Do Opioids Help Injured Workers Recover and Get Back to Work? The Impact of Opioid Prescriptions on Duration of Temporary Disability

• Date: 01 October 2019
• Author: David Neumark,Randall Lea,Bogdan Savych
• Published date: 01 October 2019
• DOI: http://doi.org/10.1111/irel.12243

Do Opioids Help Injured Workers Recover and

Get Back to Work? The Impact of Opioid

Prescriptions on Duration of Temporary

Disability

BOGDAN SAVYCH, DAVID NEUMARK and RANDALL LEA

*

We estimate the effect of opioid prescriptions on the duration of temporary dis-

ability benefits among workers with work-related low-back injuries, based on

variation in local opioid prescribing patterns, which predict whether injured work-

ers receive opioid prescriptions. More longer term opioid prescribing leads to con-

siderably lengthier durations of temporary disability.

Introduction

When used in accordance with evidence-based guidelines, opioids are an impor-

tant tool for health-care professionals in their quest to provide compassionate and

reasoned care. However, what began as a well-intentioned effort to relieve under-

treated pain has developed into an opioid crisis of epidemic proportions (U.S.

Department of Health and Human Services 2018). Aside from the direct adverse

effects of opioid use and addiction, there may be important consequences for key

economic outcomes. Indeed, some speculate that the rising use of opioids has con-

tributed to the decline in U.S. labor-force participation of men (Krueger 2017).

While short-term opioid therapy can provide pain relief to some patients,

the benefits of opioid therapies must be weighed against risks of addiction

*The authors’affiliations are, respectively, Workers Compensation Research Institute, Cambridge, Mas-

sachusetts. E-mail: bsavych@wcrinet.org; UCI, Irvine, California, NBER, Cambridge, Massachusetts, IZA,

Bonn, Germany, and Workers Compensation Research Institute,Cambridge, Massachusetts. E-mail: dneu-

mark@uci.edu; and Workers Compensation Research Institute, Cambridge, Massachusetts. E-mail: rd-

lortho@aol.com. The authors would like to thank Richard Butler, Frank Neuhauser, John Ruser, Vennela

Thumula, Dongchun Wang, and anonymous referees for helpful feedback and comments on this work. Help-

ful comments were also received from many practitioners, and from participants at the Association for Pub-

lic Policy Analysis and Management Fall Conference, the American Economic Association Annual

Meetings, and the American Society of Health Economists Annual Conference. Any errors or omissions

remaining are the responsibility of the authors. The views are their own, and do not necessarily reflect those

of the Workers Compensation Research Institute.

INDUSTRIAL RELATIONS, DOI: 10.1111/irel.12243. Vol. 58, No. 4 (October 2019). ©2019 The Regents of

the Univer sity of California . Published by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148,

USA, and 9600 Garsington Road, Oxford, OX4 2DQ, UK.

549

development, abuse, or other potential side effects (see the review in Noble

et al. 2010). Evidence suggests that patients who receive multiple opioid pre-

scriptions or patients who are prescribed large daily doses are at greater risk of

experiencing overdoses (Bohnert et al. 2011; Dunn et al. 2010; Franklin et al.

2005; Gomes et al. 2011; Paulozzi et al. 2012). Opioid use is associated with

nonfatal overdose hospitalizations (Coben et al. 2010), increased likelihood of

emergency department visits (Braden et al. 2010), and historically has been a

major cause of deaths from unintentional poisoning (Paulozzi 2012; Paulozzi

and Annest 2007).

Opioid use is common among workers injured at work. Recent studies show

that more than half of injured workers who are absent from work more than

seven days, who do not receive surgery but receive pain medications, received

an opioid prescription. Many of these workers received opioids on a longer

term basis (Thumula, Wang, and Liu 2017). This raises a question about the

potential impact of opioid prescriptions on outcomes that workers experience

after an injury: Given the risks from opioids, are there nonetheless important

benefits that might make the trade-off worthwhile? We focus on a key poten-

tial benefit from the point of view of workers’compensation policy—the dura-

tion of temporary disability, a proxy for time away from work while workers

are recovering from their injuries. Although some of the adverse effects of opi-

oid use would be expected to lengthen duration of temporary disability, there

could also be some benefits, via pain reduction, which enable faster return to

work (which can be viewed as a benefit as long as it does not impede maxi-

mum recovery).

