Opioid Prescribing Among Hospitalized Patients in Tertiary Care Hospitals: A Retrospective Cohort Study
Caitlin Roy, BSP, ACPR, MSc, Holly Mansell, PhD, Charity Evans, PhD, Shenzhen Yao, PhD, Casey Phillips, BSP, PharmD, Carmen Johnson, MD, David Blackburn, BSP, ACPR, PharmD 2024-05-28 13:48:30
ABSTRACT
Objectives: Hospitalization may be an important source of opioid prescriptions in the community. We aimed to describe opioid prescribing for inpatients of 2 tertiary care hospitals in Canada.
Methods: We conducted a retrospective cohort study in 2 Canadian hospitals using electronic discharge abstracts and inpatient prescription records for adults admitted to medicine or surgery units between 2017 and 2019. Opioid prescriptions were characterized by dosage, duration, and concomitant prescriptions. A random effects logistic regression model was built to identify independent predictors of opioid prescriptions on the day of discharge in patients with a medicine admission.
Results: Of the 56,302 patients included, the mean age was 62 years, 19,946 (52.2%) were female, 32,472 (57.7%) were admitted to a medicine unit, and 15,114 (26.8%) to surgery. At least 1 inpatient opioid prescription was observed for 65.1% of all admissions (n= 36,626/56,302). Among all patients receiving inpatient opioid prescriptions, virtually all were prescribed a strong opioid (96.8%, 35,437), and 67.8% (24,834) included an intravenous route of administration. Inpatient opioid prescriptions were active for an average of 87.1% of the hospitalization; however, most individuals received asneeded opioid prescriptions only (70.7%, 25,899). On the day of discharge, at least 1 active inpatient opioid prescription was identified in 55.2% (24,467) of all patients in the cohort. Two factors were highly predictive of an active inpatient opioid prescription on the day of discharge: duration of the opioid prescription and opioids prescribed as needed only.
Conclusions: Inpatient opioid prescriptions are currently ordered for a high percentage of hospitalized patients, and they often remain active on the day of discharge. The prescribing patterns identified provide targets for strategies to reduce unnecessary opioid exposure.
Keywords: Opioid, utilization, prescribing, inpatient, hospital, quality improvement
Affiliation: University of Saskatchewan, College of Pharmacy and Nutrition, Saskatoon, Saskatchewan, Canada, Saskatchewan Health Authority, Regina, Saskatchewan, Canada, Vancouver Coastal Health, Vancouver, British Columbia, Canada.
Corresponding Author: David Blackburn, BSP, ACPR, PharmD, FCSHP, College of Pharmacy and Nutrition, University of Saskatchewan, 2A20.20Ñ107 Wiggins Road, Saskatoon, Saskatchewan, Canada. E-mail: d.blackburn@usask.ca
This research was conducted with internal research funds.
The authors declare no conflicts of interest.
Supplemental Digital Content is available for this article. Direct URL citations are provided in the HTML and PDF versions of this article on the journal's website, www.canadianjournalofaddiction.org.
Copyright © 2024 by the Canadian Society of Addiction Medicine
DOI: 10.1097/CXA.0000000000000206*
Objectifs: L’hospitalisation peut être une source importante de prescriptions d’opioïdes dans la communauté. Nous avons cherché à décrire la prescription d’opioïdes chez les patients hospitalisés dans deux hôpitaux de soins tertiaires au Canada.
Méthodes: Nous avons mené une étude de cohorte rétrospective dans deux hôpitaux canadiens en utilisant les résumés de sorties électroniques et les dossiers de prescriptions des patients hospitalisés pour les adultes admis dans les unités de médecine ou de chirurgie entre 2017 et 2019. Les prescriptions d’opioïdes ont été caractérisées par le dosage, la durée et les prescriptions concomitantes. Un modèle de régression logistique à effets aléatoires a été construit pour identifier les prédicteurs indépendants des prescriptions d’opioïdes le jour de la sortie chez les patients ayant été admis en médecine.
