1. Original
Opioid Dose Reduction in a VA Health Care
System—Implementation of a Primary Care
Population-Level Initiative
Anders Westanmo PharmD, MBA,*
Peter Marshall MD,†
Elzie Jones PharmD,*
Kevin Burns PharmD,*
and Erin E. Krebs MD, MPH‡,§
*Department of Pharmacy, Minneapolis VA Health
Care System, Minneapolis, Minnesota, USA;
†
Department of Physical Medicine and Rehabilitation,
Comprehensive Pain Center, Minneapolis VA Health
Care System, Minneapolis, Minnesota, USA;
‡
Department of Medicine, Minneapolis VA Health Care
System, Minneapolis, Minnesota, USA; §
Department of
Medicine, University of Minnesota Medical School,
Minneapolis, Minnesota, USA
Reprint requests to: Peter Marshall, MD, Minneapolis
VAMC One Veterans Drive Minneapolis, MN 55417.
E-mail: peter.marshall@va.gov.
Disclosure: None of the authors has a conflict of inter-
est with the contents of this article. All authors contrib-
uted to the study design, interpretation of results, and
review, revision, and approval of the final manuscript.
ABSTRACT
Objective. To describe processes and outcomes of
a health system quality improvement initiative
designed to reduce opioid-related harms.
Design. The initiative was a primary care population-
level intervention to reduce high-dose opioid pre-
scribing, which was locally defined as >200
morphine-equivalent mg (MED) daily. We describe
the implementation process and report prescribing
rates and primary care provider (PCP) attitudes and
beliefs before and after implementation.
Setting. A VA health care system comprising one
large, urban teaching hospital and 11 outpatient
clinics in surrounding suburban and rural
locations.
Subjects. All patients who received any prescrip-
tion from the outpatient pharmacy (unique phar-
macy patients) were included in the population.
PCPs at the main hospital were surveyed.
Methods. Prescribing outcomes were determined
from merged VA databases by examining rates of
opioid dispensing within 90-day time windows
before and after implementation. PCP beliefs and
attitudes were evaluated with preimplementation
and postimplementation surveys.
Results. Following implementation, the number of
patients prescribed >200 MED daily decreased from
342 (0.65% of unique pharmacy patients) to 65
(0.12%). Overall, the number of unique pharmacy
patients who received at least one opioid prescrip-
tion within 90 days decreased from 6,942 (13.7%) on
April 1, 2011 to 5,981 (11.0%) on October 1, 2014
(13.8% decrease). Most PCPs agreed it was reason-
able for the medical center to set a 200 MED limit
(76% at baseline and 87% at follow up).
Conclusion. Opioid Safety Initiative implementation
was associated with a substantial reduction in high-
dose opioid prescribing. Factors that contributed to
initiative success included leadership support and
active clinical pharmacy engagement
Key Words. Chronic Pain; Opioid Analgesics;
Implementation; Primary Care
Introduction
The number of drug overdose deaths in the United
States increased every year between 2000 and 2010,
with much of this increase attributed to prescription
opioid analgesics [1,2]. Among people receiving opioids
for chronic pain, a growing body of literature has
1
Pain Medicine 2015; 00: 00–00
Wiley Periodicals, Inc.

2. demonstrated increasing risk of serious harms as the
dose of opioids increases [3–7].
The evidence for long-term benefits of opioids for
chronic noncancer pain is limited. A Cochrane review of
long-term opioids (>6 months) for chronic noncancer
pain concluded that the evidence for pain relief was
weak and evidence for improved quality of life or func-
tion improvement inconclusive [8]. Furthermore, avail-
able evidence does not suggest that treatment with
higher dose opioids confers greater benefit than treat-
ment with lower doses [9,10].
Various population based strategies have been pro-
posed to address risks of high-dose opioid prescribing
[11–14]. The Veterans Health Administration (VHA) has
recently initiated a nationwide Opioid Safety Initiative
(OSI) that includes goals of decreasing high-risk opioid
prescribing practices, including prescribing of high-dose
opioids. Prior to the national VHA initiative, the
Minneapolis VA Health Care System (MVAHCS) imple-
mented a primary care population-based OSI aimed pri-
marily at reducing high-dose opioid prescribing. This
article describes implementation processes and evalua-
tion of the Minneapolis OSI initiative.
