Opioid Prescribing on an Internal Medicine Teaching Unit
Samuel Quan, Cheyenne Lawton, Allison Budd, MD, FRCPC
Department of Medicine, University of Saskatchewan, Saskatoon, Canada
Author for correspondence: Allison Budd: acb782@mail.usask.ca
Received:
13 July 2020; Accepted after revision: 24 November 2020; Published: 21 June 2021
DOI:
https://doi.org/10.22374/cjgim.v16i2.443
ABSTRACT
Objective
To investigate the rationale and timing of opioid prescriptions for Internal Medicine inpatients
in an academic center in Saskatoon, Canada.
Methods
We performed a cross-sectional study of Internal Medicine inpatients that were prescribed
opioids in Saskatoon. We examined documentation of clinical rationale and timing of opioid
initiation or first escalation.
Results
Of 57 patients, 34 (60%) were opioid naive prior to admission and 48 (84%) had opioid doses
either initiated or escalated. Of these 48 patients, 27 (56%) occurred during on-call hours.
Rationale for escalation was documented in 31 cases (65%), with reasons including terminal
care (17%), musculoskeletal pain (15%), and skin and soft tissue infections (13%).
Conclusion
Rationale for opioid use was frequently not documented. Initial decision to change opioid dose
occurred equally between daytime and on-call hours.
RÉSUMÉ
Objectif
Étudier la justification et le moment choisi pour prescrire des opioïdes chez les patients hospitalisés
en médecine interne dans un centre universitaire de Saskatoon (Canada).
Méthodologie
Nous avons mené une étude de prévalence sur des patients hospitalisés en médecine interne chez
qui on a prescrit des opioïdes à Saskatoon. Nous avons examiné la documentation concernant
la justification clinique et le moment choisi pour entreprendre le traitement par les opioïdes ou
effectuer la première augmentation de dose.
14 V o l u m e 1 6 , I s s u e 2 , 2 0 2 1 C a n a d i a n J o u r n a l o f G e n e r a l I n t e r n a l M e d i c i n e
Original Research
CJGIM_2_2021_177932.indd 14CJGIM_2_2021_177932.indd 14 07/06/21 3:48 PM07/06/21 3:48 PM
Introduction
Canada is in the midst of an opiate crisis. Hospitalizations due to
opiate poisoning have been increasing in Canada.
1,2
This crisis
is in part driven by legal prescriptions,
2
making stewardship
of opioid prescribing more important than ever. As part of
the effort to mitigate opioid-related morbidity and mortality,
Canadian guidelines
3,4
recommend cautious opioid prescribing
and optimizing adjuvant pain control prior to starting opioids.
Multiple studies address opioid prescribing on surgical
units, but few studies have evaluated opioid use among medical
inpatients.
5
Internal medicine (IM) units care for patients with
a variety of presenting illnesses, many of which involve acute
or chronic pain. Nonopioid analgesics are recommended by
multiple guidelines prior to and along with opioids,
3,4
but this
practice is not well studied within the IM inpatient population.
Our study investigated the rationale, timing, and use of adjunct
therapies associated with opioid prescriptions on IM inpatients
in an academic hospital in Saskatoon, Canada.
Methods
A retrospective chart review with a cross-sectional design was
conducted investigating events surrounding the initiation or
escalation of opioid-prescribing practices for IM inpatients.
Population and Data Sources
We identified inpatients admitted to an IM clinical teaching unit
(CTU) at Royal University Hospital (RUH) between July 1, 2017
and June 30, 2018 who were given at least one in-hospital opioid
prescription. RUH is a tertiary academic center in Saskatoon,
Canada. Patients who underwent neuraxial or general anesthesia
within 30 days of admission or were prescribed methadone or
suboxone in the community were excluded. Opioids given for
major procedures (e.g., cardiac catheterization) were excluded.
Patients were identified using the RUH inpatient pharmacy
database, BDM Pharmacy. BDM Pharmacy tracks all inpatient
medication orders as well as the inpatient unit that patients are
discharged from. A list of inpatients discharged from an IM CTU-
specific ward with opioid prescriptions was generated by BDM
Pharmacy, and 60 unique IM patients were chosen by simple
random sample for chart review. If patients had multiple CTU
admissions within the study period, the first one was selected.
Chart review and Variables Collected
Details of opioid initiation or first escalation including the opioid
dose, route, frequency, service, who wrote the order, clinical
rationale for opioid use, and whether opioids were ordered during
daytime or on-call coverage hours were gathered from chart
review. Regular hours were defined as daytime (07:30–16:59) or
on-call hours (17:00–07:29). Medication reconciliation of home
medications, including opioids, was routinely performed for all
inpatients at the time of admission. Patients were considered
opioid naive if they did not have any active opioid prescriptions
at the time of admission. Opioid prescriptions given at the time
of discharge were also recorded.