Résultats: Sur les 56 302 patients inclus, l’âge moyen était de 62 ans, 19 946 (52,2%) étaient des femmes, 32 472 (57,7%) ont été admis dans une unité de médecine et 15 114 (26,8%) en chirurgie. Au moins une prescription d’opioïdes pour les patients hospitalisés a été observée pour 65,1% de toutes les admissions (n =36 626/56 302). Parmi les patients ayant reçu une prescription d’opioïdes en hospitalisation, la quasi-totalité s’est vu prescrire un opioïde fort (96,8%, 35 437), et 67,8% (24 834) ont été administrés par voie intraveineuse. Les prescriptions d’opioïdes pour les patients hospitalisés ont été actives pendant 87,1% de la durée de l’hospitalisation en moyenne ; cependant, la plupart des personnes ont reçu des prescriptions d’opioïdes uniquement en cas de besoin (70,7%, 25 899). Le jour de la sortie, au moins une prescription active d’opioïdes en milieu hospitalier a été identifiée chez 55,2% (24 467) de tous les patients de la cohorte. Deux facteurs étaient hautement prédictifs d’une prescription active d’opioïdes pour patients hospitalisés le jour de la sortie : la durée de la prescription d’opioïdes et les opioïdes prescrits uniquement en cas de besoin.
Conclusions: Les ordonnances d’opioïdes pour les patients hospitalisés sont actuellement prescrites pour un pourcentage élevé de patients hospitalisés et elles restent souvent actives le jour de la sortie. Les schémas de prescriptions identifiés fournissent des cibles pour des stratégies visant à réduire l’exposition inutile aux opioïdes.
Mots clés: Opioïde, utilisation, prescription, patient hospitalisé, hôpital, amélioration de la qualité
INTRODUCTION
North America is amidst an opioid crisis, with Canada and the United States identified as the 2 largest consumers of opioids worldwide.1 High rates of opioid use are leading to unprecedented rates of opioid-related morbidity and mortality, with numbers continuing to rise. Individuals who use opioids have a higher risk for hospitalizations, emergency department visits, and use of illicit drugs. Of the 955 opioid-related deaths reported in Ontario in 2016, 32.5% had an active opioid prescription at the time of death. Nonmedical use of opioids often occurs in people with a previous medical prescription intended for the individual, a friend, or a family member. Indeed, up to 10% of Canadians who use opioid medications engage in some form of problematic use,7 including a possible risk of escalating to heroin use.
Despite their widespread use for the medical management of pain, evidence suggests opioids may not be superior to alternative therapies. Chronic pain guidelines recommend nonpharmacological and nonopioid pharmacotherapies as first-line options. Opioids have demonstrated a pain relief response or improvement in physical function comparable to alternative therapies, such as nonsteroidal anti-inflammatory drugs (NSAIDs), anticonvulsants, tricyclic antidepressants, or nabilone (a synthetic cannabinoid). Opioids should be reserved for patients with inadequate pain relief despite optimization of these first-line therapies and, when used, in limited amounts. These recommendations are influenced by the association between prescribing intensity and opioid related harm.
Up to 13% of community-dwelling Canadians are prescribed an opioid every year, many of whom receive prescriptions that exceed recommended doses or duration of therapy.1 Although the source of the first prescription in these patients is often unknown, recently hospitalized patients may be an important contributor to the epidemic. It is well known that opioids are prescribed following hospital discharge, particularly in the surgical population. However, few studies are available to describe current inpatient opioid prescribing practices in surgical and nonsurgical patients. Herzig et al (2009- 2010) and Donohue et al (2010-2014) both found that opioids were prescribed to half of all inpatients in US hospitals, often in combinations, and at moderately high doses. Not only are these studies outdated, but their results are difficult to interpret given the potentially safe environment for opioid administration in hospital settings. Hospitalized patients are acutely ill, so high rates of opioid use could be acceptable if individuals are weaned off before discharge. In the current study, we measured overall opioid utilization based on each inpatient hospitalization; however, we also measured the utilization of opioids restricted to active prescriptions on the day of discharge. It was felt the latter endpoint may be a better risk indicator for inpatients who may be discharged with opioid prescriptions for use at home.
