Article Text
PDF
Abstract
Background Excess opioid prescribing after surgery can lead to prolonged opioid use and diversion. We interviewed surgeons who were part of a three-group cluster-randomised controlled trial aimed at reducing prescribed opioid quantities after surgery via two versions of a monthly emailed behavioural ‘nudge’ (messages encouraging but not mandating compliance with social norms and clinical guidelines around prescribing) at the end of the implementation year in order to understand surgeons’ reasoning for changing or continuing their prescribing behaviour as a result of the intervention and the context for their rationale.
Methods The study took place at a large healthcare system in northern California with surgeons from three surgical specialties—orthopaedics, obstetrics/gynaecology and general surgery. Following the intervention period, we conducted semistructured interviews with 36 surgeons who had participated in the trial, ensuring representation across trial arm, specialty and changes in prescribing quantities over the year. Interviews focused on reactions to the nudges, impacts of the nudges on prescribing behaviours and other factors impacting prescribing. Three study team members coded and analysed the transcribed interviews.
Results Nudges were equally effective in reducing postsurgical opioid prescribing across surgical specialties and between intervention arms. Surgeons were generally receptive to the nudge intervention, noting that it reduced the size of their discharge opioid prescriptions by improving their awareness and intentionality around prescribing. Most were unaware that clinical guidelines around opioid prescribing existed. Some had reservations regarding the accuracy and context of information provided in the nudges, the prescription quantities encouraged by the nudges and feelings of being watched or admonished. A few described discussing the nudges with colleagues. Respondents emphasised that the prescribing behaviours are informed by individual clinical experience and patient-related and procedure-related factors.
Conclusions Surgeons were open to learning about their prescribing behaviour through comparisons to guidelines or peer behaviour and incorporating this feedback as one of several factors that guide discharge opioid prescribing. Increasing awareness of clinical guidelines around opioid prescribing is important for curbing postsurgical opioid overprescribing.
Trial registration number NCT05070338.
- Surgery
- Randomised controlled trial
- Decision making
Data availability statement
Data are available upon reasonable request. We are able to provide (1) deidentified demographics (without site), but no contact information; (2) interview protocols; and (3) the qualitative analysis codebook. Reuse is permitted for similar qualitative studies with appropriate citation. Data are available from Meghan Martinez (meghan.martinez@sutterhealth.org).
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Previous research has shown interventions aimed at reducing postsurgical opioid prescribing, a contributor to the ongoing opioid epidemic, to be successful, yet few have qualitatively examined surgeons’ feelings about the interventions and what factors influence the decision to reduce (or not) prescribed opioid quantities.
WHAT THIS STUDY ADDS
Our study sought to understand how surgeons in three surgical service lines in one healthcare hospital system responded to monthly email reports comparing opioid prescribing behaviours to either peers or guidelines, and other considerations at play in postsurgical prescribing.
Both peer comparison and guideline comparison reports were equally effective and generally acceptable across all three service lines through increasing awareness around prescribing habits; however, guidelines and peer behaviours with respect to opioid prescribing are only one of many factors surgeons consider when prescribing opioids.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
In general, awareness of opioid prescribing guidelines for postsurgical discharge among surgeons was low, and more education on appropriate prescribing should be shared with postsurgical care teams.
Administrators need to understand that while surgeons are open to learning about their prescribing behaviour and incorporating this feedback as one of a suite of factors that guide discharge opioid prescribing, surgeons must be made aware of and understand the evidence base for the guidelines and be involved in development of the intervention.
Background
The US continues to experience a crisis of opioid-related morbidity and mortality, with unused opioids prescribed by healthcare providers playing a major role in access to opioids.1 Among healthcare providers, surgeons and others involved in postsurgical care are responsible for approximately 10%–20% of all opioid prescriptions in the US.2–5 Overuse and diversion of unused opioids may contribute to the opioid epidemic,6 leading some to speculate as to whether surgeons are the ‘gatekeepers of the opioid crisis’.7 Importantly, over half of postoperative opioids prescribed by US surgical providers are reportedly never used by the patient, making many prescriptions excessive.8 Taken together, this suggests that reducing surgical providers’ opioid prescribing is an important step in addressing the US opioid crisis.