Opioid escalation was defined as either: (i) an increase in
opioid dose or frequency, (ii) administration of intravenous
(IV) opioid, (iii) new breakthrough opioid dose prescribed
(different from a patients home opioid regimen), or (iv) a change
in opioid type with a greater maximum dosage when converted
to morphine equivalents.
Patient details were gathered, including age, sex, date of
admission and discharge, past history of any substance abuse,
and past or current liver disease. Prescriptions of adjunct
analgesics prior to opioid initiation or escalation, including
acetaminophen, nonsteroidal anti-inflammatory medications
(NSAIDs), pregabalin, or gabapentin, were recorded.
Statistical Analysis
Descriptive results were reported in medians with interquartile
ranges (IQR) or percentages. Characteristics were compared
Résultats
Des 57patients, 34 (60%) navaient jamais pris dopioïdes avant leur hospitalisation et 48 (84%)
ont reçu leur première dose d’opioïdes ou une augmentation de dose. De ces 48patients, 27
(56%) ont ru leur dose durant les heures de garde. La justification de laugmentation de dose
est documentée dans 31cas (65%), les raisons étant les soins de fin de vie (17%), la douleur
musculosquelettique (15%) et les infections de la peau et des tissus mous (13%).
Conclusion
Souvent, la justification de lutilisation des opioïdes nest pas documentée. Le moment où la
décision initiale de modifier la dose d’opioïde est prise est réparti de façon égale entre le jour et
durant les heures de garde.
C a n a d i a n J o u r n a l o f G e n e r a l I n t e r n a l M e d i c i n e V o l u m e 1 6 , I s s u e 2 , 2 0 2 1 15
Quan S et al.
CJGIM_2_2021_177932.indd 15CJGIM_2_2021_177932.indd 15 07/06/21 3:48 PM07/06/21 3:48 PM
between opioid-tolerant (with escalated doses) and opioid-naive
patients using student’s t-test or chi-square tests. A sensitivity
analysis was performed excluding patients who died in-hospital.
STATA version 13 was used.
Ethics Approval
Ethics approval was granted by the University of Saskatchewan
Health Research Ethics Board (#15–90).
Results
BDM pharmacy identified 704 IM patients admitted to CTU
and were prescribed opioids in-hospital, and 60 (8.5%) were
randomly selected for review. We excluded three patients due to
recent surgery. In this patient sample, the majority were elderly
(median age 69) and opioid naive prior to admission (60%;
Table 1). Most patients, 48 of 57 (84%), had opioid doses either
initiated or escalated from baseline (Table 1).
Among IM inpatients, opioid escalations occurred similarly
between on-call and day-time hours (56% occurred on-call; 95%
CI: 42–70%). Opioids were primarily prescribed by IM services
(63%; 95% CI: 49–76%). Among opioid escalations, hydromorphone
was most commonly prescribed (78%), followed by morphine
(21%); some patients had more than one opioid prescribed. After
Table 1. Characteristics of Internal Medicine Patients Who Received at Least
One Opioid Prescription in Hospital
Variable n (%)
Sample size n = 57
Female 29 (51%)
Age Median: 69, IQR: 49–81
Opioid naïve 34 (60%)
Opioid prescribed at home
a
Hydromorphone 11 (19%)
Codeine 4 (7%)
Fentanyl 3 (5%)
Oxycodone 3 (5%)
Tramadol 2 (4%)
Morphine 2 (4%)
Frequency of home prescription
Scheduled 7 (12%)
As needed 10 (18%)
Both 4 (7%)
History of substance abuse 13 (23%)
History of liver diseases 9 (16%)
Death in-hospital 14 (25%)
a
Some patients were prescribed more than one type of opioid.
excluding palliative cases and in-hospital deaths, four of the 24
(17%) patients who were opioid-naive prior to hospitalization
were given a new opioid prescription at discharge.
Documented rationale behind why opioids were started
or escalated was missing for 17 of 48 (35%; 95% CI: 22–49%)
cases (Figure 1). Among documented cases, rationale included:
terminal care (17%), musculoskeletal pain (15%), and skin and
soft tissue infections (13%).
Opioid-naive patients had similar characteristics compared
to opioid tolerant patients (Supplement 1), except that escalation
was less likely for skin and soft tissue infections (P < 0.05). A
sensitivity analysis was conducted, and results remain similar
after excluding in-hospital deaths (Supplement 2).