Given the significant attention to opioid prescribing in recent years, it is possible that opioids are no longer prescribed at such a high rate in hospital settings. We aim to describe and characterize inpatient opioid prescribing practices in a large cohort of hospitalized patients admitted to tertiary care hospitals in Canada. Our analysis was exploratory in nature to provide a general indication of current practice with respect to inpatient opioid prescribing.
METHODS
Study design and data sources
This retrospective cohort study linked patient-level discharge abstracts with inpatient prescriptions from 2 acute care, tertiary care hospitals in Saskatchewan, Canada. Both hospitals provided acute care services to a city of ~342,000 residents and its surrounding area, with 450 beds in hospital 1 and 275 beds in hospital 2. Discharge abstracts generated after every discharge, transfer, or inpatient death contained patient demographics, hospitalization details (eg, type of admission, clinical service, admit/discharge dates, disposition), and up to 25 diagnoses coded using the International Classification of Diseases, 10th Revision (ICD-10). For every patient, all inpatient prescriptions occurring between the admission and discharge date were captured by BDM Pharmacy (version 10.33), a software program used by hospital pharmacists to process inpatient opioid prescriptions. The BDM database contained details about medications and prescribed regimens but did not confirm medication administration. Opioid inpatient prescriptions were defined as medications used primarily for pain listed on the hospital’s formulary including codeine, fentanyl, hydromorphone, meperidine, morphine, oxycodone, hydrocodone, pentazocine, remifentanil, sufentanil, tapentadol, and tramadol. Medications for opioid use disorder (buprenorphine and methadone) were excluded to avoid overestimation of the results.
Population
Eligible patients were adults (18 y or older) admitted between 2017 and 2019, hospitalized at least one night, and admitted under an eligible service (ie, medicine, surgery, cardiology, neurology, nephrology, hematology, obstetrics postpartum, alternate level of care, or trauma). We excluded patients receiving palliative care (ICD-10 code Z51.5), oncology care (admitted to an oncology ward or patient admissions with oncology diagnosis), inpatient psychiatry, antepartum care, labor and delivery, patient admissions without any inpatient prescriptions, admissions longer than 365 days, and those who expired during the hospitalization. Patients with an admission date falling within one day after a previous discharge were linked as one hospitalization. Patients were allowed multiple admissions during the study period.
For each patient, we collected demographics and hospitalization details (eg, admission type, length of hospitalization, and discharge disposition; Appendix A for definitions, Supplemental Digital Content 1, http://links.lww.com/CJA/A33). Admission type was categorized as medicine (including general medicine, nephrology, hematology, cardiology, and neurology), surgery (including general, orthopedic, vascular, cardiovascular, neuro-, and gynecology surgery, and urology), trauma (fractures, injuries, and burns), or other. For patient transfers, the first admission type was carried forward for the remaining records within that transfer. Admitting diagnosis was reported according to broad ICD-10 categories (Appendix A, Supplemental Digital Content 1, http://links.lww.com/CJA/A33). Comorbidities likely to be associated with opioid use were also reported: musculoskeletal pain conditions, depression, alcohol use disorder, drug dependence/opioid poisoning, and other mental health disorders.
Opioid prescription characterization
For each hospital admission, we identified all inpatient prescriptions for opioids and further characterized them as (a) either scheduled-dose (eg, 10 mg every 12 h) or as-needed–dosing (eg, 10 mg PRN); (b) type and route of administration; (c) the percentage of total hospitalized days with an active opioid prescription; (d) occurrence of active orders on the same day for 2 or more types of opioids; and (e) use of nonopioid analgesics and benzodiazepines. Nonopioid analgesics captured included acetaminophen, NSAIDs (including COX-2 inhibitors and acetylsalicylic acid 325 mg or greater), topical therapy, oral antidepressants, and enteral anticonvulsants. See Appendix A, Supplemental Digital Content 1, http://links.lww.com/CJA/A33 for definitions of all variables.
