Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jan 1.
Published in final edited form as: Am J Ind Med. 2020 Nov 24;64(1):48–57. doi: 10.1002/ajim.23202

Predictors of long-term opioid use and opioid use disorder among construction workers: Analysis of claims data

Ann Marie Dale 1, Skye Buckner-Petty 1, Bradley A Evanoff 1, Brian F Gage 1
PMCID: PMC7799490  NIHMSID: NIHMS1652785  PMID: 33231876

Abstract

Background:

Construction workers have high rates of work-related musculoskeletal disorders, which lead to frequent opioid use and opioid use disorder (OUD). This paper quantified the incidence of opioid use and OUD among construction workers with and without musculoskeletal disorders.

Methods:

We conducted a retrospective study using union health claims from January 2015 through June 2018 from 19,909 construction workers. Claims for diagnoses of chronic musculoskeletal disorders, acute musculoskeletal injuries, musculoskeletal surgery, and other conditions were linked to new opioid prescriptions. We examined the effects of high doses (≥ 50 morphine mg equivalents per day), large supply (more than 7 days per fill), long-term opioid use (60 or more days supplied within a calendar quarter), and musculoskeletal disorders on the odds of a future OUD.

Results:

There were high rates (42.8% per year) of chronic musculoskeletal disorders among workers of whom 24.1% received new opioid prescriptions and 6.3% received long-term opioid prescriptions per year. Workers receiving opioids for chronic musculoskeletal disorders had the highest odds (95% CI) of future OUD: 4.71 (3.09-7.37); workers prescribed long-term opioids in any calendar quarter had a nearly 10-fold odds of developing an OUD.

Conclusions:

Among construction workers, opioids initiated for musculoskeletal pain were strongly associated with incident long-term opioid use and OUD. Musculoskeletal pain from physically demanding work is likely one driver of the opioid epidemic among occupations like construction. Prevention of work injuries and alternative pain management is needed for workers at risk for musculoskeletal injuries.

Keywords: blue collar worker, musculoskeletal disorders, opioid prescriptions, pain treatment

INTRODUCTION

Alleviation of pain is one of the most common reasons why patients visit healthcare providers. 41% of insured Americans who start an opioid for acute musculoskeletal pain receive at least 1 refill.1 In fact, musculoskeletal pain is the most frequent diagnosis leading to an opioid refill.1 Although most patients stop their opioids once their acute pain resolves, 1%-12% of patients continue opioid use beyond 6 months.2 Furthermore, this risk of long-term opioid use is greater among working-age men, especially those who have recurrent acute musculoskeletal pain and/or chronic pain.3, 4 In addition to musculoskeletal conditions, opioids are prescribed for many other conditions, including cancer, dental procedures, migraines, abdominal conditions, and kidney stones.5

Recent studies show a trade-off between pain relief and risks of OUD.3, 4 Many patients (35%) with long-term opioid therapy use disorder are subsequently diagnosed with opioid use disorder (OUD), a “problematic pattern of opioid use leading to clinically significant impairment or distress.”6, 7 There are substantial risks of adverse events from long-term opioid use: overdose, suicide, drug-craving, opioid withdrawal, and side effects (e.g. depression, hyperalgesia syndrome, and constipation).8 How much recurrent acute musculoskeletal pain or chronic musculoskeletal pain increases the risk of long-term opioid use and OUD is unclear, especially because opioids prescribed for any condition, even transient pain, can lead to long-term opioid use.9 This long-term use is even more distressing given that long-term opioids do not relieve chronic pain.10 However, many patients resist stopping their opioids, and alternative pain treatments may not be available, accessible, or affordable.11–13

Construction workers have high rates of opioid use and overdose related to opioids.12, 14 Among all occupations in the US, construction workers had the highest mortality rate from opioids.15, 16 A study of death certificates in Massachusetts from 2011-2015 found that construction workers accounted for 24% of all opioid-related deaths among the working population--a rate of fatal overdose five times higher than the rate for all workers.16 A national sample from 2011-2017, showed that 10.4% of construction workers were prescribed opioid analgesics. Construction workers have several factors that may contribute to their high rates of opioid use and misuse.14 They have among the highest rates of work injury,17 and chronic or recurrent pain related to their physically demanding work and frequent musculoskeletal injuries.18

Construction workers have higher rates of musculoskeletal disorders than the general working population19 and thus more opportunities to be prescribed opioids. These workers also possess characteristics typical of those at high risk for OUD: predominantly male, with high rates of alcohol and other substance use, and financial insecurity.20–26 Little is known about the course of progression to OUD among these workers.

The purpose of this study was to quantify the risks from musculoskeletal disorders and new opioid prescriptions on incident long-term opioid use and OUD among construction workers. We examined two hypotheses: (1) among construction workers presenting to healthcare providers, those with acute or chronic musculoskeletal disorders will be more likely to initiate an opioid than those presenting with other diagnoses; and (2) among workers initiated on opioids, those with chronic musculoskeletal disorder diagnoses will be more likely to develop OUD than opioid users without chronic musculoskeletal disorders.

MATERIALS & METHODS

Study population:

We obtained de-identified pharmacy and health claims from the Saint Louis-Kansas City Carpenters Regional Benefit Plan, which manages the health coverage for workers and their dependents covering Missouri, Kansas, and southern Illinois. Health coverage eligibility requires active union membership status and an initial accrual of 500 work hours in a 6-month period, with an additional accrual of 300 work hours per quarter or 1200 hours per year for ongoing eligibility. The Plan created a dataset consisting of all outpatient health claims (dates of service, ICD-9/ICD-10 codes, procedure code, service provider), and pharmacy data (dates of fill, National Drug Code [NDC] code, drug names, fill number, quantity dispensed, days’ supply per prescription, and pharmacy identifier), and demographics (age, gender). These data covered all visits and services in the period from January 2015 through June 2018. The de-identified claims dataset was restricted to active union workers (n= 20,824) who were 18 years of age or older. Spouses and dependents were not included in the dataset received from the Health Plan. We excluded females (N = 915; 4.4% of workers) from the database, leaving 19,909 active union male construction workers in the study sample. Institutional review and ethics approval was obtained from the Human Subjects Office at Washington University in St. Louis, Missouri, USA. There was a waiver of consent as this was a secondary data analysis and the data was de-identified.