Although acetaminophen was prescribed in 71% of the
cases, only 31% of the patients had received acetaminophen
within 4 h prior to opioid escalation, and only 48% had received
acetaminophen within the previous 24 h (Figure 2). Acetaminophen
≥2 g was rarely given in the 24 h prior to prescription (8%).
Among the other adjunct agents, pregabalin/gabapentin (17%)
and oral NSAIDs (15%) were infrequently prescribed (Figure 3).
Discussion
We conducted a cross-sectional study evaluating opioid-prescribing
patterns among IM inpatients in Saskatoon, Canada. Indications
for opioid initiation or escalation varied, with terminal care
being the most common reason, although documentation
was not recorded in a third of cases. Opioids were initiated or
escalated relatively equally between the on-call and day-time
primary teams.
Opioids are frequently prescribed by IM physicians,
highlighting the importance of increasing the understanding of
opioid-prescribing patterns in IM. In a study examining over 56
million inpatient and outpatient opioid prescriptions, IM provided
the second highest number of opioid prescriptions.
6
Another
study of over 1 million inpatients at 286 hospitals showed that
opioids were used in 51% of nonsurgical admissions.
7
Documentation of clinical rationale for opioid prescriptions
is important, as opioids can be prescribed for a variety of
syndromes. Previously opioid-naive patients sent home with
opioid prescriptions are at an increased risk of opioid dependency.
8
Improved communication techniques, such as writing the
clinical indication in the order or the discharge prescription,
may also help facilitate transfers of care, opioid tapering, and
eventual discontinuation. Future studies should investigate the
influence of improved documentation on opioid de-escalation
and discontinuation.
Management of acute pain can be challenging on an inpatient
service. Severe pain may require urgent initiation of analgesic,
and patients preconceived notions of nonopioid analgesics may
16 V o l u m e 1 6 , I s s u e 1 , 2 0 2 1 C a n a d i a n J o u r n a l o f G e n e r a l I n t e r n a l M e d i c i n e
Internal medicine teaching unit
CJGIM_2_2021_177932.indd 16CJGIM_2_2021_177932.indd 16 07/06/21 3:48 PM07/06/21 3:48 PM
Figure 1. Indications documented for either initiation or escalation of opioids in-hospital.
influence choice of agent. Internists can face multiple competing
demands, such as: imminently ill patients, a busy ward, or multiple
consults for admission. Barriers to why nonopioid analgesics are
underutilized prior to and in combination with opioid use on
medical wards is a complex topic that warrants further study.
Canadian practice guidelines advise against initiating
opioid therapy in chronic, noncancer pain unless nonopioid
pain management strategies have been optimized.
3,4
However,
nonopioid analgesics were frequently underutilized in our study.
Acetaminophen was included as an option on both ER and IM
order sets at our institution. Acetaminophen was often ordered
on an as-needed basis for patients in this sample, yet rarely
administered. When given, it was in low doses; frequently less
than half of the maximum daily dose and below limits considered
safe even for use in patients with underlying liver dysfunction. A
large cohort study of multiple hospitals in Pennsylvania
5
found a
minority, approximately 20%, of nonintensive care inpatient opioid
users received either acetaminophen or NSAIDs. The authors
state that although some inpatients may have contraindications
to certain analgesics (e.g., peptic ulcer disease for NSAIDs), the
authors stated the prevalence of these conditions would not alone
explain the discrepancy in nonopioid analgesic use. We recognize
there are specific safety concerns related to NSAIDs, pregabalin,
and gabapentin on IM units, where the majority of patients are
elderly
9
and/or have multiple comorbidities, but opioids are also
associated with serious risks and adverse effects.
3,4
A substantial proportion (17%) of opioid escalations in
this study was provided for terminal care; this may be partially
explained by the increased breadth of IM practice at our center.
At RUH, IM frequently provides terminal care, with palliative
care available on a consulting basis. Intensive therapies, such
as noninvasive ventilation, are provided in observation units
on the ward under the care of IM, increasing the proportion
of patients with severe illness in relation to centers where that
level of care is provided in intensive care units.
Limitations of our study include its small sample size and
focus on opioid prescribing at a single academic hospital. Future
research conducted at other centers would be valuable to better
understand opioid prescription practice patterns on IM wards.
Contraindications to nonopioid analgesics, such as peptic ulcer
disease or renal dysfunction for NSAIDs, were beyond the
scope of this study. Finally, we did not investigate how inpatient
C a n a d i a n J o u r n a l o f G e n e r a l I n t e r n a l M e d i c i n e V o l u m e 1 6 , I s s u e 2 , 2 0 2 1 17
Quan S et al.
CJGIM_2_2021_177932.indd 17CJGIM_2_2021_177932.indd 17 07/06/21 3:48 PM07/06/21 3:48 PM