In recent years, much work has been done to understand surgeons’ prescribing behaviours and develop interventions to reduce overprescribing of opioids. This body of work has illuminated several system and individual-level barriers to reducing postsurgical opioid prescribing,9 including inadequate time for counselling on opioids, the expense of offering non-opioid alternatives (eg, intravenous acetaminophen), worries over patient satisfaction scores and providers’ emotional responses to patients’ pain.9 It has also described several interventions that have successfully reduced postsurgical opioid prescribing.10 However, there is little information as to how successful interventions have managed to overcome these barriers to prescribing reduction and how surgeons have perceived these interventions. To our knowledge, only three studies on postsurgical discharge opioid prescribing interventions have explored surgeons’ reactions, finding that surgeons were generally supportive of the interventions but had concerns about adequate pain management and the interventions’ appropriateness.11–13
In this study, we sought to build on this literature by providing insights into factors driving postsurgery opioid prescribing behaviours and responses to a prescribing reduction intervention implemented in a randomised controlled trial across a large healthcare system. Our intervention was not only effective in reducing prescribing14 but also unique in comprising two versions of behavioural ‘nudges’, which use social norms to guide individuals towards reducing postsurgical opioid prescribing without impinging on autonomy and have been shown to scale easily.15 Because our intervention is promising in its effectiveness and scalability, understanding reasons why it transformed surgeon prescribing behaviour (or not) as well as its acceptability is particularly important, and whether there were any differences by type of nudge or surgical specialty. We conducted qualitative interviews to investigate surgeons’ reasoning for changing their prescribing behaviour or not as a result of our intervention and the context for this reasoning after implementing the nudges for 1 year.
Methods
Setting and intervention context
This study was conducted at Sutter Health, a large, multispecialty delivery system in northern and central California that serves approximately 3.5 million patients. The qualitative interviews were part of a larger study described in detail in Kirkegaard et. al. 2022.16 Nineteen Sutter Health hospitals and their associated surgeons from 3 surgical specialties (general, obstetric/gynecologic and orthopaedic) were included in the study, which was a three-arm cluster-randomised controlled trial of two behavioural nudges compared with usual postsurgical care (control). One nudge compared opioid prescribing behaviour relative to guidelines endorsed by Sutter Health (guideline arm), while the other provided feedback on prescribing behaviour relative to peers in the same specialty (peer comparison arm). During the intervention period of October 2021–October 2022, nudges were emailed monthly to surgeons who were outside of recommended ranges at least two times in the previous month. Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.
Recruitment
We recruited surgeons for interviews through a multistep process. First, we reached out to hospital leadership (eg, chief medical officers) to ensure their buy-in and encourage participation among any contacted surgeons. Next, we followed a purposive sampling strategy intended to elucidate (1) why some surgeons with high baseline prescribing improved their behaviour after receiving our intervention, (2) why other surgeons with high baseline prescribing did not change their behaviour, (3) why some surgeons who did not receive our intervention still decreased their prescribing and (4) interest, concerns and confusion regarding the intervention. We aimed to achieve distribution across the three surgical specialties and intervention arms so that we could detect any differences in prescribing behaviour and reactions to the nudge interventions across these dimensions. In order to recruit a sufficient number of surgeons to achieve thematic saturation, we recruited (1) surgeons who received at least one nudge email during the study period, oversampling those who had high prescribing quantities at baseline, (2) anyone in the control group who decreased their prescribing from baseline or (3) anyone who had actively reached out to the study team about the nudge emails with questions or feedback on receiving them. We invited 245 surgeons, emailing them up to three times.
