Evaluation of Prescribing Patterns Following Surgical Procedures in Opioid Naïve Patients at a Veterans Affairs Teaching Hospital

ABSTRACT

Objectives

To evaluate facility postoperative opioid prescribing patterns in comparison to published guidelines and adherence to opioid safety mandates.

Methods

This quality analysis was performed between November 2019 and March 2020. Patients were identified to have been opioid naïve prior to receiving a new opioid prescription postoperatively during the study period. Patient charts were reviewed, and patients were contacted to collect desired data. Statistical analysis was performed to evaluate distributions of morphine equivalent daily dose and opioid day supply prescribed across study subpopulations.

Results

Ninety-four of 100 prescriptions evaluated were determined to be within quantity or duration recommendations of the selected guideline. Statistical analysis found no significantly different distributions between the duration and quantity of opioid prescribed at discharge and patient-specific risk factors. Forty-eight patients did not use the entire quantity of the initial opioid prescription dispensed. Of those patients, 26 still had opioids within the home. Opioid risk review documentation was completed in 19 of 65 patients indicated for documentation.

Conclusion

Most opioid prescriptions provided within the study period aligned with recommendations from author-selected guidelines. However, a review of risk prior to opioid prescribing frequently was not performed. The number of patients utilizing less than 50% of prescribed opioids, and few refills indicate that reductions in opioids prescribed would improve safety for both patients and the surrounding community without increasing the risk for the under-treatment of postoperative pain. Improved prescribing habits and patient safety will be targeted through provider education regarding risk review documentation in opioid naïve patients.

INTRODUCTION

Patients using higher opioid doses for longer periods of time are at increased risk of developing an opioid use disorder.¹ Overall, 6-10% of opioid naïve individuals started on opioid therapy will be maintained on opioid therapy 3-6 months after surgery.^2,3 Studies have shown that the rate of progression to long-term opioid use is influenced by the opioid supply provided in the initial prescription. For patients dispensed a 5-day supply of opioids, 10% will go on to long-term use.^2,3 Long-term use is defined as having filled 10 or more prescriptions or more than 120 days’ supply within 1 year following a surgery.⁴ Risk for long-term use increases to 50% when dispensed a 30-day supply.^2,3 Even in patients not progressing to long-term use, there remains the risk of diversion within the community due to overprescribing,⁵ as 70-80% of opioid prescribed at discharge go unused.⁶ It is estimated that 71% of chronic opioid users obtain opioids through methods of diversion.^7,8 In addition to the risk of diversion, extended duration of use after surgery is strongly associated with an increased rate of opioid misuse.⁹ As most misused opioids initially come from a medical source, there is an increased emphasis on reducing quantities prescribed postoperatively.¹⁰

The need for opioid risk evaluation has been highlighted within the literature.^2,3,11 In 2020, a Veterans Health Administration (VHA) study observed significant associations between starting or stopping opioid treatment and increased risks for death from overdose or suicide.¹² Limiting the number of opioids prescribed in order to decrease unnecessary exposure is therefore pivotal in reducing this risk.

While guidelines to direct opioid prescribing in opioid naïve patients exist and have demonstrated reductions in opioid prescribing,^13–15 no guideline has been recognized as standard nationally. With no standardized guidance, many surgeons are uncertain of the appropriate quantity of opioids to prescribe or how much of the prescribed opioids are actually taken by patients.¹⁶ To help providers prescribe appropriate opioid quantities to opioid naïve patients following a surgical procedure, several institutions have developed guidelines to improve prescribing practices.^6,9,17–20 In addition to improved prescribing practices, it is important to provide patient education regarding the proper use, storage, and disposal of opioids. This has been emphasized by The Joint Commission²¹ as a way to improve opioid safety.