Methods
Concerns about risks of high-dose opioids for chronic
noncancer pain led to an internal evaluation of opioid
prescribing practices in the MVAHCS in 2010 and to
implementation of the OSI in 2011. The primary focus of
the OSI was to reduce use of high-dose opioid therapy,
which was locally defined as >200 morphine-equivalent
mg (MED) daily. The initiative included a simultaneous
effort to shift prescribing away from oxycodone sus-
tained action (SA), due to a change in the health system
formulary. This project was conducted as a quality
improvement initiative. It was reviewed by the
Minneapolis VA IRB and determined not to be research.
Setting
At the time of the project, the MVAHCS comprised one
large, urban, tertiary care teaching hospital, and 11
community-based outpatient clinics (CBOCs) in sur-
rounding suburban and rural locations. Geographically,
the system extends from Western Minnesota through
the Western part of Wisconsin and from the Canadian
to the Iowa borders.
VHA primary care is delivered using a patient-centered
medical home model [15]. In this model, patients’ care
is managed by a primary care team that includes a pri-
mary care provider (PCP), along with dedicated nursing,
pharmacy, and mental health staff. Clinical pharmacists
are colocated in primary care clinics and have a broad
scope of practice, which includes medication manage-
ment for chronic conditions such as hypertension, dia-
betes, tobacco dependence, and pain. In the vast
majority of cases, long-term opioid therapy is prescribed
and managed in the primary care setting. At the time of
the initiative, a broad range of specialty services were
available, including orthopedics, physical medicine and
rehabilitation, and addiction psychiatry; there was no
multidisciplinary pain clinic on site.
In 2011, approximately 68,000 patients were enrolled
with a PCP within MVAHCS. Of these, 55% were seen
primarily at clinics located within the main hospital facil-
ity and 45% at one of the CBOCs. The OSI applied to
all clinic sites, although CBOCs received less support
for OSI implementation, as described below.
Implementation
The OSI included two main phases. A 1-year prepara-
tory phase included planning, pain management educa-
tion and training, and baseline evaluation. This was
followed by the implementation phase, in which active
support for patients and PCPs was provided, and opioid
dose reduction strategies were implemented.
In the preparatory phase, meetings were held with stake-
holders to ensure that they were informed and to clarify
their role in the OSI. Stakeholders included representa-
tives from primary care, addiction services, mental health,
pharmacy, laboratory, facility police, patient advocates,
public relations, and veterans’ service organizations.
Seminars were presented to Minneapolis facility-based
PCPs. These seminars addressed chronic pain manage-
ment and opioid utilization topics, as well as quality
improvement initiatives from other health systems. CBOC
providers did not have access to these presentations. All
PCPs were given a book on opioid management [16]. A
clinical pharmacist with training in pain management met
with the primary care-based clinical pharmacists to dis-
cuss OSI goals and processes and review clinical proce-
dures for opioid dose tapering [17].
Information about existing nonpharmacological pain
management services, including cognitive behavioral
therapy and yoga classes for chronic pain, was dissemi-
nated to PCPs. For patients, a full day pain education
course was developed to provide an introduction to
nonpharmacological pain management strategies. This
class was offered to all interested patients on a monthly
basis and was taught by a team that included psycholo-
gists, physicians, pharmacists, and physical therapists.
The implementation phase included endorsement from
executive leadership of the health care system, monitoring
of PCP prescribing with feedback to primary care teams,
and pharmacist support for pain medication changes and
opioid tapering. At the start of the implementation phase,
the Chief of Staff issued a memorandum to all PCPs estab-
lishing a health system goal that all patients receiving long-
term opioids for chronic noncancer pain should have their
daily dose reduced to less than 200 MED. The memo
explicitly stated that patients could request a new provider,
but opioid tapering goals would be consistent irrespective
of provider. Additional memos acknowledging progress
Westanmo et al.
2

3. and reiterating goals were sent from the Chief of Staff to
PCPs 2–3 times per year.