Interview protocol
We developed a semistructured interview guide to elicit surgeons’ reactions to the nudge interventions, including what aspects they liked and disliked, whether they found the information presented novel, what they thought about the guidelines or peer comparisons presented, whether they discussed the nudges with colleagues, whether the nudges prompted them to reconsider their prescribing habits and whether they would want to continue receiving similar information in the future. To contextualise these reactions, we also included questions about surgeons’ general opioid prescribing workflows, awareness of clinical guidelines and any factors that may have impacted their opioid prescribing over the past 5 years, including any disruptions to care delivery from the COVID-19 pandemic. Surgeons in our control arm were only asked these contextual questions.
In order to achieve thematic saturation, we conducted 36 interviews (response rate=15%). Each interview lasted between 25 and 60 min and all were conducted by MM between September 2022 and January 2023. Interviews were conducted over the phone and were audio-recorded and transcribed for accuracy. Providers received US$100 in compensation for their time.
Analysis
Transcripts were managed and analysed in Dedoose V.9.0.17 We created an initial codebook deductively based on the research questions, then inductively added subthemes and additional codes as novel concepts emerged. Using a thematic analysis process,18 MM and KB both reviewed and coded three transcripts as the reference coders, then met to refine the codes and address discrepancies. The remaining 33 transcripts were divided for coding among MM, KB and AK. The coding team met regularly to review documented points of ambiguity and reach consensus. Demographic characteristics for respondents were obtained from administrative data, where available.
In our analysis, we grouped surgeons into three categories based on the difference between each surgeon’s percentage of eligible procedures with discharge opioid prescriptions within recommended quantity ranges during the intervention year and during the year prior to the intervention. Greater than 20 percentage points was defined as a large reduction, 5–20 points as a medium reduction and less than 5 points as a small reduction or increase (ie, did not change prescribing behaviours).
Results
We interviewed 36 surgeons—15 general surgeons, 13 obstetric/gynecologic surgeons and 8 orthopaedic surgeons. Among the 36 participants, 21 participants (58.3%) were men, almost two-thirds had been in practice between 10 and 29 years and the majority worked in urban hospital settings, which is consistent with the demographics of our overall study sample (table 1).
Study sample
Reactions to the nudge interventions and reasoning given for changing or not changing prescribing
Surgeons’ reactions to the emailed reports varied widely, with no overarching trends by surgical specialty or intervention arm or by type of nudge (table 2). Five main themes emerged in relation to surgeons’ reactions: improved awareness and intentionality with prescribing, discussions with colleagues, questions about the information presented in reports and whether a change was necessary, doubts about the recommended amounts and need for refills, and feelings of being watched or admonished.
Reactions to the nudge interventions and reasoning given for changing or not changing prescribing
Improved awareness and intentionality with prescribing
Many surgeons spoke positively about the reports and articulated finding the information to be very useful in increasing their intentionality and awareness of their prescribing habits, resulting in moderate to large reductions in prescribing across surgical specialties. Some surgeons who received the nudge based on peer comparison mentioned how surgeons tend to be competitive by nature and being compared with colleagues spurred a reduction. Among those who decreased their quantities, surgeons were at times surprised that their patients tolerated the decreased opioid quantities without any issues or known requests for refills.
Discussions with colleagues
Surgeons mentioned how the intervention increased discussions with colleagues, sometimes in more positive ways than others. An obstetric/gynecologic surgeon (peer comparison arm) who moderately decreased their prescribing noted that the peer comparison spurred discussion within their immediate care team in which the group decided to set their own prescribing standards. In contrast, a general surgeon (guideline arm) who did not decrease their prescribing described how other surgeons had suggested ignoring the guidelines altogether.
