS), prayer or medication, relaxation, rest or sleep, therapy, complementary and alternative medicine, dietary and herbal supplements, diet modification, and others [12]. Of these, exercise and massage were among the most commonly used [12].
Studies have investigated the association between frequent exercise and healthcare expenditures. For example, a recent MEPS study found that US adults aged ≥50 years who did at least 30 min of moderate-vigorous intensity physical activity at least five times a week had 15% lower annual prescription medication expenditures in 2018 compared to those who did not meet this level of exercise [30].
With regards to massage, a recent randomized controlled trial of patients with low back pain received deep tissue massage or deep tissue massage with NSAIDs. The trial found that deep tissue massage had a positive effect on reducing pain among patients with chronic low back pain [31].
Non-pharmacological therapies have shown moderate effectiveness for chronic low back pain at the short and intermediate term, but most of their effects independently were small and with little long-term evidence [32]. Though additional investigation is needed, several other forms of non-pharmacological treatment may be more effective as a potential adjunct to prescription drugs [33]. Non-pharmacological treatment may not act as a replacement for prescription medications, but rather serve to complement medications to help manage pain and reduce reliance on medications [34,35].
Previous research as investigated the association of various interventions on health status. For instance, one study of older women in Spain who received motivational aquatic resistance training found that several factors, including greater satisfaction, self-selection, volition, and autonomy, were associated with greater well-being, adherence, and health outcomes [36]. Another study found traditional Chinese medicine, such as acupuncture, herbal medicine, and dietary therapy was perceived to improve health and reduce the stress and side effects of treatments among cancer survivors in Australia [37].
Advantages of this study included the large nationally representative sample of US adults with pain, and the availability of data for many potentially confounding variables that could be controlled for in analyses.
Limitations of the study included the self-reported nature of the MEPS data, which could lead to biases. Variables in the study were a combination of objective measures and subjective measures; although the subjective variables allowed for the inclusion of patientreported data, they may also have led to biases. To be included in the study, individuals needed to have positive annual prescription medication expenditure, thus older individuals that had pain but did not have positive prescription medication expenditures were not captured in the results. The study also did not include institutionalized or non-civilian individuals; thus, the findings cannot be generalized to that population.
Greater effort is therefore needed to help improve the perceived health status of older US adults with pain to help reduce their need for prescription medications, and further research is needed to assess strategies that may help improve pain management without adversely affecting health outcomes.

Conclusions
The adjusted linear regression analyses in this retrospective, cross-sectional study found that relative to fair/poor perceived health status, excellent perceived health status was associated with lowest prescription medication expenditures, followed by very good and good perceived health status. The findings from this study add contemporary information to the literature about the ever-increasing prescription medication costs in the US. These findings also suggest that more effort is warranted to help improve the health of older US adults with pain to help reduce their need for prescription medications, and there is a need for additional research to assess strategies that may help improve pain management without adversely affecting health outcomes.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. Group, for work not related to this study. The remaining authors have no conflicts of interest to disclose. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.