Patients prescribed >200 MED daily were identified, and
baseline opioid prescribing patterns were evaluated by
location and provider. PCPs at the Minneapolis facility and
CBOCs were given individual prescribing practice reports
including their patients receiving over 200 MED daily
(Figure 1). For each patient receiving opioids, OSI reports
provided the average daily MED on the date of the report
and at 3-, 6-, and 12-month prior, the date of the last urine
toxicology screen (within 1 year), and the date of the next
primary care appointment. Reports were later revised to
include information about active benzodiazepine prescrip-
tions, methadone prescriptions, and, for those on metha-
done, whether an ECG had been performed in the past
year. Providers were encouraged to review the reports
with the entire primary care team and develop tapering
plans if appropriate.
Additionally, the numbers of patients prescribed >200
MED daily and prescribed oxycodone SA were added
to monthly PCP clinical performance reports at the
Minneapolis facility. These reports allowed PCPs to
compare their opioid prescribing practices to those of
other PCPs in the facility. OSI measures were presented
in a manner consistent with that of other local perform-
ance measures, but had no impact on performance pay
or other positive or negative consequences.
Prescribing reports were generated using prescription,
provider, and laboratory data obtained from the VHA
Corporate Date Warehouse (CDW) using Structured
Query Language. The CDW contains extracts from VHA
clinical and administrative systems that contain com-
plete clinical data from October 1999 [18]. MED were
calculated using Washington Agency Medical Director’s
Group (AMDG) conversion ratios [19] and included all
prescriptions released from the pharmacy to an individ-
ual patient in the 90 days prior to the evaluation date.
Because the initiative was focused on opioid therapy for
chronic pain, not short-term prescriptions for acute indi-
cations, patients needed to have at least two opioid
prescriptions released in a 90-day evaluation period to
be considered for inclusion on prescribing reports. Lists
of patients were generated and sorted by assigned
PCP. If a patient did not have an assigned PCP at the
MVAHCS, additional steps were taken to identify the
opioid prescriber. Patients without one clearly defined
opioid prescriber were counted in the overall measure
and were reviewed by OSI team. If the assigned PCP
was not managing the patient’s pain or if the treatment
was for cancer pain or palliative care, the PCP was
advised to contact the OSI team with this information so
reports could be corrected.
Clinical pharmacists provided substantial support for the
OSI at the main facility and at some CBOCs; approxi-
mately 75% of primary care teams had an affiliated clini-
cal pharmacist located in the primary care clinic.
Primary care pharmacists received OSI reports as
described above for the PCPs in their clinics.
Pharmacists were encouraged to develop and discuss
tapering plans with PCPs and other primary care team
members. If requested, pharmacists provided case
management to assist with implementing tapers.
Evaluation
The primary outcome for this analysis, consistent with
the primary target of the OSI project, was prescribing of
high-dose opioid therapy, defined as >200 MED daily.
Because studies have demonstrated increased harms
associated with a variety of “high dose” thresholds [3–5]
we also report on prescribing of >50 MED, >100 MED,
>120 MED, and >400 MED daily. Prescribing of spe-
cific opioid drugs (e.g., oxycodone SA, morphine SA,
methadone) over time is reported as a secondary
outcome.
Data Analysis
Prescribing outcome data were obtained from the VHA
CDW, as described above, and evaluated within 90-day
cross-sectional time windows. Within each evaluation
window, all patients who received any prescription from
the outpatient pharmacy (unique pharmacy patients)
were included in the population. All opioid analgesics
that were US DEA controlled substances were included
unless they were used for addiction treatment (i.e.,
buprenorphine/naloxone and concentrated 10 mg/mL
methadone); tramadol was not included because it was
not a controlled substance at the time of the initiative.
Mean prescribed opioid daily doses for each patient
were calculated within evaluation windows as the sum
of all MED released from the pharmacy in the time win-
dow, divided by 90 days. Rates of each dose and drug
outcome were calculated for pre-OSI (90 days prior to
April 1, 2011) and post-OSI (90 days prior to October 1,
2014) evaluation windows as the proportion of all unique
Figure 1 Sample OSI PCP prescribing practice report.
Implementation of Opioid Dose Reduction
3

4. pharmacy patients meeting each outcome. The absolute
change in the proportion of patients meeting each out-
come was calculated by subtracting the pre-OSI rate
from the post-OSI rate. The relative change is the abso-
lute change divided by the pre-OSI rate.
