Patterns of opioid use for lower limb trauma patients during the first 6 months after discharge

Context

This study forms part of a programme of work to understand how opioid-naïve patients use opioids within the first few months of prescription. Opioid-naïve patients are people who have not received opioids in the 30 days before an acute event or injury (Pino et al, 2024). The results will direct practice and inform further proposed study regarding intervention development in relation to clinician education and discharge planning. Research ethics committee (19/LO/0549) and Health Research Authority (251431) approvals were granted for the study in 2019, which was conducted in line with the R&D procedures of the hospital. The study has been presented to and discussed with a patient and public involvement group and deemed highly relevant to the general population.

Aims

The research question was: How do patients admitted with lower limb injuries continue to use opioid analgesics after they are discharged from hospital?

The aims of the study were:

Design

Recruitment, consent, and data collection occurred at one major trauma centre within a large teaching hospital (November 2019 to September 2020). The participants' medical records were screened by the researcher, a nurse specialist from the pain service, during a direct care visit. Adults with leg injuries admitted to the trauma wards were asked to participate in this study if they had been opioid naïve before their admission. The researcher asked all eligible people if they would like to be involved, provided written and verbal information, and took written consent. Reflexivity was considered throughout the study as the researcher's professional experience influenced the data collection and analysis. Confidentiality was guaranteed. Purposeful sampling guided the selection and was drawn from the population of adult patients admitted to hospital with acute pain following a leg injury. The sample is representative of all patients admitted with leg injury and incorporates variables that influence opioid use. Ten people declined involvement in the study, 30 people were recruited. Initial data were collected as soon as consent was given.

The participants completed a short questionnaire about their expectations concerning the injury and how long they expected to take analgesia. They were asked to record whether their pain had increased or decreased during their hospital stay, rather than using a more formal pain score, because it was not possible to record pain scores after discharge. The questionnaire was completed on the ward and collected by the nurse specialist. An itemised checklist, created for the study, was completed by the nurse specialist on known factors that may indicate a propensity to long-term or problematic opioid use, which included diagnosed psychological/mental health issues, a person or family history of substance misuse or alcohol use, increasing analgesia at the point of discharge, pain in more than three sites and the presence of significant pain in response to non-painful stimuli (allodynia). The nature of the injury and opioid medication use while an inpatient were recorded. The researcher reviewed GP and hospital care records 6 months after discharge to record discharge medication, observe whether opioids were still being used and how often the patients were reviewed by the GP practice or hospital. Opioid conversion was not done in this study because the aim was to see whether opioids were being used commonly after discharge, not to describe the opioid ‘burden’. Descriptive statistics were used to summarise the basic characteristics of the data due to the small purposeful sample size.

Results

The participants in this study were all 18 years of age or older and able to consent. They had experienced leg injuries including at least one of the following: femoral fractures, tibia and fibula fractures, ankle fractures, swollen and infected knees, and patella tendon injury. Thirty patients who met the inclusion criteria were enrolled in the study, three were lost to follow-up. The study sample included 15 men and 15 women. Mean age was 50 years (range: 19-81). The following mechanisms of injury were represented: fall (17), road traffic accident (3), sport (4), infection (3), pothole (1), jump from height (1), hit by an object (1). Twenty-five people had injuries that required surgery. The mean length of hospital stay was 17 days (range: 2–100 days). At the time of the injury 20 people were employed, unemployed or engaged in study, and 10 were retired. All participants were discharged directly home.

Thirty seven per cent (11/30) of patients were taking simple non-opioid analgesia, for example paracetamol, before admission. All patients were prescribed immediate-release (IR) morphine during their hospital stay; every patient prescribed codeine also had IR morphine prescribed (Table 1).

Half (15/30) of the patients were independent in self-care needs. Those people who were dependent had tibia and ankle fractures in the main; this group included more women with an older average age of 66 years (10/15). All patients engaged with physiotherapy and rehabilitation. Three people described having psychological issues pre-admission; no patients identified having a personal or family history of substance misuse. Ten per cent (3/30) of patients declared a history of personal alcohol misuse.

Eighty per cent (24/30) of patients reported that their pain had decreased during their hospital stay; only one person felt that their pain had increased during their admission. Sixty seven per cent (8/12) of women with decreasing pain had reduced their opioid consumption over the course of their hospital stay, compared with 17% (2/12) of men. Fifty seven per cent (17/30) of patients expected to use analgesia for up to 2 months, 40% (12/30) were unsure, whereas one patient expected to be using analgesia forever. Twenty three per cent (7/30) of patients experienced allodynia, a term used to describe the phenomenon when non-painful stimuli are felt as painful; this was not related to a specific site or mechanism of injury, and predominantly occurred in women (5/7).

Before admission nearly half of the patients had not been in formal employment. Of those working, 56% (9/16) expected to return to work within 12 weeks of injury; 13% (2/16) were less specific and thought it would take a few months; and 31% (5/16) were unsure. Most people could estimate healing time and return to activity, with 63% (19/30) expecting to heal fully by 10 weeks from injury and return to normal activity by 12 weeks. Four people expected to take up to 1 year to heal but anticipated that it would take less than 6 months to return to normal activity.