1

To address this question, we use detailed workers’compensa-

tion claims data to examine the relationship between multiple measures of

opioid prescribing and the time that injured workers spend on temporary dis-

ability benefits while recovering from an injury.

Work-related injuries represent a substantial share of injuries that occur to

working adults. Nearly half of all trauma injuries to working adults in 2008–

2010 were deemed to be work related and were covered by workers’compen-

sation insurance, and one in five injuries for soft-tissue conditions were

deemed work related (Victor, Fomenko, and Gruber 2015). This suggests that

opioid medications provided in the workers’compensation system are a non-

trivial part of the prescribing to which working adults are exposed.

Several studies establish a correlation between opioid prescribing and longer

durations of temporary disability benefits. However, this relationship could be

1

Many other aspects of return to work that may be important to injured workers and employers are not

examined in this research. Those may include the nature of employment at the time of return to work, the

nature of interactions with other workers, and the need for specific job modifications or accommodations

(for a review, see MacEachen, Irvin, and Franche 2006).

550 / BOGDAN SAVYCH,DAVID NEUMARK AND RANDALL LEA

noncausal, driven by prescribing of opioids for more severe injuries that, inde-

pendently, are associated with longer durations of disability. While past studies

tried to control for injury severity, unmeasured injury severity may be corre-

lated with both opioid prescribing and the duration of temporary disability.

Moreover, opioid prescriptions may be a marker for worker characteristics that

result in longer time away from work unrelated to the actual opioids, or, con-

versely, workers who want to return to work more quickly may choose to use

opioids to speed up their return to work. Thus, a causal analysis of the effects

of opioid prescriptions on the duration of disability is needed.

In this paper, we use empirical methods designed to estimate the causal effect

of opioid prescriptions on the duration of disability. We rely on an instrumental

variables (IV) approach that isolates the variation in individual opioid prescrip-

tions that is driven by local prescribing patterns, rather than by individual charac-

teristics, preferences, or decisions of workers (or their providers) that pertain to

injury severity or desire for faster return to work. This strategy assumes that local

prescribing patterns influence whether an injured worker receives prescriptions

for opioids, in a manner that is independent of the characteristics, preferences, or

decisions of workers or their providers that directly affect the duration of disabil-

ity.

2

Our results reflect a combination of the effects opioids may have on the

duration of disability because of reductions in pain as well as the potential addic-

tive nature of opioids—effects that presumably act in opposite directions. We do

not examine therapeutic, pain management effects of opioids.

The assumption underlying our identification strategy is plausible because,

given the rich controls included in our models, we should account for injury

severity, treatment differences, and other sources of variation in the propensity

to return to work that could otherwise be correlated with both opioid prescrib-

ing and the duration of temporary disability. Moreover, we present a number

of supplemental analyses intended to explore threats to the validity of our IV

—and these analyses closely replicate our main results. That said, one can

never prove that an identifying assumption is valid, and we note possible

remaining factors that could raise questions about a causal interpretation of our

IV estimates. Our evidence is nonexperimental, and hence how compelling this

identification is rests on how one views the validity of our instrumental vari-

ables research design, and the analyses we present exploring threats to this

validity. Nonetheless, if one is skeptical, our evidence is still informative about

the associations between local opioid prescribing patterns and the duration of

2

This is consistent with prior studies documenting sizable differences in physician practice patterns

across regions that are not explained by differences in case mix, other treatment differences, or measurable

differences in health outcomes, but are related to variation in regional practice norms (Pasley et al. 1987;

Yasaitis, Bynum, and Skinner 2013).

Opioids Prescribing and the Duration of Temporary Disability / 551