There is no standardized guidance for postoperative opioid prescribing at the William S. Middleton Memorial Veterans Affairs Hospital (William S. Middleton VA). In an effort to reduce risks to patients prescribed opioids, a 2017 mandate was put in place requiring prescribers at the William S. Middleton VA to perform and document a review of opioid-associated risks whenever prescribing opioids for an opioid naïve patient (defined by the national VA as no prescribed opioids within the past year). Specific patient characteristics, comorbid conditions, medications, and health care utilization rates have been used to create clinical decision support for clinicians to better quantify risks for opioid adverse effects. The Risk Index for Overdose or Serious Opioid-induced Respiratory Depression (RIOSORD) has been demonstrated to accurately predict the incidence of adverse outcomes with long-term opioid use,²² and when utilized as a part of opioid risk review, has been shown to reduce the number of opioids prescribed in the chronic pain setting.²³ The most common way for providers to assess opioid risk at the William S. Middleton VA is by utilizing the Stratification Tool for Opioid Risk Mitigation (STORM) which calculates risks related to opioid use within 1 and 3 years based on data from the electronic medical record and RIOSORD score. Risk factors tracked by the tool include comorbidities that impact mental health, substance use, chronic diseases, pulmonary risks, and concomitant sedating medications. Access to this tool is available to all providers from within the electronic medical record. Prior to this quality analysis, there had been no formal evaluation of provider adherence to the 2017 mandate or of prescribing patterns across various disciplines.

At the William S. Middleton VA, postoperative opioids are prescribed on a patient-by-patient basis by either the attending physician, surgical residents, physician assistants, or nurse practitioners. Postoperative opioids are not prescribed through the utilization of order-sets. Prior to leaving the facility after surgery, patients prescribed opioids are provided education regarding the appropriate use of opioids, risks of acute and prolonged opioid use, and instructions for proper storage and safe disposal of leftover opioids. For patients presenting for same-day surgery and not formally admitted, education is provided in a written format. Patients formally admitted are provided face-to-face and written education by a pharmacist prior to discharge. At discharge, naloxone dispensing for patients prescribed opioids may be offered on a patient-by-patient basis but is not done as a standardized process. While pharmacy staff ensures prescriptions are reasonable and patient education is provided at the time of discharge, little is known as to how the opioids are used and disposed of after exiting the facility.

The purpose of this quality analysis was to collect postoperative data to inform future directions targeting improved patient safety and opioid stewardship. The primary outcome of this quality analysis was to evaluate facility postoperative opioid prescribing patterns in comparison to published guidelines. As there is no nationally recognized standard, the Washington State Agency Medical Director’s Group (AMDG),²⁴ Johns Hopkins,¹⁹ and the Michigan Opioid Prescribing Engagement Network (OPEN)¹⁸ recommendations for postoperative opioid prescribing in opioid naïve patients were selected for comparison. Secondary outcomes were the evaluation of prescriber adherence rates to national and local VA opioid safety mandates, and the impact of education on patient adherence to appropriate opioid use and disposal.

METHODS

Design

This quality analysis was performed between November 2019 and March 2020.

Part 1: Subject Identification

Patients were initially identified for inclusion in this quality analysis using a facility-level acute opioid prescription dashboard. To be flagged by the dashboard the following criteria had to be met: (1) Outpatient opioid prescription received after the procedure; (2) Procedure within 30-90 days prior to chart review.

Part 2: Confirm Inclusion and Investigate Outcome of Initial Opioid Prescription

Once identified by the dashboard, team members reviewed patient charts at random to determine eligibility and contacted patients via phone to collect the following data: (1) Was the entire opioid supply used? (2) If the entire supply was not used, how many opioid tablets remained at the time of discontinuation? (3) What did patients do with the remaining supply? Consent was assumed if patients responded to survey questions. This project was classified as a quality analysis project and therefore IRB exempt. The departments of pharmacy, surgery, and nursing sponsored this quality analysis. Patients were excluded if any of the following criteria were met: (1) the patient was deceased; (2) the opioid prescription was not intended for the treatment of postoperative pain; (3) the patient was not opioid naïve (defined by the study team as prescribed any opioid for ≥90 days continuously in the past 6 months); (4) patients could not be contacted by phone; or (5) patients could not provide requested data at the time of contact.

Part 3: Full Patient Chart Review and Comparison to Guidelines

Once 100 patients had been contacted and determined to be eligible, the following additional data were collected through patient chart review in order to investigate patterns of opioid prescribing and determine impacts on the quantity or duration of opioids prescribed:

• the specific procedure, prescribing service (distribution can be found in Table III), procedure type (outpatient vs. inpatient), documentation of Opioid Risk Review Note, prescribed morphine equivalent daily dose (MEDD), opioid day supply prescribed, specific opioid prescribed, additional opioid prescriptions dispensed, and patient age. Patient sex was not collected as the selected guidelines for postoperative opioid prescribing do not differ based on sex.