Measures

Outpatient health claims:

The international classification diagnostic (ICD) system changed to ICD-10 codes on October 1 of 2015, using ICD-9 codes prior to that date. First we selected the codes from each system. For claims beginning in October 2015, we identified diagnoses from relevant ICD-10 codes: acute musculoskeletal injuries (ICD10:S00-T14), chronic musculoskeletal disorders (ICD10:M.x [x = any number], G54-G57), dental procedures, and painful conditions of the abdominal region (ICD-10: kidney stones (N20) and abdominal or pelvic pain (R10)). All other diagnostic conditions were assigned to “other”. For the claims up to the end of September 2015, we used ICD-9 codes to identify comorbid conditions; we generated relevant ICD-9 codes by converting the ICD-10 codes using the general equivalence mapping (GEM) tool.27 For all data, we identified musculoskeletal surgery using Current Procedural Terminology (CPT) codes (CPT: 20005-21010; 21500-22900; 62263-63746; 64400-64999). Finally, we classified musculoskeletal diagnoses by one of three body regions: back/torso, upper extremity, lower extremity. We assigned body region to acute musculoskeletal injuries based on the first 2-digits of the ICD-10 S.x codes. ICD-10 chronic musculoskeletal disorder codes (ICD-10: M.x and G.x) were assigned to a body region by consensus of two authors (AMD, BFG).

Pharmacy data:

In the pharmacy claims, we identified prescribed opioids and benzodiazepines from the drug name and the corresponding NDC.28 Medication assisted treatment included all medications containing buprenorphine. We calculated morphine milligram equivalent (MME) doses using the CDC opioid conversion table29 and reported doses above the CDC thresholds of 50 and 90 MME/day.30 We used the number of days’ supply in a single prescription to identify supplies of more than seven days. Long-term opioids were defined as 60 or more days supplied within a calendar quarter.31

We linked opioid prescriptions to diagnoses occurring within 3 days prior to the fill. For workers with claims for more than one condition within the 3-day window, we linked the opioid prescription to the worker’s diagnoses by acuity using the following order of priority as recommended in the Centers for Medicaid and Medicare Services (CMS) guidelines:32 1) musculoskeletal surgery, 2) acute musculoskeletal injury, 3) chronic musculoskeletal disorder or recurrent injury, and 4) other (non-musculoskeletal conditions). When two or more opioid drugs were filled on the same day, the total days’ supply was based on the drug with the longest supply length (days’ supply from only one of the opioids) and the daily dose of the multiple opioids was based on the sum of the MME divided by the opioid with the longest supply length (number of days).

Data analysis:

For each member, we restricted analysis to the continuous period of coverage after becoming benefits eligible or starting January 2015 if already enrolled in the plan at study initiation. We calculated the annual average prevalence of opioid use stratified by mental health conditions, musculoskeletal conditions, other non-musculoskeletal health conditions, opioid-related conditions, and age (in 10-year increments) across the entire period. Because workers in the health plan contributed differing periods of time, we reported annual average results to aid comparison to other literature. For all analyses, we excluded workers with diagnoses of any of the following conditions prior to the initial opioid prescription: malignant or metastatic cancer (ICD-10: C00.x–C26.x, C30.x–C34.x, C37.x– C41.x, C43.x, C45.x–C58.x, C60.x–C76.x, C77.x–C85.x, C88.x, C90.x–C97.x); opioid use disorder (ICD-10: F11); poisoning from opioids (ICD-10: T40.0, T40.1, T40.3,T40.4 and T40.6); other psychoactive substance related disorders (ICD-10: F13, F14, F15, F16, F18, F19); or prescriptions for buprenorphine, a sedative, or a stimulant. All models were adjusted for age and the following mental health disorders: alcohol-related disorders (F10), nicotine dependence (F17), cannabis-related disorders (F12), mood affective disorders (F30-F39), and anxiety and other nonpsychotic mental disorders (F40-F49).

We quantified the annual incidence of receiving high-dose, large supplies (> 7 days), or long-term opioids by each health condition among opioid-naïve patients, defined as eligible to file a claim but who had not filled an opioid prescription for at least 60 days before the first opioid fill during the study, and had no history of OUD (F11), poisoning, or other illicit drugs (F13-F16, F18, F19). For single prescriptions, we computed the risk of receiving a high daily dose of opioids (>50 MME/day) and a supply of more than 7 days in multivariable log-binomial regression models for each health condition among opioid-naïve patients. We ran the same model to compute the risk of long-term opioid use within a year from the first opioid fill among opioid naïve users.

Finally, we quantified the odds ratio between diagnostic conditions and developing an opioid use disorder among all workers prescribed opioids at any point in the study period without regard to being opioid naive. We defined opioid use disorder as receiving an ICD-9/10 diagnosis of opioid use disorder (including poisoning), or receiving a buprenorphine prescription. All analyses were carried out with R statistical software.33

RESULTS

There were 19,909 male construction workers in the cohort during the 3.5-year period of observation, with an average of 16,707 workers eligible to file a claim each year (Table 1). On average, 92% of eligible workers filed a claim annually. The largest portion (46.4%) of claims were musculoskeletal conditions (N = 7752) of which 2,672 (34%) received an opioid prescription. Of 1644 workers undergoing surgery for musculoskeletal conditions, 1002 (61%) started opioids. On average, three percent of workers per year had new long-term opioid use (received a supply of at least 60 days of opioids in a calendar quarter). Nearly 3% of workers also received a prescription for a benzodiazepine within 30 days of filling an opioid prescription. In the 3.5-year study, 304 workers (1.5% of the total sample) received a diagnosis of opioid use disorder; 116 of these individuals had a diagnosis of OUD within 12 months of their first opioid prescription.