Assessment of Provider Beliefs and Attitudes
A paper-based survey was administered to PCPs at the
Minneapolis hospital before and after the OSI. These sur-
veys were disseminated at provider meetings and by fol-
low up email at the start of the initiative (February 2012)
and again 2 years later (January 2014). The survey was
designed evaluate beliefs and attitudes about opioid pre-
scribing and to cover the following domains: importance
of standards of care, perceived barriers to and benefits
of dose limits on opioid prescribing, and adequacy of
training and support for pain management. PCPs were
asked to indicate how strongly they agreed or disagreed
with 18 statements about use of opioids to treat chronic
nonmalignant pain. Response options were “strongly
agree,” “agree,” “neither agree or disagree,” “disagree,”
and “strongly disagree.” For reporting purposes,
responses of “strongly agree” and “agree” were catego-
rized as agreement. Results were used to inform OSI
planning and implementation. Surveys were anonymous,
so individual responses could not be paired for analysis.
Results
Opioid Prescribing
In the 90 days prior to the start of the OSI in April 2011, the
MVAHCS provided any prescription medication to 50,749
individual patients (i.e., unique pharmacy patients); at that
time, 342 individual patients (0.67% of unique pharmacy
patients) received more than 200 MED of opioids per day.
At follow up in October 2014, the number of unique phar-
macy patients was 54,636, and 65 of these patients
(0.12%) were receiving more than 200 MED of opioids
daily. Figure 2 shows the number of patients receiving
more than 200 MED daily over the entire follow up period.
Additionally, we observed decreases in prescribing at all
MED thresholds examined, with greater decreases seen at
higher MED thresholds (Table 1).
We observed an overall decrease in opioid prescribing
rates and doses over the study period. The number of
patients who received at least one opioid prescription
within 90 days decreased from 6,942 (13.7% of unique
pharmacy patients) on April 1, 2011 to 5,981 (11.0%)
on October 1, 2014 (13.8% decrease). The mean dose
among those who received an opioid prescription
decreased from 43 to 23 MED daily (47% decrease).
The number of patients receiving oxycodone SA
decreased from 292 to 3 over the study time period.
The number of patients receiving other long-acting
opioids, as well as hydrocodone/acetaminophen, hydro-
morphone, and oxycodone/acetaminophen also
decreased (Table 2).
Provider Perceptions
Survey response rates were 74% (34 completed/46 pro-
viders) at baseline and 64% (31 completed/48 providers)
at follow up. The vast majority of respondents agreed at
both time points that standards of care for opioid prescrib-
ing were important and that more support was needed to
care for patients with chronic pain (Table 3). The proportion
of PCPs who agreed that it was reasonable for the medical
center to set a 200 MED/day limit was 76% at baseline
and 87% at follow up. A minority was satisfied with the
care being provided to patients with chronic pain at base-
line (9%) and follow up (26%). The two most commonly
endorsed barriers to lowering doses <200 MED/day were
patients becoming upset (62% baseline and 64% follow
up) and pressure from patient service representatives or
the administration (59% baseline and 22% follow up). The
most commonly endorsed benefit was improving patient
safety (85% at baseline and 87% at follow up).
Discussion
The Minneapolis VAHCS OSI was associated with a
substantial reduction in high-dose opioid prescribing.
We also found reductions in overall dispensed opioid
doses and overall rates of opioid receipt, although these
were not the focus of the OSI.
We believe many factors contributed to the apparent suc-
cess of the initiative. One key factor was explicit and sus-
tained support for OSI goals by clinical leaders and the
organization as a whole. Pre-OSI surveys demonstrated
that the majority of PCPs were concerned about pressure
from patients or patient advocates to continue prescribing
high doses of opioids. Prior reports have similarly found
that some PCPs feel pressured to prescribe opioids to
maintain patient satisfaction and avoid being “fired” [20].
The OSI proactively addressed these concerns by obtain-
ing endorsements from clinical leadership, providing
education for patient advocates and veterans’ service
Figure 2 Number of patients receiving >200
MED daily. Daily morphine-equivalent doses (MED)
were calculated for the 90-day window preceding
each date.
Westanmo et al.
4

5. organizations, and updating policies to discourage pro-
vider switching due to opioid-related disagreements.
Although patient complaints and requests for provider
switches were not systematically assessed, they
occurred less often than the project team anticipated.