To investigate if comorbidities are known to influence the risk for opioid-induced respiratory depression, the following concurrent diagnoses were collected:

• substance use disorder (opioid, alcohol, other), mood disorder (post-traumatic stress disorder, major depressive disorder, anxiety, bipolar disorder, Attention Deficit Disorder (ADD)/Attention Deficit Hyperactivity Disorder (ADHD), insomnia), chronic pain diagnosis, obesity, obstructive sleep apnea, chronic obstructive pulmonary disease, cardiovascular disease, diabetes, neurologic disease, renal disease, and concomitant benzodiazepine prescription.

Upon completion of the chart review, prescription quantities and durations were evaluated against three guidelines for postoperative opioid prescribing. Guidelines were selected by authors prior to initiation of data collection and included the Washington State Agency Medical Director’s Group (AMDG),²⁴ Johns Hopkins,¹⁹ and the Michigan Opioid Prescribing Engagement Network (OPEN).¹⁸ The aforementioned guidelines were selected based on key features (e.g., publication between 2015 and 2019) and recommendations (e.g., recommendations for procedures pertinent to this analysis, recommendations based on current best clinical and scientific evidence, and expert consensus). The authors acknowledge that additional guidance is available, but further guidelines were not selected as they neither provided recommendations for pertinent procedures, nor did they add additional recommendations to supplement the above-included guidelines. It is important to note that there is overlap and variation in the recommendations among the selected guidelines. As there is no specific guideline nationally recognized as standard, prescriptions were determined to be of appropriate quantity or duration if they aligned with any of the above guidelines.

A total of 303 prescriptions, dispensed between August and December of 2019, were generated by the acute opioid report. These prescriptions were then assessed at random until a preselected target of 100 prescriptions was determined to meet inclusion criteria. A total of 173 prescriptions were assessed to reach this endpoint. Seventy-three prescriptions were excluded for the following reasons: opioids prescribed for a reason other than postoperative pain (14); the patient was deceased (1); the patient could not be reached (53); the patient could not provide requested data at the time of contact (5). The review concluded once data from the available population was gathered from a random sample of 100 prescriptions. The remaining 130 prescriptions were not assessed. Prescriptions were written by orthopedics, podiatry, neurosurgery, cardiac, thoracic, vascular, otolaryngology, plastics, and general surgery providers. Sixty-five of the included prescriptions were provided after outpatient procedures, and 35 were prescribed after procedures requiring formal admission.

Statistical Analysis

Distributions of MEDD and opioid day supply prescribed, were determined using the Wilcoxon Rank-Sum Test across study subpopulations binned by their risk factor status in regard to opioid-induced respiratory depression or overdose. To account for the small sample size and multiple comparisons, a Bonferroni correction was calculated indicating a threshold for significance of P = 0.00128. The remainder of statistics were descriptive in nature.

Results All one-hundred prescriptions were evaluated per AMDG guidelines, with 23 additionally evaluated per Johns Hopkins, and 20 additionally evaluated per OPEN guidelines. Ninety-four prescriptions were determined to be within quantity or duration recommendations of any guideline. Results of prescription evaluations for specific guidelines are presented in Table I. Oxycodone 5 mg tablets were most frequently prescribed at discharge with hydrocodone, tramadol, morphine, hydromorphone, and codeine also prescribed. The average quantity of opioid tablets prescribed at discharge was 23 (range 1-120), and an average of 5 days duration was supplied (range 1-30); the mean prescribed oral morphine equivalent was 45 mg (range 10 mg—180 mg). Statistical analysis found no significantly different distributions between the duration and quantity of opioids prescribed at discharge as stratified by patient-specific risk factors or comorbidities. Additional details of the risk factors associated with opioid use can be found in Table II. Forty-eight (48%) patients did not use the entire quantity of the initial opioid prescription dispensed. Of those patients, 26 (54%) still had opioids within the home. A total of 20 patients received an additional supply of opioids after the initial prescription for postoperative pain. Sixty-five patients that received an opioid for postoperative pain were eligible for a documented opioid risk review note which was completed in 19 of the 65 patients. Additional details of secondary outcomes are presented in Table III.