Table I.

Average annual construction worker demographics, diagnoses, and opioid prescriptions data from January 2015 to June 2018

Average per year Received prescription opioids per year
Eligible workers 16707 3443
had medical claim 15364 (92%) 3432 (99.7%)
Age at start of year
  under 30 2836 (17%) 354 (10.3%)
  30s 4334 (25.9%) 851 (24.7%)
  40s 3758 (22.5%) 788 (22.9%)
  50s 4372 (26.2%) 1095 (31.8%)
  60+ 1407 (8.4%) 354 (10.3%)
Mental health conditions
  Alcohol disorders (F10) 307 (1.8%) 120 (3.5%)
  Nicotine dependence (F17) 1483 (8.9%) 659 (19.1%)
  Mood disorders (ie depression, bipolar) F30-F39 773 (4.6%) 286 (8.3%)
  Nonpsychotic mental disorders (ie anxiety) F40-F48 1206 (7.2%) 438 (12.7%)
Cannabis disorders 56 (0.3%) 23 (0.7%)
Other illicit drug disorders 91 (0.5%) 35 (1%)
Other (non-MSK) health conditions
 Cancer (metastatic or malignant) 281 (1.7%) 132 (3.8%)
 Kidney stones 251 (1.5%) 168 (4.9%)
 Abdominal pain 802 (4.8%) 378 (11%)
 Dental visit (prevention or treatment) 6418 (38.4%) 1852 (53.8%)
Musculoskeletal (MSK) conditions
  Any MSK condition 7752 (46.4%) 2672 (77.6%)
 Chronic MSK disorder or recurring MSK injury 7151 (42.8%) 2504 (72.7%)
  Chronic MSK disorder* 7106 (42.5%) 2490 (72.3%)
  Recurring MSK injury* 686 (4.1%) 316 (9.2%)
 Acute MSK injury* 2734 (16.4%) 1195 (34.7%)
 MSK Surgery 1644 (9.8%) 1002 (29.1%)
  MSK Surgery and Chronic MSK disorder or recurring MSK injury 1473 (8.8%) 883 (25.7%)
  MSK Surgery and Acute MSK injury 394 (2.4%) 307 (8.9%)
MSK body region (acute or chronic)
 Back/torso MSK injury or disorder 5025 (30.1%) 1788 (51.9%)
 Lower extremity MSK injury or disorder 2435 (14.6%) 1020 (29.7%)
 Upper extremity MSK injury or disorder 2507 (15%) 1066 (31%)
Opioid-related conditions
 Opioid use disorders (OUD) (F11) 132 (0.8%) 61 (1.8%)
 Opioid poisoning (T40.0, T40.1, T40.3, T40.4, T40.6) 11 (0.1%) 3 (0.1%)
 Prescription for buprenorphine 67 (0.4%) 22 (0.6%)
Total opioid-related conditions (OUD, poisoning or buprenorphine) 151 (0.9%) 69 (2%)
 Long-term opioid use (>=60days supply in a calendar quarter) 518 (3.1%) 518 (15%)
 Benzodiazepine fill within 30 days of opioid fill 430 (2.6%) 430 (12.5%)
Open in a new tab
*

MSK definition: musculoskeletal; MSK surgery (CPT codes: 20005-21010; 21500-22900; 62263-63746; 64400-64999); acute injury (ICD10: S00.x-S99.x; T00.x-T14.x); chronic MSK (ICD10: M00-M25, M30-M36, M40-M54, M60-M79, M80-M94, G54.0-G57.0) or recurring acute (acute injury code repeated after 30 days)

The condition most likely (46%) to be treated with opioids was kidney stones (Figure 1), but this condition was rare (prevalence 1%; n = 251) and rarely led to long-term opioid use (2.7%). In contrast, chronic musculoskeletal disorders or recurring musculoskeletal injury was common (43%; n=7151) and workers with this diagnosis received prescriptions for opioids 24.1% of the time, and of those who started an opioid, 6.3% developed long-term opioid use. Musculoskeletal surgery was less common (prevalence 10% per year) but 40% of these surgical cases received an opioid prescription, with 7.4% developing chronic use. In contrast, only 1% of workers who initiated opioids for dental visits ended up with chronic use.

Figure 1:

Figure 1:

Open in a new tab

Prevalence of construction workers who received opioids within 3 days of diagnosis by the proportion with each diagnosis

Gray columns show the fraction of workers who were initiated on opioid therapy and black columns show the fraction of workers who had long-term opioid use, where long-term is defined as >= 60 days of prescription opioids supply in a subsequent 90-day period.

Column widths are proportional to the number of workers within each diagnostic category (see text for ICD-9 and -10 codes).

Musculoskeletal (MSK) Surgery column includes workers with MSK surgery, with or without chronic MSK disorder or acute MSK injury.

ICD: International Classification of Disease

Among the 5033 workers initiating opioids (Table 2), high-dose opioids (≥50 MME/day) were most commonly initiated for those undergoing musculoskeletal surgery (54%) compared to those without surgery (23%; relative risk (RR) 1.92; 95%CI 1.30-2.83). Fifteen percent of workers with chronic musculoskeletal disorders who started opioids began with >=50 MME/day. Opioid supply for > 7 days was common for those with chronic musculoskeletal disorders (41%, RR 2.65; 95%CI 1.48-4.73), and those with musculoskeletal surgery (28%, RR 1.67; 95%CI 1.23-2.27). Long-term opioid use (> 60 days per calendar quarter) was common (RR 4.34; 95%CI 1.81-10.40) among those with chronic musculoskeletal disorders within one-year of the first opioid prescription for this condition.