Finally, we identified all patients with an active inpatient opioid prescription on their day of discharge. These orders were characterized as scheduled dose versus as-needed and the total daily dosage for all scheduled doses in oral milligram morphine equivalents (MME) was calculated. The dosage calculation excluded patients receiving patient-controlled analgesia (PCA), pain pumps, neuraxial, or topical administration as they could not be well characterized with the databases available. The dosage estimation on the day of discharge only captured scheduled doses (ie, excluded as-needed doses) with a fixed dosage (ie, 10 mg). Prescriptions were excluded from the dosage calculation if they contained a dose range (eg, “5-10 mg” were not captured by the database) and those reported as zero (order entry error or dose indicated in directions text). Data analysis Baseline characteristics and opioid utilization indices were presented descriptively using means and SDs for normally distributed variables or medians with interquartile ranges (IQRs) for nonparametric data. Continuous variables with skewed distribution or high variance were transformed into categorical variables (Appendix A, Supplemental Digital Content 1, http://links.lww.com/CJA/A33](http://links.lww.com/CJA/A33)). Categorical variables were reported using frequencies and proportions. A random effects logistic regression model was built to examine predictors of an active inpatient opioid prescription on the day of discharge. The patient ID variable was assigned as the random effect since certain individuals were admitted more than once during the study period. This final analysis was carried out on a subgroup of patients with a medicine-type admission, had at least 1 opioid prescription during the hospitalization, and were discharged home. All baseline characteristics were tested for an association with the outcome. Those with a significance level less than 0.10 on univariate analysis were entered into the multivariable model in a forward stepwise approach and maintained if they improved model fit as measured by − 2Loglikelihood ratio. Multicollinearity was identified by a variance inflation factor >2.5.18 First-order interactions were examined for variables in the final model. Statistical significance was defined as P<0.05. All analyses were conducted utilizing SAS Software v9.4 (SAS Institute Inc, Cary, NC, USA). This study was approved on scientific and ethical grounds by the University of Saskatchewan Biomedical Research Ethics Board (ID 1512) with waiver of informed consent.
RESULTS
The complete dataset contained 2,402,812 inpatient prescription observations for 54,385 patients admitted between January 1, 2017, and December 31, 2019. A total of 16,167 patients (607,876 drug prescriptions) were excluded (Figure 1). Ultimately, 38,218 patients were included in the study, representing 56,302 admissions (ie, 9,954 patients with more than 1 admission).
Patients included in the study were a mean age of 62 years; 50.3% (19,241/38,218) were 65 years of age or older, and just over half were female (52.2%, 19,946/ 38,218) (Table 1). For the 56,302 patient admissions, the most common admission type was medicine (57.7%, 32,472/56,302), followed by surgery (26.8%, 15,144/ 56,302). Most patients were discharged to home (78.6%, 44,252/56,302) (Table 1).
Two-thirds (65.1%, 36,626/56,302) of all hospital admissions received at least 1 inpatient opioid prescription. The highest opioid prescription rates were in patients admitted to surgery (96.0%, 14,502/15,144) and trauma services (96.0%, 4603/4795). Three strong-potency opioid agents accounted for over 90% of agents prescribed (hydromorphone, morphine, and fentanyl) (Table 2). In almost half of cases (46.3%, 16,955/36,626), combination opioids were prescribed. Intravenous was the most common route of administration (67.8%, 24,834/36,626). Patient-controlled analgesia (PCA) was prescribed for 2224/36,626 (6%) patient admissions, mostly in those admitted to a surgical unit [1894 (85%) in surgery; 186 (8%) in trauma; and 83 (4%) in medicine]. Almost all patient admissions with an opioid also had an inpatient prescription for acetaminophen at some point during their hospitalization (94.7%, 34,694/36,626).