Anecdotally, a small number of patients were very dis-
satisfied with the initiative. Some of these patients had
been on stable high-dose regimens without adverse
effects for many years and did not view their treatment
as unsafe, given their own experience. Other patients
reported a positive experience with opioid dose reduc-
tion, despite initial misgivings.
A second major factor that contributed to the success
of the OSI was the involvement of clinical pharmacists
embedded in primary care teams. Evidence of phar-
macist management effectiveness is well established
for chronic conditions such as diabetes and hyperten-
sion and is accumulating for chronic pain [21–24]. For
example, a recent meta-analysis found that chronic
pain educational interventions by pharmacists led to
improvements in both adverse events and patient sat-
isfaction [25]. In the Minneapolis VAHCS, primary care
pharmacists already had a broad scope of practice
that included pain management. They were able to
refer to other clinical services (e.g., physical therapy,
mental health), order laboratory tests, and prescribe
nonopioid medications to aid in pain management. In
addition, they had previously received training in moti-
vational interviewing, which was viewed as helpful in
facilitating patients’ active role in medication changes.
Referrals from PCPs to primary care pharmacists for
opioid tapers and other analgesic management
increased early in the OSI implementation period.
Pharmacists followed tapering guidance included in
the VA/DOD Clinical Practice Guideline for
Management of Opioid Therapy for Chronic Pain [17],
but typically used less aggressive taper schedules
than the 20–50% weekly reductions recommended in
Table 1 Change in prescribing rates at selected high-dose thresholds
Number (%) of Patients
Pre-OSI (April 1, 2011) Post-OSI (October 1, 2014) Absolute Change (%) Relative Change (%)
>50 MED 1256 (2.47) 811 (1.49) 20.98 239.68
>100 MED 712 (1.40) 303 (0.55) 20.85 260.71
>120 MED 626 (1.23) 229 (0.42) 20.81 265.85
>200 MED 342 (0.67) 65 (0.12) 20.55 282.09
>400 MED 126 (0.25) 11 (0.02) 20.23 291.89
* Pre-OSI and post-OSI rates are proportions of all unique pharmacy patients (pre-OSI n 5 50,749 and post-OSI n 5 54,636)
receiving prescribed opioid doses above the given threshold. The absolute change is the post-OSI rate minus the pre-OSI rate.
The relative change is the absolute change divided by the pre-OSI rate.
Table 2 Changes in prescribing of specific opioid medications*
Number (%) of Patients
(4/1/2011)
Pre-OSI, n (%)
(10/1/2014)
Post-OSI, n (%)
Absolute
Change, %
Relative
Change, %
Long-acting opioids
Fentanyl TD 94 (0.19) 94 (0.17) 20.01 27.11
Methadone 286 (0.56) 164 (0.30) 20.26 246.74
Morphine SA 831 (1.64) 770 (1.41) 20.23 213.93
Oxycodone SA 292 (0.58) 3 (0.01) 20.57 299.05
Short-acting opioids
Hydrocodone/APAP 4058 (8.00) 3480 (6.37) 21.63 220.34
Hydromorphone IR 164 (0.32) 150 (0.27) 20.05 215.04
Morphine IR 100 (0.20) 113 (0.21) 0.01 4.96
Oxycodone/APAP 1562 (3.08) 1345 (2.46) 20.62 220.02
Oxycodone IR 679 (1.34) 1124 (2.06) 0.72 53.76
APAP 5acetaminophen; IR 5 immediate release; OSI 5 Opioid Safety Initiative; SA 5 sustained action; TD 5 transdermal.
*Pre-OSI and post-OSI rates are proportions of all unique pharmacy patients (pre-OSI n5 50,749 and post-OSI n5 54,636). The
absolute change is the post-OSI rate minus the pre-OSI rate. The relative change is the absolute change divided by the pre-OSI rate.
Implementation of Opioid Dose Reduction
5

6. the guideline. Anecdotally, a typical pharmacist-
managed taper schedule for a patient on high-dose
opioids was 10% decrease per month.
Finally, the OSI attempted to increase awareness and
use of nonpharmacological pain management strategies
by providing education for both primary care teams and
patients. Despite these strategies and a perception of
increased nonpharmacological pain service use, we
found no pre-OSI to post-OSI improvement in the 90%
of PCPs who reported a need for additional pain man-
agement resources. We believe that successfully shifting
the dominant model of care from a pharmacologically-
based approach to an integrated or behaviorally-based
model will require more dedicated resources than this
initiative was able to provide.