DISCUSSION

The purpose of this quality analysis was to determine how postoperative opioid prescribing patterns for opioid naïve patients at the William S. Middleton VA aligned with recommendations published by nationally recognized institutions. We found that although there was a wide range of opioid doses and durations prescribed at discharge, greater than 90% of prescriptions were within the recommendations of the referenced guidelines. Interestingly, even though prescribing fell within recommendations for most patients, 35% of patients did not use even 50% of the opioid supply provided at discharge. This suggests that a more individualized prescribing guidance may be needed to reduce the quantity of opioids available for diversion within the community. Hill et al. proposed a guideline to individualize postoperative prescriptions based on inpatient opioid use rather than based on procedure alone and concluded that this method would reduce the total quantity of opioid prescribed.²⁵ Considering the large quantity of opioids that went unused after a surgery, even though prescribed quantities fell within guidelines, patient-specific guidance would be a reasonable consideration in the future, and steps to implement this at the William S. Middleton VA have been taken following the completion of this analysis.

The optimal duration of postoperative opioid therapy is uncertain.¹³ With more focused prescribing guidelines, it is expected that fewer opioids would be prescribed^6,13,26–31 and the risk of undertreating postoperative pain would increase. However, our results are promising in that reduction in opioid quantity and duration prescribed may be unlikely to undertreat pain based on the quantity of unused opioids in combination with the low number of patients requiring additional prescriptions. Only 20% of patients evaluated here received refills. This is consistent with evidence that prescriptions aligning with the selected guidelines will provide adequate analgesia without the need for refills in 75% of patients.^18,24 Supporting this conclusion are previous studies that documented no change in refill patterns or patient satisfaction after implementation of guidelines which reduced prescribed opioid quantities by > 50%.^27,32

Another finding of this study is the under-documentation of opioid risk reviews. Prescribers may review STORM, or perform opioid risk review internally, however, this was not always documented as required by National VA mandates. The results showed no significant differences between opioid quantities prescribed to patients with risk factors for an opioid-related adverse event, such as respiratory depression or overdose, than those without risk factors. This highlights the importance of increased opioid stewardship and performance (and documentation) of risk review prior to prescribing. Although this quality analysis did not follow patients for an adequate duration to determine how many went on to long-term opioid use, it does identify a need for prescriber education regarding documentation of opioid risk reviews to help prevent progression to long-term opioid use and to improve patient safety.³⁰

Since the completion of this quality analysis, the William S. Middleton VA has made changes to improve postoperative opioid prescribing. Notably, the facility hired a pharmacist to work in the perioperative setting to meet with patients prior to surgery and develop a customized postoperative opioid prescribing plan and set pain management expectations. The pharmacist also ensures that any needed monitoring is completed and communicates recommendations to the surgical team. Ongoing education to the nurse practitioners, who complete the majority of preoperative appointments, has also been implemented. These interventions have not yet been evaluated but have been well received by the surgical teams and patients. The William S. Middleton VA is currently up-to-date with mandated opioid monitoring.

LIMITATIONS

There are several limitations to our analysis. In addition to the small sample size, this is a single institution, non-randomized, retrospective study conducted over 5 months with no long-term follow-up. Data collected regarding the use and disposal of opioids by patients after discharge were based on patient reports and verified when available, through review of prescription fill histories. This limitation is consistent with the previous studies^3,17 and authors believe that although not ideal, a more accurate method for collecting this data is unknown. Lastly, patients were not followed for a duration adequate to determine if patients went on to long-term use as has been defined in previous studies as 1 year after the initial prescription.^2–4

CONCLUSION

Most opioid prescriptions provided within the study period aligned with recommendations from author-selected guidelines. However, a review of risk prior to opioid prescribing in many patients was not documented. Prescribers likely did perform risk reviews internally, however, the results showed no significant differences between opioid quantities prescribed to patients with different risk factors for an opioid-related adverse event, such as respiratory depression or overdose, than to those without risk factors. Furthermore, the number of patients using less than 50% of prescribed opioids, and relatively low number of refills indicate that reductions in opioids prescribed would improve safety for both patients and the surrounding community without increasing the risk for the under-treatment of postoperative pain. These results indicate that the logical starting point for improving prescribing and patient safety is through the development of provider education targeted at increasing documentation of opioid risk review prior to prescribing opioids in opioid naïve patients. As a result, new guidance has subsequently been implemented at the William S. Middleton VA.

ACKNOWLEDGMENTS

None.

FUNDING

None.

CONFLICT OF INTEREST STATEMENT

None declared.

REFERENCES

Author notes