Among all hospital admissions receiving opioid prescriptions, 29.3% (10,727/36,626) were written for a scheduled dose, while 70.7% (25,899/36,626) were prescribed as needed only. Only 8.3% (887/10,727) of admissions received scheduled-dose opioids in the absence of additional as-needed prescriptions. On average, opioid prescriptions were active for 87.1% (SD 22.8) of the admissions (Table 2).
On the day of discharge, over half (55.2%) of all admissions had an active inpatient opioid prescription on their profile (Table 3). While rates were highest in trauma and surgery units (92.2% 2384/2,586 and 87.8% 11,892/13,537, respectively), active opioid prescriptions on the day of discharge were also observed for over one-third (36.5% 9,763/26,739) of patients admitted to medicine units. Most active opioid orders on the final day were for asneeded doses only (82.4%, 20,171/24,467). In the 3531 patients receiving scheduled opioid doses on the day of discharge, the median MME on the day of discharge was 30 (IQR 30 to 60 excluding as-needed and missing doses); in 18.6%, the daily dose was higher than 89 MME.
For the 10,057 patients discharged home from a medicine unit, the strongest predictor of an active opioid on the final day of hospitalization was the duration of the inpatient prescription (Table 4). The odds of an opioid prescription remaining active on the day of discharge were 8.14 times higher for those with opioid prescriptions lasting more than 75% of their hospitalization versus less than 25% (62.4% vs. 17.1%, OR 8.14, 95% CI: 6.64-9.97). In addition, the odds were also significantly increased for those who received asneeded opioid prescriptions only during their hospitalization versus those with scheduled prescriptions also (77.8% vs. 22.2%, OR 1.56, 95% CI: 1.43-1.69). Factors associated with a reduced odds of an active opioid on discharge were diagnosis of a comorbidity of interest (psychiatric-related vs. none, 28.6% vs. 39.3%, OR 0.35 95% CI: 0.27-0.44), longer duration of hospitalization ( >9 d vs. ≤4 d, 18.1% vs. 80.9%, OR 0.05 95% CI: 0.05-0.06), and more recent year of admission (2019 vs. 2017, 49.1% vs. 56.6%, OR 0.76 95% CI: 0.70-0.83). Multicollinearity and significant interactions were not detected in the model.
DISCUSSION
This is the first Canadian study to describe opioid prescribing practices for a generalized population of hospitalized patients. Of all patient admissions, two-thirds were prescribed at least one opioid (65.1%) during hospitalization, and over half had an active opioid prescription on the day of discharge (55.2%). The majority of these were prescribed for as-needed–dosing only and for nearly the entire hospitalization. Strong opioids were prescribed in nearly all cases (96.8%), and two-thirds of all opioid use involved the IV route. The odds of having an active opioid prescription on the day of discharge were the highest for those with active opioid prescriptions lasting throughout their hospital stay and those receiving as-needed doses only.
Overall, opioid utilization estimates in our study were of similar magnitude or higher compared to those from almost a decade ago. For example, in a US study of opioid-naïve medical and surgical patients from 2010 to 2014, 32.7% to 59.1% of the 148,068 patient admissions were administered at least 1 opioid dose. Another US study of a nonsurgical patient population from 2009 to 2010 reported that 51% of the 1.1 million patient admissions received at least 1 opioid claim. A small Canadian study reported an opioid prescription rate of only 22% in opioidnaïve medicine patients. However, important differences must be noted. First, previous studies measured doses administered to patients, whereas our study measured the number of opioids prescribed. Our results include as-needed prescriptions that were available but never actually administered to patients. Second, previous studies typically restricted their study population to opioid-naive patients only. We believe the inclusion of all hospitalized patients in our study is more generalizable to the entire inpatient population. Regardless, our findings demonstrate a very high percentage of hospitalized patients are likely receiving or given the opportunity to receive opioids during their stay. A higher risk of severe opioid-related adverse events has been observed at hospitals with high opioid prescribing rates. In addition, receipt of an opioid as an inpatient has been associated with a two-fold higher relative risk for outpatient use compared with no outpatient use.