Table II.

Among 5033 workers initiating opioids, relative risk of high opioid dose and supply in first prescription and long-term opioids by diagnoses
Diagnoses
Chronic MSK disorder Acute MSK injury MSK surgery Dental visit Abdominal pain Kidney stones
Prevalence with diagnosis n (%) 892 (19%) 771 (16%) 654 (14%) 1384 (29%) 197 (4%) 179 (4%)
50 MME or more per day in first fill
  with diagnosis n(%) 130 (15%) 170 (22%) 356 (54%) 336 (24%) 52 (26%) 56 (31%)
  without diagnosis n(%) 1182 (30%) 1142 (28%) 956 (23%) 976 (29%) 1260 (27%) 1256 (27%)
  RR 0.52 0.81 1.92 0.89 1.07 1.16
  p-value <0.001 0.01 <0.001 0.09 0.59 0.21
>7 days supply in first fill
  with diagnosis n(%) 361 (41%) 95 (12%) 180 (28%) 22 (2%) 21 (11%) 13 (7%)
  without diagnosis n(%) 452 (12%) 718 (18%) 633 (15%) 791 (23%) 792 (17%) 800 (17%)
  RR 2.65 0.83 1.67 0.10 0.51 0.37
  p-value <0.001 0.13 <0.001 <0.001 0.001 <0.001
Long-term opioids* in a calendar quarter
  Prevalence of diagnosis among those with 1 year follow-up 671 (20%) 544 (16%) 452 (13%) 982 (29%) 134 (4%) 126 (4%)
  with diagnosis n(%) 76 (11%) 18 (3%) 9 (2%) 9 (1%) 4 (3%) 0 (0%)†
  without diagnosis n(%) 56 (2%) 114 (4%) 123 (4%) 123 (5%) 128 (4%) 132 (4%)
  RR 4.34 1.19 0.78 0.33 0.65 --
  p-value <0.001 0.59 0.53 0.004 0.39 --
Open in a new tab

Log-binomial multiple regression models, adjusted for age, mental health; Diagnoses are those within 3 days prior to opioid fill. Assignment of MSK diagnoses for multiple diagnoses in same claim by acuity: MSK surgery before acute injury before chronic MSK disorder; Opioid naïve defined as eligible to file a claim but opioid free for 60 days before fill and no history of opioid use disorder (F11), poisoning, or other illicit drugs (F13-F16, F18, F19).

*

long-term opioids are at least 60 days’ supply in a single calendar quarter within a year from first opioid fill

†

excluded from model

We also examined the odds of having an opioid use disorder within one year after filling an opioid prescription for a musculoskeletal condition versus several non-musculoskeletal diagnoses (kidney stones, abdominal pain, or a dental visit) (see Figure 2). Those diagnosed with chronic musculoskeletal disorders had the highest OR (95% CI) of developing opioid use disorder 4.71 (3.09-7.37) followed by those with musculoskeletal surgery 1.89 (1.26-2.79) or acute musculoskeletal injury 1.53 (1.02-2.26). Other diagnoses that were often prescribed opioids showed no elevated odds of developing opioid use disorder. MSK disorders or injuries of the back or torso were more frequently associated with incident opioid use (26.1%) than lower or upper extremity MSK disorders or injuries. Workers who received multiple opioid drugs in the same prescription were nearly five times more likely to develop an opioid use disorder (OR 4.84, 95% CI 2.96-7.71). Workers given high doses (>=50 MME/day) doubled their likelihood of opioid use disorder (OR 2.06, 95th CI 1.39-3.08), and those prescribed > 7 days’ supply were seven times more likely (OR 7.2, 95th CI 4.51-12.02). Workers with long-term opioid prescriptions were nearly ten times more likely (OR 9.95, 95th CI 6.66-14.94) to be treated for OUD.

Figure 2.

Figure 2.

Open in a new tab

Odds of opioid use disorder among those receiving opioid prescriptions (n=5057)

Forest plot of multivariable logistic regression models for the outcome of opioid use disorder (diagnosis of opioid use disorder or poisoning, or receiving buprenorphine) and a predictor variable (musculoskeletal condition, other diagnosis, or the opioid prescriptions) after adjusting for age, gender, mental health among workers who received opioids and followed for up to one-year after the fill of the first prescription.

Each opioid prescription was linked to a visit up to three days prior to the prescription. If more than one diagnosis was assigned on a single visit or claim, the assigned claim followed the order of acuity (musculoskeletal (MSK) surgery, acute MSK injury, chronic MSK disorder, other diagnosis). The odds ratio shows the odds of developing an opioid use disorder within one year after filling an opioid prescription assigned to one or more of the health conditions or to the dose and supply of opioids. The first opioid prescription filled in the database for each worker was used in this analysis.