Seventy seven per cent (23/30) of patients were discharged with an opioid in their prescription (Table 2) which included 93% (14/15) of male participants. Codeine and IR morphine were the preferred discharge opioids, which mirrors the inpatient prescriptions, and 30% (7/23) received both on discharge.

Analgesia prescribed in the 6 months after discharge included a broader range: codeine, tramadol, IR morphine or slow-release (SR) morphine, and IR/SR oxycodone (Table 3). Data were missing for three people who were lost to follow-up. Twenty seven per cent (8/30) of people who received opioid on discharge from the hospital did not receive any further prescription from their GP; 17% (5/30) had one prescription for opioid issued by their GP but did not go on to receive repeat prescriptions; and 47% (14/30) received repeat prescriptions for opioid that were still active at the time the records were reviewed at 6 months post discharge. Fifty per cent (7/14) of people on repeat prescription had been reducing opioid use while an inpatient, with most having stated that the pain was the same or had decreased prior to discharge. Of this group one person had a history of alcohol misuse, and one person had had psychological issues before admission.

Almost half of the people had access to codeine following discharge and 69% (9/13) of this cohort continued with repeat prescriptions over the first 6 months following discharge from hospital. Two people continued to use both codeine and IR morphine; both drugs had been initiated in hospital and were continued on repeat over 6 months. Three people were prescribed tramadol by their GP; one person continued this on repeat prescription; and one person progressed from tramadol to SR morphine. Repeat prescriptions of IR morphine were given to 20% (6/30) of people; two people were prescribed SR morphine by their GP, one of whom had been discharged with it. Following discharge one person was prescribed IR and SR oxycodone on repeat by their GP (Table 4).

In the 6 months following discharge from hospital 33% (9/27) attended fewer than four appointments with their GP practice, 52% (14/27) attended between four and ten appointments, with 15% (4/27) attending more than ten appointments. Forty eight per cent (13/27) attended up to two hospital review appointments with their orthopaedic team, 44% (12/27) attended between three and six appointments, and 7% (2/27) attended up to ten appointments. Those individuals who attended hospital appointments more than three times had also visited their GP practice an average of seven times.

The people attending more than four GP appointments and more than two hospital appointments included all of those with pain in more than three sites; two of the three people who had psychological issues; and 86% (6/7) of those with allodynia. This group also included most people who were unsure about healing time, duration of analgesia and return to normal activity. However, at the point of discharge from hospital 95% (18/19) of the frequently attending patients had decreasing pain.

It was not deemed appropriate to ask trauma patients about adverse childhood experiences without immediate access to psychological therapy, and for this reason an opioid risk tool was not included in the study. However, the researcher was able to undertake a hypothetical analysis using the Opioid Risk Tool (ORT) (Webster and Webster, 2005) from the data collected and speculated that 77% (23/30) of patients would potentially be classified as low risk.

Discussion

Codeine and morphine were prescribed predominantly, with only a small proportion of people requiring an alternative opioid. This is not unexpected because pain is a known consequence of injury, and orthopaedic traumas are among the most painful (Jones et al, 2017). In studies of patients with traumatic orthopaedic injuries, 59% reported moderate to severe pain at the time of discharge from hospital (Archer et al, 2012), with 50% of patients requiring opioid medication on discharge (Chaudhary et al, 2017). Effective analgesia is fundamental to trauma care, and opioids continue to play an important part in treating acute pain.

Most people reported that their pain had decreased during their hospital stay, and almost half were able to reduce their opioid consumption before discharge home. Three-quarters of participants were discharged with opioid medication, and more than one-quarter were discharged with two opioids. The evidence supports the use of opioid for acute pain, tapering over days to a few weeks (Faculty of Pain Medicine, 2024), with weaning beginning as soon as function is recovering, and pain is subsiding (Hegman et al, 2014b; Quinlan et al, 2019). However, nearly half the patients continued to receive repeat prescriptions for opioids from their GP in the 6 months following discharge, including the group of people who had reduced their opioid use in hospital and declared that their pain had reduced in severity.

Persistent postoperative opioid use (PPOU) is a term coined to describe the situation when individuals continue use of prescribed opioids beyond 3 months after injury or surgery. It has previously been reported that 4-8% of opioid-naïve patients following total knee or total hip replacement surgery were still using opioids at 6 months (Goesling et al, 2016). One explanation for the substantial difference in figures between this study (47%) and the Goesling et al (2016) study could be that the injuries were related to trauma rather than elective surgery. Risk factors for PPOU include an over-reliance on opioid analgesia, unlimited amounts of opioid dispensed and automatic repeat prescription (Goesling et al, 2016). In a study reviewing postsurgical opioids, Brat et al (2018) reported that risk of overdose or misuse increased by 44% for each refill of a prescription, and by 20% for each additional week of use. A substantial number of participants in this study could be described as having developed PPOU, and a broader range of opioids were prescribed for them by GPs in the months following discharge, thus exposing them to a greater risk of harm.