Prior clinical initiatives have focused on deescalating
opioid prescribing. An important early effort was led by
the Washington state AMDG, which implemented a
guideline and web-based tool kit beginning in 2007 and
updated in 2010 [19]. A major feature of the AMDG initia-
tive was a “yellow-flag” warning if opioid dose exceeded
120 daily MED without a pain specialty consultation. The
Washington state initiative was strengthened by a law
passed in 2010, which directed additional resources
toward support and tools for guideline implementation,
including simple tools for tracking pain and function, an
Table 3 PCP beliefs and attitudes before and after OSI implementation
Number (%) Agree or Strongly Agree
Pre-OSI Post-OSI
Absolute
Change,* %
Importance
I’m satisfied with the care being provided to patients with chronic nonmalig-
nant pain
3 (9) 8 (26) 17
I have adequate training and skills to care for my patients with chronic pain 11 (32) 9 (29) 23
I could use more help caring for my patients with chronic pain 31 (91) 28 (90) 21
It is important for the medical center to have a consistent standard of care
for opioid prescribing
33 (97) 31 (100) 3
It is important for the primary care section to have a consistent standard of
care for opioid prescribing
34 (100) 31 (100) 0
It is important for me to have a consistent standard of care for opioid
prescribing
34 (100) 31 (100) 0
It is reasonable for the medical center to set a dose limit of 200 ME/day for
safe opioid prescribing in patients with chronic noncancer pain
26 (76) 27 (87) 11
Barriers
There are no good alternatives to using high doses of opioids for my chronic
pain patients
12 (35) 7 (23) 212
If I try to lower doses to <200 ME/day, my patients may become upset and
perhaps threatening or violent
21 (62) 20 (64) 2
There is insufficient evidence that prescribed opioid doses >200 ME/day
increase the risk of overdose and accidental death
2 (12) 5 (16) 4
If I try to limit opioid doses to <200 ME/day I’ll have to deal with pressure
from the patient service representatives or administration
20 (59) 7 (22) 237
Keeping prescribed opioid doses <200 ME/day will worsen pain control for
many patients
2 (6) 4 (13) 7
It will be difficult for me to determine how many ME my patients are getting 9 (27) 6 (19) 28
Benefits
Keeping prescribed opioid doses <200 ME/day will improve patient safety
and reduce the risk of accidental deaths
29 (85) 27 (87) 2
Keeping prescribed doses <200 ME/day will improve patients’ quality of life 20 (59) 17 (55) 24
Keeping prescribed doses <200 ME/day will reduce utilization of health care 13 (39) 11 (36) 23
Keeping prescribed doses <200 ME/day will reduce drug diversion 16 (47) 16 (52) 5
Keeping prescribed doses <200 ME/day will help protect me as a prescrib-
ing physician
22 (65) 20 (65) 0
ME 5 morphine equivalent mg; OSI5 Opioid Safety Initiative. The absolute change is the post-OSI rate minus the pre-OSI rate.
Westanmo et al.
6

7. online MED calculator, and patient education materials.
Additional direction and guidance was provided for phar-
macy benefit managers, insurance companies, managed
care organizations, and third party administrators to track
high-dose patients, offer assistance to high prescribers,
and provide incentives for best practices. These interven-
tions achieved a 35% reduction in the percent of patients
receiving >120 MED per day among workers in the
Washington State workers’ compensation system.
Furthermore, although effects of the AMDG initiative can-
not be distinguished from those of other policy and prac-
tice changes, a 50% decrease in the death rate among
injured workers occurred after its implementation [26].
Other opioid management improvement initiatives have
aimed to increase opioid safety by focusing on stand-
ardizing pain care processes and increasing use of
opioid monitoring practices. Group Health Cooperative
implemented a multifaceted opioid risk reduction initia-
tive that included guidelines for prescribing, physician
training, peer support, and electronic medical record
tools [27]. This initiative resulted in increased use of
urine drug testing and, although dosing was not an
explicit target of the initiative, a 17% decrease in the
mean prescribed daily opioid dose [28]. A VA regional
health system located in California and Nevada imple-
mented an electronic medical record-based opioid
safety dashboard to provide prescribers with access to
customizable reports on high dose prescribing and
patient risk factors (e.g., depression, substance use dis-
orders). Implementation of this dashboard was also
associated with decreased prescribing of high-dose
opioids [29].