MME: Morphine milligram equivalents, MSK: musculoskeletal

DISCUSSION

This study showed that nearly half of the construction workers were diagnosed with chronic musculoskeletal disorders, and one-quarter of those with chronic musculoskeletal disorders received an opioid prescription of whom 6.3% developed long-term opioid use. Those diagnosed with chronic musculoskeletal disorders were more than 4 times as likely to develop an OUD than those started on opioids for other diagnoses. Those initiating opioids for other diagnoses (including kidney stones, abdominal pain and dental visits) were not at increased risk of having an opioid dependency. High doses, prescriptions greater than 7 days, and prescriptions more than 60 days in a calendar quarter were strongly associated with a new OUD (approximately 2, 7, and nearly 10 times greater respectively, compared to those prescribed fewer opioids). Our first hypothesis was that workers with acute or chronic musculoskeletal disorders were more likely to initiate an opioid than those with other conditions. Instead, we found that those with a diagnosis of kidney stones and abdominal pain were most likely to receive opioids, although few workers had claims for these diagnoses. However, the greatest number of workers who received opioids had a musculoskeletal diagnosis. The data supported our second hypothesis: among workers initiated on opioids, workers with chronic musculoskeletal disorders were more likely to develop OUD than those prescribed opioids for other conditions including a (non-recurring) acute injury.

Our results showed high dosage of MME was not commonly prescribed to patients with chronic musculoskeletal disorders, differing from findings in older studies.30 Among workers initiating opioids, the first prescription of opioids exceeded the 50 MME threshold promulgated by the CDC30 in 15% of patients with chronic musculoskeletal disorders and in 54% of those who underwent musculoskeletal surgery. Like other studies,3, 34 we found that multiple prescriptions of opioids and long-term prescriptions were more important risk factors for developing OUD than high daily dose. Opioids were commonly prescribed for musculoskeletal surgery (61%), consistent with other studies.35 Our findings show that conditions associated with acute pain, like dental procedures, carry a relatively low risk for long-term opioid use. In contrast, chronic musculoskeletal disorders have a high risk of long-term opioid use. Others have shown a higher rate of opioid overdose among patients with long-term opioid use 36–39 further highlighting the risk of long-term opioid use. Our study also found that MSK disorders or injuries of the back or torso were more frequently associated with incident opioid use. Ly reported40 that among patients with claims for recurrent low back pain, 26.2% received tramadol and 32.3% received non-tramadol opioids. Furthermore, Ly found that these patients often receive opioids prior to trying physical therapy or NSAIDs.

Our results are consistent with studies that show opioids are commonly prescribed for acute and chronic musculoskeletal pain.41–43 However, the prevalence of opioid use and opioid use disorder among this construction cohort is much higher than other populations. Moshfegh and colleagues found that the risk of developing long-term opioid use among opioid-naïve patients with new musculoskeletal disorders was less than 1%,44 --far below the 6.3% risk we found among construction workers diagnosed with chronic musculoskeletal disorder and the 2.4% risk among the workers with acute musculoskeletal injuries. Another study,34 showed that 12% of construction workers received chronic prescriptions, suggesting workers seek treatment for ongoing pain. Notably, we used general health insurance data, not workers compensation data, and found high rates of acute and chronic musculoskeletal conditions associated with opioid use. The higher rates of musculoskeletal conditions captured in general health insurance data are likely the consequence of underreporting of work-related injuries in Workers’ Compensation data, reflecting cost-shifting of claims to general health insurance,45, 46 partly due to the difficulty of attributing chronic conditions to specific work incidents. In addition, the average annual prevalence of OUD in this sample of 0.8% is likely underestimated since only workers who seek medical treatment could receive a diagnosis and some workers may seek treatment for health conditions unrelated to OUD so fail to receive a diagnosis for OUD.47 Past studies suggest that other reasons for the underestimation of OUD diagnosis among construction workers is due to failure to seek treatment from lack of health insurance,48 suffering from mental distress so avoid seeking treatment,49 and low job security so fear of taking time off work.50–52 Despite this limitation, health claims provide a means to show associations between prescription opioids and development of OUD among those workers who seek treatment.

Due to the physically demanding nature of their work, construction workers are at high risk for chronic musculoskeletal disorders. Construction and mining industries have the highest rates of opioid prescriptions for work-related injuries.34 Healthcare providers should use caution when initiating opioids in construction workers and other laborers, as these patients have high physical exposures at work. The high rates of opioid use and overdose seen in hazardous occupations also emphasizes the need for primary prevention, to reduce physical exposures that lead to initial or recurring painful health conditions.53 Ideally, physical, complementary, and psychological therapies would be substituted for chronic opioid therapy.54

These findings highlight an important role that work injuries play in driving the opioid epidemic among hazardous or physically demanding occupations. A previous study of opioid overdoses in Utah found that half of those who died from an opioid related overdose had previously been treated for a job related injury.55 Data from Massachusetts12, 16 showed that the occupational groups with the highest rates of fatal overdose have high rates of musculoskeletal injuries. These data also showed that workers in occupations with lower availability of paid sick leave and lower job security had higher opioid-related mortality. These findings suggest that the need to return to work soon after an injury or to work while in pain might contribute to the high rates of opioid use and opioid-related overdose deaths seen among construction workers and other workers who often lack paid sick leave and have intermittent or insecure work.

Limitations

Health claims cannot account for all contributing factors to opioid misuse, including mental health conditions and other substance use, which are incompletely captured by claims. Long-term opioid use and opioid use disorder may be undercounted by health claims. These outpatient claims would not capture opioid prescriptions given before discharge from the hospital. Likewise, OUD is not always captured by diagnostic codes or prescription buprenorphine. Data from a single healthcare provider and pharmacy benefit plan are unable to capture all opioids taken by the workers, who may fill prescriptions outside of their health plan, or use illicit opioids. Workers may have filled opioid prescriptions from other providers or paid cash for prescriptions, which would not be recorded by the pharmacy benefit provider. We classified workers as “opioid naïve” if they had not been prescribed any opioid within a 60-day window before the first opioid prescription in our database, but acknowledge that some of these workers may have received opioids earlier. It is not known how many opioids were actually consumed from the filled prescriptions.