The high rate of as-needed inpatient opioid prescriptions in our study suggests some may be prescribed for patients without an indication. Although pre-emptive prescriptions and routine order sets on admission or after surgical procedures have advantages for efficiency and timely responses to patients’ pain, a disadvantage is the risk of over-use in applying to all patients without an adequate assessment of individual patient needs. After an opioid is prescribed to an inpatient, it should be reassessed regularly and discontinued once pain requirements allow. A previous study found that opioid prescriptions lasting more than 75% of a patient’s stay is associated with a higher risk of long-term use compared with those who received an opioid for 1% to 25% of their stay (RRR 1.35% at 365 d).
Automatic stop-order policies have been part of hospital practice for decades and can reduce unnecessary drug use by forcing regular assessments of prescriptions. Essentially, every prescription expires after a fixed number of days, so no further doses can be administered without renewal. The hospitals in our study employed automatic stop dates of 5 days for all opioid prescriptions. However, this policy was often irrelevant since a large percentage of inpatients were discharged before their prescriptions expired. In 2017, an automatic stop-order policy was shown to reduce the duration of antibiotic exposure among infants with very low birth weight. Perhaps the number of active opioid prescriptions on the day of discharge could be reduced by a simple adjustment to the existing policy. In addition, educating patients and other health care professionals on reasonable expectations for pain reductions may further support appropriate prescribing and understanding of the limitations of current drug therapies to minimize unnecessary use.
Nearly half of all patients in our study were prescribed 2 or more different types of opioids and most patients received at least 1 intravenous opioid prescription. Although intravenous opioid formulations are an essential component of pain control for some individuals, they are up to 3 times more potent and faster acting than oral, increasing the risk of adverse events, including accidental overdose or addiction. In addition, prescribing multiple types and routes of opioids poses safety risks from challenges in converting between the prescriptions and discerning the patient’s opioid needs. A survey by the Institute for Safe Medication Practices Canada found “a concerning number of respondents in every discipline … demonstrated knowledge deficit in differentiating between morphine and HYDROmorphone.” Limiting inpatient prescriptions to 1 type of opioid and avoiding IV administration when possible may allow the care team to become familiar with specific agents and support safe and effective pain relief for their patients. Hospital formularies and prescribing policies could play a role in standardizing the approach to routine pain needs.
An encouraging finding was a very high rate of nonopioid analgesic prescribed during the hospital stay (ie, 97% of included admissions), commonly accounted for by either acetaminophen (94.7%) or an NSAID (32.4%). These rates are much higher than reported in previous studies (22.6% to 83%) and appear to align with Choosing Wisely and chronic pain guidelines, which recommend optimizing nonopioid pharmacotherapy before trialing an opioid. These medications can be maximized through standard orders on order sets, which are provided to almost all patients on admission or perioperatively in the studied hospitals. Order sets likely contributed to the high rates of prescriptions observed. However, our study did not capture the actual administration of medications; thus, we cannot confirm the extent to which patients received nonopioid analgesia before an opioid prescription. Regardless, the high frequency of these nonopioid prescriptions is a positive signal, suggesting that care teams and patients are supported in minimizing opioid use.
Over half of patients discharged from medicine units in our study had an inpatient opioid prescription remaining active on their final day of hospitalization, corresponding to 44,305 individuals over a 3-year period. Although a large percentage of these individuals had an as-needed prescription only, those with scheduled doses on the day of discharge were prescribed a mean of 57.3 MME per day. This opioid dose could not be stopped abruptly due to the potential consequences such as withdrawal or loss of pain control; thus, any tapering regimen must be managed by primary care providers who are not always readily accessible. Among patients receiving opioids within 90 days of discharge from a Canadian hospital, opioid initiation appeared to occur during hospitalization in 51.5% of medical patients and 88.2% of surgical patients. This is an important metric suggesting the role of hospitalizations as a significant contributor to outpatient opioid prescriptions and reinforces the need for stewardship to prevent unnecessary continuation at home or harmful abrupt discontinuation.