Our study has important limitations. First, this is an
uncontrolled evaluation of a quality improvement project
that describes prescribing changes over time at a health
system level, not among a specific cohort of patients.
We cannot exclude secular effects or other confounding
factors as the cause of observed changes; likely, grow-
ing general awareness of opioid-related harms contrib-
uted to the changes we observed. We are also unable
to exclude the possibility that changes in prescribing
were due to changes in the population seen in the
health system over time; however, we do not think this
is likely. The overall number of patients seen increased
over the time period covered, but there was no sub-
stantial change in the overall demographics or charac-
teristics of the population during this time. Second, a
major limitation of our OSI evaluation is the lack of
patient-reported outcomes, such as pain severity, func-
tional status, patient satisfaction, and clinical outcomes,
such as overdose deaths. As a result, we cannot deter-
mine whether patients benefited from the observed
changes in prescribing practices. Additionally, we can-
not rule out the possibility of unintended patient harms
associated with the initiative. Although we are not aware
of any serious harm events among patients who were
receiving high-dose opioids at the start of the initiative, it
is possible that undetected harms occurred. Finally, we
describe a project conducted in one VA health care sys-
tem. Although our project was implemented at multiple
clinical sites with differing resources, our results may not
generalize to other VA and nonVA settings.
Findings from the Minneapolis VAHCS OSI may inform
other VA and non-VA health systems in their efforts to
change opioid prescribing practice. In early 2014, the
Veterans Health Administration announced systemwide
goals of a nationwide OSI that include decreasing high-
risk opioid prescribing practices, including prescription of
>200 MED daily and coprescription of opioids and ben-
zodiazepines, and increasing use of opioid risk manage-
ment tools such as urine drug testing at all VA facilities.
In summary, we found that a local VA-based OSI was
associated with reduced opioid daily doses, consistent
with the primary goal of the initiative. Furthermore, we
observed substantial changes in opioid prescribing prac-
tice that extended beyond the specific goals of the initia-
tive. However, most PCPs continued to express a need
for additional pain management resources, suggesting
that augmented pain services should be included in
future initiatives. Future research should evaluate effects
of prescribing practice changes on patient outcomes.
References
1 Paulozzi LJ, Jones C, Mack K, Rudd R. Centers for
Disease Control and Prevention (CDC). Vital signs:
Overdoses of prescription opioid pain relievers—
United States, 1999–2008. MMWR Morb Mortal
Wkly Rep 2011;60(43):1487–92. Available at: http://
www.cdc.gov/mmwr/preview/mmwrhtml/
mm6226a3.htm (accessed October 2014).
2 Okie S. A flood of opioids, a rising tide of deaths. N
Engl J Med 2010;363(21):1981–3.
3 Bohnert ASB, Valenstein M, Bair MJ, et al.
Association between opioid prescribing patterns and
opioid overdose-related deaths. JAMA 2011;
305(13):1315–21.
4 Dunn KM, Saunders KW, Rutter CM, et al. Opioid
prescriptions for chronic pain and overdose: A
cohort study. Ann Intern Med 2010;152(2):85–92.
5 Gomes T, Mamdani MM, Dhalla IA, Paterson JM,
Juurlink DN. Opioid dose and drug-related mortality
in patients with nonmalignant pain. Arch Intern Med
2011;171:686–91.
6 Bohnert ASB, Ilgen MA, Trafton JA, et al. Trends and
regional variation in opioid overdose mortality among
Veterans Health Administration patients, fiscal year
2001 to 2009. Clin J Pain 2014;30(7):605–12.
7 Edlund MJ, Martin BC, Russo JE, et al. The role of
opioid prescription in incident opioid abuse and
Implementation of Opioid Dose Reduction
7

8. dependence among individuals with chronic non-
cancer pain. Clin J Pain 2014;30:557–64.
8 Noble M, Treadwell JR, Tregear SJ, et al. Long-term
opioid management for chronic noncancer pain.
Cochrane Database Syst Rev 2010;(1):CD006605.
9 Naliboff BD, Wu SM, Schieffer B, et al. A randomized
trial of 2 prescription strategies for opioid treatment of
chronic nonmalignant pain. J Pain 2011;12(2):288–96.