A strength of the study is the large sample of workers with individual level medical claims and prescriptions. Our study found robust associations between treatment for chronic musculoskeletal disorders and musculoskeletal surgery with long-term opioid use and opioid use disorder. Construction workers are at high risk given their high rates of overdose deaths.15 Clinicians and employers should consider their roles in preventing long-term opioid use and misuse. Employers should explore means to prevent work conditions that lead to pain among workers; clinicians should follow recommended guidelines,30 prescribing doses < 50 MME/day and supplies < 7 days of opioids, with careful monitoring and substituting alternative treatments for pain when feasible.

Conclusions

Opioids are commonly prescribed for pain relief among construction workers through their personal health insurance. A substantial number of workers who initiate opioids become long-term users, and these long-term users have a nearly 10-fold risk of developing opioid use disorder within a short period of time. Opioids should be more judiciously prescribed, particularly for workers in hazardous and physically demanding occupations, with greater use of alternative pain management treatments.

Acknowledgements:

The authors would like to acknowledge the St. Louis-Kansas City Carpenters Regional Benefit Plan who provided the data for this research study. Without their participation, these studies are not possible. We would also like to thank Lisa Hayibor from the Washington University in St. Louis Occupational Safety and Health Research Lab for her assistance with manuscript preparation.

Funding: Washington University Institute of Clinical and Translational Sciences grant from the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH); Grant number: UL1 TR000448.

Footnotes

Institution and Ethics approval and informed consent: Institutional review and ethics approval was obtained from the Human Subjects Office at Washington University in St. Louis, Missouri, USA. There was a waiver of consent as this was a secondary data analysis and the data was de-identified.

Disclosure (Authors): The authors declare no conflicts of interest.