Strengths and limitations
Our study had several limitations. We were only able to capture inpatient opioid prescribing but could not confirm administration. Therefore, it must be acknowledged that an active opioid prescription on the day of discharge may not increase the risk for outpatient use. That said, we believe our study provides important information on the frequency of inpatient opioid prescribing and the prevalence of default orders for these unique medications. The data was abstracted from 2 hospitals within the same health system and were in the same city. As a result, we cannot be sure they represent the practices across all Canadian hospitals. Important patient characteristics that may affect opioid prescribing, such as opioid exposure before admission, pain severity, opioid indication, and the specific opioid doses administered, were not available. Within our administrative databases, the validity of the studyÕs clinical and health-services variables could not be verified. In addition, dosing information for asneeded prescriptions was often missing (68.3%) and, therefore was not included in the calculation of total daily exposure. Most importantly, the actual number of patients who received an outpatient opioid prescription to be taken at home was unknown.
Several strengths should also be noted. To our knowledge, this is the first Canadian study of a large cohort describing the role of hospital prescribing in the opioid epidemic. The recent study period allows for timely interventions and benchmarking. This study connected data from 2 hospital databases, supporting a deeper understanding of patient and system-related factors. Lastly, the patient population represents a typical overall acute care population regardless of admission type.
CONCLUSION
Inpatient opioid prescriptions are exceedingly common in hospitalized patients, including on the day of discharge. This high rate of opioid prescriptions may be contributing to the existing opioid crisis in Canadian communities, highlighting the need for ongoing opioid stewardship to limit opportunities for unnecessary use. Areas of focus should include prescription durations as well as high rates of intravenous and as-needed prescriptions, which may be tackled through policy development, formulary restrictions, and provider education. Future research is required to determine opioid appropriateness, associate inpatient opioid prescribing with outpatient use, and to evaluate strategies to reduce inpatient opioid prescriptions without negatively impacting patient outcomes.
REFERENCES
1. Canadian Institute for Health Information. Opioid Prescribing in Canada: How Are Practices Changing?; 2019.
2. Alsabbagh, MW, Chang, F, Cooke, M, et al. National trends in population rates of opioid-related mortality, hospitalization and emergency department visits in Canada between 2000 and 2017. A population-based study. Addiction 2021;116:3482–3493.
3. Special Advisory Committee on the Epidemic of Opioid Overdoses. Opioid- and Stimulant-related harms in Canada - Public Health Infobase. Ottawa: Public Health Agency of Canada; 2022. Accessed December 9, 2022. https://health-infobase. canada.ca/substance-related-harms/opioids/graphs
4. Evans-Shields, J, Brummett, CM, England, C, et al. Health Care Burden Associated with Outpatient Opioid Use Following Inpatient or Outpatient Surgery. J Manag Care Spec Pharm 2019;25:973–983.
5. Gomes, T, Khuu, W, Martins, D, et al. Contributions of prescribed and non-prescribed opioids to opioid related deaths: population based cohort study in Ontario, Canada. BMJ 2018;362:3207.
6. Canadian Medical Association. Harms Associated with Opioids and Other Psychoactive Prescription Drugs.; 2015.
7. Statistics Canada. Pain relief medication containing opioids, 2018. Published June 25, 2019. Accessed April 5, 2022.https:// www150.statcan.gc.ca/n1/pub/82-625-x/2019001/article/00008-eng.htm
8. Compton, WM, Jones, CM, Baldwin, GT. Relationship between nonmedical prescription-opioid use and heroin use. New England Journal of Medicine 2016;374:154–163.
9. National Pain Center. The 2017 Canadian Guideline for Opioids for Chronic Non-Cancer Pain; 2017. Accessed September 20, 2020. http://nationalpaincentre.mcmaster.ca/ documents/OpioidGLforCMAJ_01may2017.pdf
10. Dowell, D, Haegerich, T, Chou, R. CDC Guideline for prescribing opioids in chronic pain - United States, 2016. Morbidity and Mortality Weekly Report (MMWR) 2016;65:1–49.