Questions about the information presented in reports and whether a change was necessary
Surgeons in both intervention arms and across service lines noted that they wanted more information and context in the nudge emails, including who their ‘peers’ were, how comparisons were calculated and whether the guidelines accounted for patient population differences. However, two obstetric/gynecologic surgeons (peer comparison arm) who felt the emails were annoying and useless without this additional information nevertheless had moderate to large decreases in their prescribing. Several surgeons mentioned that they had already decreased their prescription quantities on account of the opioid epidemic and they did not believe they needed to reduce their prescribing any further. In some instances, the surgeon performing the procedure was not the one in charge of prescribing discharge medications. For those cases, the nudge emails sometimes resulted in the surgeon discussing opioid prescribing with the rest of the care team or ignoring the emails entirely (“I just figured it was because other people were discharging my patients…so after that, I just started ignoring them” (general, peer comparison arm, large decrease)).
Doubts about recommended amounts and the need for refills
For some surgeons, the recommended quantities seemed unreasonably low, even if they still reduced their prescribing. For example, one obstetric/gynecologic surgeon (peer comparison arm) was surprised that sending ‘2 people with 10 pills each’ was ‘overdoing it’. Another surgeon (orthopaedic, guideline arm) felt that the difference between what they were prescribing and the ‘appropriate’ range provided was minimal or insignificant (eg, two pills), yet they nevertheless had a small decrease in prescribing quantities. One surgeon (orthopaedic, guideline arm, large decrease) noted that they believed patients would still get a refill and ultimately would end up with the same number of pills, but they were willing to comply with the reduced initial quantities.
Feelings of being watched or admonished
One general surgeon (guideline arm) who did not decrease their prescribing felt like they were being lectured on how they practice without being properly informed on what expectations were. An orthopaedic surgeon (guideline arm) who also did not decrease reported feeling like they were being watched and felt frustrated by not being able to find more information on the guidelines. For a few surgeons, the nudge emails led them to fear that their current prescribing behaviour might incur repercussions either from the organisation or from state or federal monitoring boards. One surgeon (guideline arm) specifically mentioned the possibility of incurring a statement of complaint related to underprescribing and suggested that it would be helpful for Sutter Health to protect surgeons against such complaints, explaining:
I would like to also be cleared of not having to respond to any sort of statement of complaint because of it…I’ve already received a couple in my career. If I get another one…that’s more work for me to deal with that than it is to not read an email…it actually reflects on my record.
Contextual factors surrounding surgeons’ opioid prescribing
Our interviews revealed that several contextual factors impacted surgeons’ prescribing behaviours independent of our nudge intervention (table 3). When discussing these contextual factors, five main themes emerged: procedure-related considerations, patient characteristics, patient preferences regarding opioids, patient access to refills and the opioid epidemic.
Contextual factors surrounding surgeons’ opioid prescribing
Procedure-related considerations
The most commonly mentioned factor that respondents consider when prescribing opioids was the type of procedure, including the approach (eg, laparoscopic vs open), urgency (eg, emergent vs elective), location (eg, knee vs shoulder), availability of additional pain support (eg, nerve blocks) and hospital length of stay. Some procedure-related considerations were specific to surgeons’ specialties and involved the circumstances surrounding their patients’ recovery. For example, obstetric/gynecologic surgeons noted that breast feeding was often a reason to limit an opioid prescription, though patients who had just given birth also needed sufficient opioids to control their pain so that they could care for their newborns. Similarly, the exacerbation of pelvic surgery discomfort by opioid-related constipation was also given as a reason to limit opioids. Orthopaedic surgeons underscored that their procedures are often rather painful and that patients need to be able to move in order to initiate physical recovery and prevent blood clots, thus warranting more pain medication.
For patients who had an inpatient procedure, some respondents described being able to calibrate discharge opioid prescriptions based on the quantity of opioids the patient had required in the hospital. However, for same day and outpatient surgeries, surgeons typically wrote prescriptions ahead of time during the preoperative visit and determined the quantity to prescribe based on their expectations of the patient’s needs or a default quantity they had set in the electronic health record.