Publisher's Disclaimer: Disclaimer: None

References

  • 1.Mundkur ML, Franklin JM, Abdia Y, et al. Days’ Supply of Initial Opioid Analgesic Prescriptions and Additional Fills for Acute Pain Conditions Treated in the Primary Care Setting - United States, 2014. MMWR Morb Mortal Wkly Rep. February 15 2019;68(6):140–143. doi: 10.15585/mmwr.mm6806a3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Zin CS, Nazar NI, Rahman NSA, Ahmad WR, Rani NS, Ng KS. Patterns of initial opioid prescription and its association with short-term and long-term use among opioid-naive patients in Malaysia: a retrospective cohort study. BMJ open. July 2 2019;9(7):e027203. doi: 10.1136/bmjopen-2018-027203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Edlund MJ, Martin BC, Russo JE, DeVries A, Braden JB, Sullivan MD. The role of opioid prescription in incident opioid abuse and dependence among individuals with chronic noncancer pain: the role of opioid prescription. Clin J Pain. July 2014;30(7):557–64. doi: 10.1097/ajp.0000000000000021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. Opioid Overdose: Prescription Opioids: Overview. https://www.cdc.gov/drugoverdose/data/prescribing/overview.html Accessed December 23, 2019.
  • 5.Mundkur ML, Rough K, Huybrechts KF, et al. Patterns of opioid initiation at first visits for pain in United States primary care settings. Pharmacoepidemiol Drug Saf. 2018;27(5):495–503. doi: 10.1002/pds.4322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.American Psychiatric Association Diagnostic and statistical manual of mental disorders. 5 ed American Psychiatric Publishing; 2013. [Google Scholar]
  • 7.Boscarino JA, Rukstalis MR, Hoffman SN, et al. Prevalence of Prescription Opioid-Use Disorder Among Chronic Pain Patients: Comparison of the DSM-5 vs. DSM-4 Diagnostic Criteria. J Addict Dis. 2011/07/01 2011;30(3):185–194. doi: 10.1080/10550887.2011.581961. [DOI] [PubMed] [Google Scholar]
  • 8.Bedson J, Chen Y, Ashworth J, Hayward RA, Dunn KM, Jordan KP. Risk of adverse events in patients prescribed long-term opioids: A cohort study in the UK Clinical Practice Research Datalink. Eur J Pain. May 2019;23(5):908–922. doi: 10.1002/ejp.1357. [DOI] [PubMed] [Google Scholar]
  • 9.Barnett ML, Olenksi AR, Jena AB. Opioid Prescribing by Emergency Physicians and Risk of Long-Term Use. N Engl J Med. May 11 2017;376(19):1896. doi: 10.1056/NEJMc1703338. [DOI] [PubMed] [Google Scholar]
  • 10.Krebs EE, Gravely A, Nugent S, et al. Effect of Opioid vs Nonopioid Medications on Pain-Related Function in Patients With Chronic Back Pain or Hip or Knee Osteoarthritis Pain: The SPACE Randomized Clinical Trial. JAMA. March 6 2018;319(9):872–882. doi: 10.1001/jama.2018.0899. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.March L ACP Journal Club: review: in knee and hip OA, opioids reduce pain and improve function but increase adverse events. Ann Intern Med. February 17 2015;162(4):Jc8. doi: 10.7326/acpjc-2015-162-4-008. [DOI] [PubMed] [Google Scholar]
  • 12.Hawkins D, Roelofs C, Laing J, Davis L. Opioid-related overdose deaths by industry and occupation-Massachusetts, 2011-2015. Am J Ind Med. October 2019;62(10):815–825. doi: 10.1002/ajim.23029. [DOI] [PubMed] [Google Scholar]
  • 13.Shah A, Hayes CJ, Martin BC. Factors Influencing Long-Term Opioid Use Among Opioid Naive Patients: An Examination of Initial Prescription Characteristics and Pain Etiologies. J Pain. November 2017;18(11):1374–1383. doi: 10.1016/j.jpain.2017.06.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Dong XS, Brooks RD, Cain CT. Overdose Fatalities at Worksites and Opioid Use in the Construction Industry, Fourth Quarter 2019. CPWR, The Center for Construction Research and Training; Updated 2020. https://www.cpwr.com/sites/default/files/publications/Quarter4-QDR-2019.pdf. [Google Scholar]
  • 15.Harduar Morano L, Steege AL, Luckhaupt SE. Occupational Patterns in Unintentional and Undetermined Drug-Involved and Opioid-Involved Overdose Deaths - United States, 2007-2012. MMWR Morb Mortal Wkly Rep. August 24 2018;67(33):925–930. doi: 10.15585/mmwr.mm6733a3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Massachusetts Department of Public Health. Opioid-related Overdose Deaths in Massachusetts by Industry and Occupation, 2011-2015. 2018. https://www.mass.gov/files/documents/2018/08/15/opioid-industry-occupation.pdf Accessed April 23, 2019.
  • 17.U.S. Bureau of Labor Statisitics. Case circumstances and worker characteristics for injuries and illnesses involving days away from work. https://www.bls.gov/iif/soii-data.htm Accessed December 23, 2019.
  • 18.Ma CC, Gu JK, Charles LE, Andrew ME, Dong RG, Burchfiel CM. Work-related upper extremity musculoskeletal disorders in the United States: 2006, 2009, and 2014 National Health Interview Survey. Work (Reading, Mass). 2018;60(4):623–634. doi: 10.3233/wor-182770. [DOI] [PubMed] [Google Scholar]
  • 19.Dale AM, Ryan D, Welch L, Olsen MA, Buchholz B, Evanoff B. Comparison of musculoskeletal disorder health claims between construction floor layers and a general working population. Original Article-Exposure Assessment. Occup Environ Med. 2015;72(1):15–20. Downloaded from oem.bmj.com on September 16, 2014 - Published by group.bmj.com. doi: 10.1136/oemed-2014-102313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Ciesielski T, Iyengar R, Bothra A, Tomala D, Cislo G, Gage BF. A Tool to Assess Risk of De Novo Opioid Abuse or Dependence. Am J Med. July 2016;129(7):699–705.e4. doi: 10.1016/j.amjmed.2016.02.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.National Academies of Sciences Engineering and Medicine. Pain Management and the Opioid Epidemic: Balancing Societal and Individual Benefits and Risks of Prescription Opioid Use. The National Academies Press; 2017:482. [PubMed] [Google Scholar]
  • 22.Jacobsen HB, Caban-Martinez A, Onyebeke LC, Sorensen G, Dennerlein JT, Reme SE. Construction workers struggle with a high prevalence of mental distress, and this is associated with their pain and injuries. J Occup Environ Med. October 2013;55(10):1197–204. doi: 10.1097/JOM.0b013e31829c76b3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Frone MR, Brown AL. Workplace Substance-Use Norms as Predictors of Employee Substance Use and Impairment: A Survey of US Workers. J Stud Alcohol Drugs. July 2010;71(4):526–534. doi:DOI 10.15288/jsad.2010.71.526. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Dong XS, Wang X, Largay JA. Occupational and non-occupational factors associated with work-related injuries among construction workers in the USA. Int J Occup Environ Health. March 2015;21(2):142–50. doi: 10.1179/2049396714y.0000000107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.King TL, Batterham PJ, Lingard H, et al. Are Young Men Getting the Message? Age Differences in Suicide Prevention Literacy among Male Construction Workers. Int J Environ Res Public Health. 2019;16(3):475. doi: 10.3390/ijerph16030475. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Airagnes G, Lemogne C, Meneton P, et al. Alcohol, tobacco and cannabis use are associated with job loss at follow-up: Findings from the CONSTANCES cohort. PLoS One. 2019;14(9):e0222361–e0222361. doi: 10.1371/journal.pone.0222361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Centers for Medicare and Medicaid Services. 2018 ICD-10 CM and GEMs. https://www.cms.gov/medicare/coding/icd10/2018-icd-10-cm-and-gems Accessed December 23, 2019.