Excessive prescribing of opioids after surgery has been recognised as an important concern for public health and a potential contributor to patterns of opioid misuse and related harm (Neuman et al, 2019). Evidence is emerging that the quantity of opioids dispensed can be reduced without compromising effective pain control (Neuman et al, 2019), as prescription opioid consumption is frequently lower than the quantity prescribed. Studies agree that overprescription of opioids occurs for postoperative patients, with most individuals taking no opioid at all, and some taking less than 20% of what was dispensed (Hill et al, 2018; Goudra et al, 2021).

All clinicians must be aware of the potential risks and be able to identify and appropriately manage patients where there might be a risk of prolonged use or misuse. Risk is usually ascertained through the identification of depression, anxiety, pain catastrophisation, history of substance misuse and psychiatric disorders (MacIntyre et al, 2014). The Opioid Risk Tool (ORT) is a 5-item tool designed to predict the possibility of aberrant behaviour and drug misuse, and stratifies into low, moderate, or high risk. However, it relies on self-reporting and can be open to deception and manipulation. Although ORT has been shown to be an effective screening tool in chronic pain management, it has been shown to be ineffective for acute surgical patients because it does not reliably identify patients at risk of aberrant behaviour (Petersen et al, 2020). The retrospective calculation of ORT scores in this project did not elucidate any valuable information and concurs with findings from other studies; patients who potentially would have scored as medium risk were not associated with repeat opioid prescription and as such would not be identified as being at risk of persistent or problematic opioid use. The small sample size does not allow any statistical interpretation but does suggest that screening would be inconclusive. Petersen et al (2020) recommended a limit on the number of pills prescribed rather than screening patients because this is more likely to be an effective management strategy for potential opioid use and associated risks.

There was uncertainty over healing time, duration of analgesia and return to work for more than one-third of people in this study. This may be related to the fact that only half of the participants were fully independent on discharge, which is unsurprising given the nature of the injuries and the median length of stay. Following discharge two-thirds of people attended more than four appointments with their GP. The data did not include reasons for GP attendance, pain assessment or pain scores. However, this group included the people with known risk factors for long-term opioid use and misuse, such as widespread pain, allodynia, preadmission psychological issues, and uncertainty over how long opioid analgesia would be needed or when they could return to work or activity. There is the possibility that these people had developed neuropathic pain and, while they continued to experience unrelieved pain, they continued to take opioid medication. Data concerning pain assessment or scores were not available. Helmerhorst et al (2014) found that patients who continued using opioids 1–2 months after injury had higher pain scores, more psychological distress, and more disability than those who discontinued opioid treatment. Screening for these indicators and risk factors may provide greater utility in identifying patients in need of regular review, which in turn may reduce the potential for long-term and problematic opioid use.

The participants were drawn from a heterogeneous population of patients with lower limb trauma, with regard to background, age, and injury type. Limitations include the small number of participants and recruitment from a single hospital and, along with narrow inclusion criteria, the findings may not reflect experiences of all trauma patients. The results cannot be generalised in a statistical sense but can be seen as furthering the understanding of patients' experiences and opioid use. Information is known about which opioids were prescribed but the data do not include what was taken by the participants after they were discharged from hospital; conclusions cannot be drawn about excessive prescribing or the nature of the pain and its assessment, nor the reasons for seeking repeat prescription. Recommendations from this study include the need for clear discharge advice on the duration of opioid use, including when and how to taper. Development of prescribing protocols may prove useful to limit the quantity of medication dispensed, which in turn may limit the potential for misuse. Patients with the potential for PPOU should be identified before discharge so they can be directed to appropriate support services. The GP and hospital attendances should include advice and guidance on pain management to reduce the potential for long-term opioid use and offer alternative strategies.

Conclusion

This project explored how opioids were used during an acute hospital stay and in the 6 months following discharge; it identified how often patients were reviewed following discharge from hospital, and recommended that risk stratification/opioid risk assessment tools cannot be reliably used for acute pain to predict long-term use of opioid. Although the sample is small, the findings indicate that patients are frequently discharged from hospital with an opioid prescription. For most this will stop within a few weeks of discharge, however for some it will continue beyond the recommended duration of 3 months and will include dose escalation rather than dose reduction.

It is essential that appropriate analgesia is provided on discharge, but clear information regarding the ongoing management and planned discontinuation of opioids must also be provided. PPOU is a real concern and should be recognised as such so that appropriate support can be provided. Clinicians should be aware of the potential risks associated with prolonged opioid use and be able to identify which patients are at most risk. Screening for indicators for long-term use may prove more useful than the application of formal risk stratification tools in an acute pain population.

Future research should look at the transition from acute opioid use to longer term use, consider the advice and ongoing support needed for patients during this time, and explore overprescription of opioids at discharge. Research should also focus on the adoption of screening for long-term opioid use within the acute opioid-naïve patient population. This project was part of a programme of work to understand how opioid-naïve patients use opioids within the first few months of prescription, and the results will inform the development of a pathway to support healthcare practitioners as they support patients discharged with opioid medication to transition from hospital to home.

KEY POINTS

CPD reflective questions