10 Chen L, Vo T, Seefeld L, et al. Lack of correlation
between opioid dose adjustment and pain score
change in a group of chronic pain patients. J Pain
2013;14(4):384–92.
11 Perrone J, Nelson LS. Curbing the opioid epidemic
in the United States: The risk evaluation and mitiga-
tion strategy (REMS). JAMA 2012;308(5):457–8.
12 McLellan AT, Turner BJ. Chronic noncancer pain
management and opioid overdose: Time to change
prescribing practices. Ann Int Med 2010;152(2):123–4.
13 Paulozzi LJ, Weisler RH, Patkar AA. A national epi-
demic of unintentional prescription opioid overdose
deaths: How physicians can help control it. J Clin
Psychiatry 2011;72(5):589–92.
14 Sullivan MD. Limiting the potential harms of high-dose
opioid therapy. Arch Int Med 2011;171(7):691–3.
15 U.S. Department of Veterans Affairs 2010. Patient
Aligned Care Team (PACT). [Internet] Available at:
http://www.va.gov/health/services/primarycare/pact/
index.asp (accessed October 2014).
16 Fishman SM. Responsible Opioid Prescribing: A
Clinicians Guide, 2nd edition. Washington, DC:
Waterford Life Sciences, 2012.
17 The Management of Opioid Therapy for Chronic Pain
Working Group. VA/DoD clinical practice guideline for
the management of opioid therapy for chronic pain.
Version 2.0. Washington, DC: U.S. Veterans Health
Administration and Department of Defense. 2010.
18 Health Services Research & Development.
Corporate data warehouse [Internet]. Washington
(DC): Department of Veterans Affairs, Health
Services Research and Development Service; 2012
[updated 2014 Mar 28]. Available at: http://www.
hsrd.research.va.gov/for_researchers/vinci/cdw.cfm
(accessed October 2014).
19 Washington Agency Medical Director’s Group.
Interagency guideline on opioid dosing for chronic,
non-cancer pain: An educational aid to improve
care and safety with opioid therapy. 2010 update.
Available at: http://www.agencymeddirectors.wa.
gov/opioiddosing.asp (accessed October 2014).
20 Matthias MS, Parpart AL, Nyland KA, et al. The
patient-provider relationship in chronic pain: Providers’
perspectives. Pain Med 2010;11(11):1688–97.
21 Brushwood DB. From confrontation to collaboration:
Collegial accountability and the expanding the role
of pharmacists in the management of chronic pain.
J Law Med Ethics 2001;28(S4):69–93.
22 Bruhn H, Bond CM, Elliott AM, et al. Pharmacist-led
management of chronic pain in primary care:
Results from a randomised controlled exploratory
trial. BMJ Open 2013;3(4):1–12.
23 Briggs M, Closs SJ, Marczewski K, Barratt J. A feasi-
bility study of a combined nurse/pharmacist-led
chronic pain clinic in primary care. Qual Prim Care
2008;16:91–4.
24 Jouini G, Choiniere M, Martin E, et al.
Pharmacotherapeutic management of chronic non-
cancer pain in primary care: Lessons for pharma-
cists. J Pain Res 2014;7:163–73.
25 Bennett MI, Bagnall AM, Raine G. Educational inter-
ventions by pharmacists to patients with chronic
pain: Systematic review and meta-analysis. Clin J
Pain 2011;27(7):623–30.
26 Franklin GM, Mai J, Turner J, et al. Bending the pre-
scription opioid dosing and mortality curves: Impact
of the Washington state opioid dosing guideline. Am
J Ind Med 2012. 55(4):325–31.
27 Trescott CE, Beck RM, Seelig MD, Von Korff MR.
Group Health’s initiative to avert opioid misuse and
overdose among patients with chronic noncancer
pain. Health Aff (Millwood) 2011;30(8):1420–4.
28 Turner JA, Saunders K, Shortreed SM, et al.
Chronic opioid therapy risk reduction initiative:
Impact on urine drug testing rates and results.
J Gen Intern Med 2014;29(2):305–11.
29 Kryskalla J, Kern S, Gray D, Hauser P. Using dash-
board technology to monitor overdose risk. Fed
Pract 2014;31(9):32–8.
Westanmo et al.
8