11. 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 2018; 319:872–882.
12. Walensky, RP, Bunnell, R, Layden, J, et al. CDC Clinical Practice Guideline for prescribing opioids for pain - United States, 2022. Morbidity and Mortality Weekly Report 2022; 71:1–95.
13. Herzig, SJ, Rothberg, MB, Cheung, M, et al. Opioid utilization and opioid-related adverse events in nonsurgical patients in US hospitals. J Hosp Med 2014;9:73–81.
14. Donohue, JM, Kennedy, JN, Seymour, CW, et al. Patterns of opioid administration among opioid-naive inpatients and associations with postdischarge opioid use. Ann Intern Med 2019;171:81.
15. World Health Organization. ICD-10 Version:2019. Published 2019. Accessed September 26, 2022. https://icd.who.int/browse10/2019/en
16. Lexicomp. Opioid Conversion Table and Morphine Equivalent Dose Table (Lexi-Drugs). Published online May 19, 2020.
17. Guidelines & Protocols Advisory Committee. Appendix A: Equianalgesic Conversion for Morphine Morphine Equivalence Table (for chronic dosing). Accessed January 23, 2022. https://www2.gov.bc.ca/assets/gov/health/practitioner-pro/bc-guidelines/palliative2_-_pain_equianalgesic.pdf
18. Johnston, R, Jones, K, Manley, D. Confounding and collinearity in regression analysis: a cautionary tale and an alternative procedure, illustrated by studies of British voting behaviour. Qual Quant 2018;52:1957.
19. Regier, L. *CHRONIC NON-CANCER PAIN (CNCP): Overview of Drugs Used in Treatment. RxFiles; 2021. Accessed April 5, 2022. www.RxFiles.ca
20. Lail, S, Sequeira, K, Lieu, J, et al. Prescription of opioids for opioid-naive medical inpatients. Canadian Journal of Hospital Pharmacy 2014;67:337–342.
21. The Coalition for Safe and Effective Pain Management (CSEPM). Reducing the Role of Opioids in Pain Management.
Coalition for Safe and Effective Pain Management; 2019.
22. Garrison, TJ, Puckett, CF. Editorial: the automatic stop order —is it needed? Am J Hosp Pharm 1975;32:994–1000.
23. Pena, V, Guharoy, SR. Impact of a multidisciplinary, collaborative approach to implement and enforce an automatic stop order policy. J Pharm Technol 1992;8:20–22.
24. Tolia, VN, Desai, S, Qin, H, et al. Implementation of an automatic stop order and initial antibiotic exposure in very low birth weight infants. Am J Perinatol 2017;34:105–110.
25. Institute for Safe Medication Practices Canada. Report to Health Canada Hydromorphone Knowledge Assessment Survey Results; 2012. Accessed December 7, 2022. www.ismpcanada.org
26. Canadian Medical Association. 2017 Canadian Opioid Prescribing Guidelines; 2017. Accessed September 20, 2020. https://www.cma.ca/sites/default/files/2018-11/2139_cma_opioid_poster_en.pdf
27. Ciccarello, C, Leber, MB, Leonard, MC, et al. ASHP Guidelines on the pharmacy and therapeutics committee and the formulary system purpose. J Health-Syst Pharm 2021; 78:907–918.
28. Choosing Wisely Canada. Opioid Wisely. Accessed September 21, 2020. https://choosingwiselycanada.org/opioid-wisely/
29. Kurteva, S, Abrahamowicz, M, Gomes, T, et al. Association of opioid consumption profiles after hospitalization with risk of adverse health care events. JAMA Netw Open 2021;4: e218782–e218782.
©Wolters Kluwer Health. View All Articles.
Opioid Prescribing Among Hospitalized Patients in Tertiary Care Hospitals: A Retrospective Cohort Study https://www.qgdigitalpublishing.com/article/Opioid+Prescribing+Among+Hospitalized+Patients+in+Tertiary+Care+Hospitals%3A+A+Retrospective+Cohort+Study/4784216/822886/article.html