Patient characteristics
Several respondents noted that they take the patient’s age into account when prescribing, with older patients requiring fewer opioids than younger patients. Weight, smoking status, pre-existing pain conditions, intolerance of certain opioids or non-opioid pain medications and pre-existing use of opioids or other substances (eg, cannabidiol) were also mentioned as considerations, though less frequently and with less clear implications for prescribing. Some surgeons described forming an overall subjective perception of a patient’s pain tolerance and pain management needs, with one general surgeon remarking that “sometimes, it’s, you know, honestly, kind of an eyeball test in the office. If someone seems like, yeah, they're pretty tough, I don't think they’ll need as much.”
Patient preferences regarding opioids
Some surgeons mentioned encountering patients who either requested a specific opioid (eg, tramadol rather than oxycodone) due to past experiences or were uncomfortable taking any opioids due to their risks. Most surgeons who mentioned requests for specific opioids said that they accommodated the requests. For patients who were uncomfortable with taking any opioids, surgeons responded in one of three ways: allowing the patient to go home without medication, but telling them to call if they changed their mind; encouraging patients to take at least enough for a day or two; and writing the prescription so patients had it available, but noting that they did not have to fill it. Surgeons generally felt it was better to have some opioids on hand and not take them rather than need some later and not be able to get them.
Patient access to refills
Surgeons considered the burden of refills when prescribing, particularly the day of the week, because they did not want other providers to have to refill opioids during the weekend for their patients. They also did not want their patients to be in a position where they needed more opioids but lacked access to a pharmacy (eg, they lived far away from a pharmacy or their pharmacy did not provide opioids).
Impact of the opioid epidemic on prescribing workflows
Beyond surgeons’ general cognizance of the opioid epidemic, including coming to see narcotics as sometimes harmful and knowing people in their own networks who had become addicted to opioids, some mentioned specific ways in which systems-level measures taken against the opioid epidemic impacted their prescribing. Examples included issues pertaining to insurance requiring prior authorisation for opioid prescriptions, limits placed by pharmacies on how many opioid pills they will dispense, and state-level requirements to check the prescription drug monitoring programme database when prescribing more than a certain amount. However, some surgeons did not see the opioid epidemic as relevant for their prescribing, such as an obstetric/gynecologic surgeon who remarked, “I thankfully don’t have an issue with patients and opioid addiction, so I think at the end of the day I kind of was like, ‘I don’t think this applies to me.’” Others noted the ‘swinging pendulum’ or shifting norms of opioid prescribing and pain management during the course of their tenure. Multiple surgeons noted how quantities have been decreasing over the years as we move away from “pain as the fifth vital sign”—“you would just give them out like candy” back in 2007 and “I think it used to be 90, and then it was like 60, and now we’re down to 42 [pills]….”—though another surgeon expressed concerns over how “maybe it’s sort of gotten a little too restrictive.”
Discussion
As part of a larger study testing two behavioural nudges on postsurgical opioid prescribing across three surgical specialties (general, obstetrics/gynaecology and orthopaedics), we aimed to explore surgeons’ reactions to these nudges and reasons for why this intervention was or was not effective for a given surgeon. We found that the nudges were effective in reducing opioid prescribing, and there were no clear or consistent differences by surgical specialty or intervention group. Several interviews revealed that simply being made aware of prescribing more than peers or clinical guidelines was a catalyst for reducing the number of pills prescribed. However, we found the acceptability of the intervention to be impacted by a number of factors. Surgeons expressed scepticism of the accuracy and context of the information in the nudges, wanting more transparency around where the guidelines came from and how they were developed, who the comparison peers were and whether patients could handle fewer opioids at discharge without increasing refill requests. They also emphasised that their prescribing behaviour is informed by individual clinical experience and that they will adjust prescriptions according to the circumstances of the patient and procedure rather than uniformly follow clinical guidelines or the behaviours of their peers.