  • 28.National Center for Injury Prevention and Control. CDC compilation of benzodiazepines, muscle relaxants, stimulants, zolpidem, and opioid analgesics with oral morphine milligram equivalent conversion factors, 2017 version. Centers for Disease Control and Prevention; Updated 2017. https://www.cdc.gov/drugoverdose/resources/data.html [Google Scholar]
  • 29.Centers for Disease Control and Prevention. CDC Guideline for Prescribing Opioids for Chronic Pain. https://www.cdc.gov/drugoverdose/prescribing/guideline.html Accessed December 30, 2019.
  • 30.Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain - United States, 2016. MMWR Recomm Rep. March 18 2016;65(1):1–49. doi: 10.15585/mmwr.rr6501e1. [DOI] [PubMed] [Google Scholar]
  • 31.Bree Collaborative. Opioid Prescribing Metrics. http://www.breecollaborative.org/wp-content/uploads/Bree-Opioid-Prescribing-Metrics-Final-2017.pdf Accessed December 30, 2019.
  • 32.The Centers for Medicare and Medicaid Services (CMS), National Center for Health Statistics (NCHS). ICD-10-CM Official Guidelines for Coding and Reporting: FY 2018 (October 1, 2017 - September 30, 2018). https://www.cms.gov/Medicare/Coding/ICD10/Downloads/2018-ICD-10-CM-Coding-Guidelines.pdf.
  • 33.R: A Language and Environment for Statistical Computing. Version 3.3.2. R Foundation for Statistical Computing; 2016. http://www.R-project.org. [Google Scholar]
  • 34.Thumula V, Liu T-C. Correlates of Opioid Dispensing. 2018. WC-18-48 https://www.wcrinet.org/reports/correlates-of-opioid-dispensing Accessed April 26, 2019.
  • 35.Kraut A, Raymond CB, Ekuma O, Shafer LA. A comparison of opioid use between WCB recipients and other Manitobans for knee, shoulder, back and carpal tunnel release procedures. Am J Ind Med. April 2016;59(4):257–63. doi: 10.1002/ajim.22562. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Dunn KM, Saunders KW, Rutter CM, et al. Opioid prescriptions for chronic pain aInd overdose: a cohort study. Annals of internal medicine. 2010;152(2):85–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Bohnert AS, Valenstein M, Bair MJ, et al. Association between opioid prescribing patterns and opioid overdose-related deaths. JAMA. April 6 2011;305(13):1315–21. doi: 10.1001/jama.2011.370. [DOI] [PubMed] [Google Scholar]
  • 38.Garg RK, Fulton-Kehoe D, Franklin GM. Patterns of Opioid Use and Risk of Opioid Overdose Death Among Medicaid Patients. Med Care. July 2017;55(7):661–668. doi: 10.1097/MLR.0000000000000738. [DOI] [PubMed] [Google Scholar]
  • 39.Nuckols TK, Anderson L, Popescu I, et al. Opioid prescribing: a systematic review and critical appraisal of guidelines for chronic pain. Ann Intern Med. January 7 2014;160(1):38–47. doi: 10.7326/0003-4819-160-1-201401070-00732. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Ly DP. Evaluation and Treatment Patterns of New Low Back Pain Episodes for Elderly Adults in the United States, 2011–2014. Med Care. February 2020;58(2):108–113. doi: 10.1097/MLR.0000000000001244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Calvo-Alen J Opioids in chronic musculoskeletal conditions. Ther Adv Musculoskelet Dis. October 2010;2(5):291–7. doi: 10.1177/1759720x10370237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Stokes A, Berry KM, Hempstead K, Lundberg DJ, Neogi T. Trends in Prescription Analgesic Use Among Adults With Musculoskeletal Conditions in the United States, 1999-2016. JAMA Netw Open. December 2 2019;2(12):e1917228. doi: 10.1001/jamanetworkopen.2019.17228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Sites BD, Harrison J, Herrick MD, Masaracchia MM, Beach ML, Davis MA. Prescription Opioid Use and Satisfaction With Care Among Adults With Musculoskeletal Conditions. Ann Fam Med. January 2018;16(1):6–13. doi: 10.1370/afm.2148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Moshfegh J, George SZ, Sun E. Risk and Risk Factors for Chronic Opioid Use Among Opioid-Naive Patients With Newly Diagnosed Musculoskeletal Pain in the Neck, Shoulder, Knee, or Low Back. Ann Intern Med. April 2 2019;170(7):504–505. doi: 10.7326/m18-2261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Dong XS, Fujimoto A, Ringen K, et al. Injury underreporting among small establishments in the construction industry. Am J Ind Med. May 2011;54(5):339–49. doi: 10.1002/ajim.20928. [DOI] [PubMed] [Google Scholar]
  • 46.Lipscomb HJ, Schoenfisch AL, Cameron W, Kucera KL, Adams D, Silverstein BA. Contrasting patterns of care for musculoskeletal disorders and injuries of the upper extremity and knee through workers’ compensation and private health care insurance among union carpenters in Washington State, 1989 to 2008. Am J Ind Med. 2015;58(9):955–963. doi: 10.1002/ajim.22455. [DOI] [PubMed] [Google Scholar]
  • 47.Dong XS, Brooks RD, Brown S. Musculoskeletal Disorders and Prescription Opioid Use Among U.S. Construction Workers. J Occup Environ Med. 2020;doi: 10.1097/jom.0000000000002017. [DOI] [PubMed] [Google Scholar]
  • 48.CPWR - The Center for Construction Research and Training. The Construction Chartbook. 2018;2020(October 26). https://www.cpwr.com/wp-content/uploads/publications/The_6th_Edition_Construction_eChart_Book.pdf
  • 49.Eyllon M, Vallas SP, Dennerlein JT, et al. Mental Health Stigma and Wellbeing Among Commercial Construction Workers: A Mixed Methods Study. J Occup Environ Med. 2020/08// 2020;62(8):e423–e430. doi: 10.1097/jom.0000000000001929. [DOI] [PubMed] [Google Scholar]
  • 50.To Battle The Opioid Overdose Epidemic, Deploy The ‘Cascade of Care’ Model. 2017. https://www.healthaffairs.org/do/10.1377/hblog20170313.059163/full/.
  • 51.Dong XS, Brooks RD, Cain CT. Prescription opioid use and associated factors among US construction workers. Am J Ind Med. 2020;63(10):868–877. doi: 10.1002/ajim.23158. [DOI] [PubMed] [Google Scholar]
  • 52.Center for the Promotion of Health in the New England Workplace. Opioids and Work: A Formative Research Assessment to Inform Educational Outreach. 2018. https://www.uml.edu/docs/Opioids%20and%20Work%20Formative%20Research_FINAL_Jan%2030%202019_tcm18-305156.pdf.
  • 53.Dale AM, Evanoff B, Macomber M, O’Reilly M, Rosen R, Schneider S. Can Ergonomics Programs Help Solve the Opioid Crisis? The Synergist: The Synergist; 2019. [Google Scholar]
  • 54.Owen GT, Bruel BM, Schade CM, Eckmann MS, Hustak EC, Engle MP. Evidence-based pain medicine for primary care physicians. Proc (Bayl Univ Med Cent). 2018;31(1):37–47. doi: 10.1080/08998280.2017.1400290. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Cheng M, Sauer B, Johnson E, Porucznik C, Hegmann K. Comparison of opioid-related deaths by work-related injury. Am J Ind Med. March 2013;56(3):308–16. doi: 10.1002/ajim.22138. [DOI] [PubMed] [Google Scholar]

RESOURCES