That surgeons generally were supportive of the intervention and reduced opioid prescribing as a result echoes similar work done in other hospital settings where clinicians found feedback via email about their opioid prescribing behaviour useful, particularly comparative prescribing data with their peers.19 However, in our study, surgeons felt that the notion of ‘peers’ was not concretely explained or defined, and thus was not a worthwhile comparator. Because guidelines are simpler to explain and define than peer comparisons and our guideline-based nudges were equally effective as our peer comparison-based nudges across surgical specialties, future interventions may benefit from relying on guidelines. However, surgeons still need to understand the evidence base for the guidelines and ideally be involved in the development of the intervention.11
These in-depth interviews also demonstrate the myriad factors that come to bear on the amount of opioids prescribed to patients following a procedure. With regard to opioid prescribing, surgeons are in the unique position of performing procedures that are objectively painful. Our participants repeatedly noted the important role opioids can play in postsurgical recovery, but also acknowledged the desire to not overprescribe. A growing body of evidence suggests that pain can be managed with few, if any, opioids,20–26 and many of our respondents noted that they have already cut back substantially on opioids by incorporating multimodal approaches to pain management. Some participants expressed concern that patients receiving fewer initial opioids would just need additional refills, but here evidence suggests that this not to be true.27–31 As not all procedures have associated guideline-recommended prescribing ranges, clinical leadership should ensure that surgeons clearly set expectations with patients, as many of our surgeons reported doing, and make surgeons aware of the literature around pain management without opioids and the need for refills. Furthermore, while it has been suggested that surgeons are ‘gatekeepers’ to the opioid crisis, some surgeons expressed scepticism or disbelief that their patient population suffered from any opioid addictions, highlighting the importance of personal and collective responsibility among the medical community in understanding the extent of the epidemic. Wider education campaigns around the role that excess opioids resulting from postsurgical prescriptions play in the opioid crisis would improve consistency across clinical staff prescribing.
This study has limitations. First, the intervention was conducted in a single healthcare system in northern California that may not be representative of other systems throughout the country. However, the organisation does cover a wide geographic region including urban, suburban and rural areas with a diverse patient population and all insurance types. Second, during our interviews, we did not reference the amounts by which surgeons had changed their prescriptions during the intervention year. Though we saw this as important for maintaining rapport and for surgeons to not feel as though they were being criticised for their prescribing behaviours, it also prevented us from discussing with surgeons how their perceptions of their prescribing changes aligned with their actual prescribing changes.
Conclusions
This qualitative research explored surgeons’ reactions to a nudge-based intervention targeting excessive postsurgical opioid prescribing across three surgical specialties. It serves as a complement to the results observed following a 12-month nudge-based intervention to curb opioid overprescribing. The findings suggest that surgeons are open to learning about their prescribing behaviour and incorporating this feedback as one of a suite of factors that guide discharge opioid prescribing.
Data availability statement
Data are available upon reasonable request. We are able to provide (1) deidentified demographics (without site), but no contact information; (2) interview protocols; and (3) the qualitative analysis codebook. Reuse is permitted for similar qualitative studies with appropriate citation. Data are available from Meghan Martinez (meghan.martinez@sutterhealth.org).
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by Sutter Health Institutional Review Board and RAND Corporation’s Human Subject Protection Committee (study number 2018-0988-CR05). Participants gave informed consent to participate in the study before taking part.
References
Footnotes
Contributors Study conception and design: MM, KB, XSY and KEW. Coordination and implementation of the study: MM, AK, KB, ZW and XSY. Data collection: MM. Data management, analysis and validation: MM, AK and KB. Drafted manuscript: MM, AK and KB. All authors read, edited and approved the final manuscript. MM accepts full responsibility for the work and had access to the data. MM, KB, and KEW jointly controlled the decision to publish. We would like to thank all the surgeons who took the time to contribute to this work.
Funding This study was funded by National Institute on Drug Abuse (R01DA046226).
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.