Article Text
PDF
Abstract
Objectives The COVID-19 pandemic and related lockdown measures disrupted global healthcare provision, including opioid prescribing. In North America, opioid sales declined while opioid-related deaths increased. In Europe, the effect of the pandemic on prescribing is not yet known. Given the ongoing increase in opioid-related harm and mortality, it is crucial to analyse the impact of the COVID-19 crisis and lockdown measures on opioid prescribing. Therefore, the objective of this study was to characterise opioid prescribing in the Netherlands during the COVID-19 pandemic.
Design A nationwide register-based study characterising opioid prescribing using aggregated insurance reimbursement data.
Setting Dutch healthcare during the first 2 years of the COVID lockdown.
Participants The whole Dutch population.
Primary and secondary outcome measures Comparing the number of opioid prescriptions during the pandemic with a prepandemic period using a risk ratio (RR), with separate analysis on the prescription type (first-time or repeat prescription), patients’ sex, age and socioeconomic status. We also explored lockdown effects.
Results During the first lockdown, the total number of new opioid prescriptions and prescriptions to young patients (briefly) decreased (RR 0.88, 95% CI 0.88 to 0.89 and RR 0.73, 95% CI 0.70 to 0.75, respectively), but the overall number of opioid prescriptions remained stable throughout the pandemic compared with prepandemic. Women, older patients and patients living in lower socioeconomic areas received more opioids per capita, but the pandemic did not amplify these differences.
Conclusions The pandemic appears to have had a limited impact on opioid prescribing in the Netherlands. Yet, chronic use of opioids remains an important public health issue.
- COVID-19
- PAIN MANAGEMENT
- EPIDEMIOLOGY
Data availability statement
Data are available upon reasonable request.
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
STRENGTHS AND LIMITATIONS OF THIS STUDY
This study included opioid prescription data during the COVID-19 pandemic for the entire Dutch population.
This study used weekly data and was therefore able to monitor the immediate effects of COVID-19 lockdowns.
This study used aggregated data and could not track individual patients over time.
It was unknown for which indication opioids were prescribed.
Introduction
The COVID-19 pandemic significantly disrupted healthcare provision; global care utilisation decreased by about a third during the first months of the COVID-19 pandemic.1 Primary care was more often delivered remotely,2 3 elective surgeries were postponed,4 and some patients avoided using healthcare altogether.5 In the Netherlands alone, postponing surgeries led to a loss of 320 000 quality-adjusted life years.4 In both Dutch and US surveys, 40% of respondents reported experiencing negative health effects due to postponed care.6 7 These changes in healthcare provision resulted in, among other things, changes in drug-prescribing behaviour, including opioids.8–10
The COVID-19 pandemic also had detrimental effects on the US opioid epidemic.11 In the USA, the number of unintentional opioid-related deaths rose by 64% in 2021 after they had stabilised between 2017 and 2019.11 In Canada, opioid-related deaths increased by around 70%–80% in 2020–2022 compared with 2017–2019.12 Remarkably, at the same time, global opioid sales declined by 5.2% in the first months of the COVID-19 pandemic9 and studies from the USA, Canada and Australia show a decline in opioid prescribing during the COVID-19 pandemic.9 13–18
A combination of factors may explain these reductions. With the increased awareness of opioid-related risks and the consequent changes in clinical practice guidelines, this may simply be an acceleration of the downward prescription trends that have been present since 2016.11 Furthermore, the postponement of elective surgeries4 and the reduction in acute trauma-related care may also explain the reduction in prescribing.19 Yet, although the decrease in prescribing over the last few years was a necessary response to the overprescribing of opioids, it may also have caused patients to seek out illicit opioids,20 potentially explaining this increase in opioid-related mortality in North America.
Given the risks related to opioid use and the ongoing increase in opioid-related harm and mortality worldwide,21 it is relevant to monitor opioid-prescribing practices. While numerous data are available from the COVID-19 pandemic in North America and Australia, there are limited data from Europe. One study from the UK showed that opioid prescribing remained stable during the COVID-19 pandemic.22 Data from the Netherlands shows that during the first few months of the COVID-19 pandemic, general practitioners (GPs) prescribed less oxycodone and non-steroidal anti-inflammatory drugs (NSAIDs) to new patients.23 24 It is not yet known how the pandemic has affected opioid prescribing in the later stages of the pandemic nor is it known if lockdown periods specifically affected prescribing.
Furthermore, a Dutch study showed that healthcare reductions due to the COVID-19 pandemic were disproportionately distributed between groups. There were more reductions in care for women, patients with older age, patients with migrant backgrounds and those living below the poverty line.19 It is important to study whether the COVID-19 pandemic amplified existing variance and inequalities in opioid prescribing.
The aim of the current study is, therefore, to (1) analyse the impact of the COVID-19 pandemic and related lockdown measures on opioid-prescribing practices throughout the pandemic compared with the prepandemic and (2) explore whether there was any change in variability in opioid prescribing regarding age, sex and socioeconomic status (SES).
Methods
Data sources and sample
We conducted a nationwide register-based study characterising opioid prescribing for the complete Dutch population. Pharmaceutical reimbursement data were obtained from the Drug Information Project (Genees-en hulpmiddelenInformatie Project (GIP)) from the Dutch National Healthcare Institute (Zorginstituut Nederland). This database covers prescriptions dispensed for extramural use, covering 96% of the Dutch population.
Data on the number of prescriptions and defined daily doses (DDDs) were requested for all opioids that can be prescribed in the Netherlands (online supplemental table 1). Codeine was excluded due to its primary indication for cough (Anatomical Therapeutic Chemical (ATC) code R05) and because the combination paracetamol/codeine (N02AJ06) is not reimbursed in the Netherlands and is thus not present in our dataset.
Supplemental material
These data were provided in aggregated form, broken down by dispensing week, patient characteristics (sex, age and SES), opioid type (ATC code and administration route), prescriber specialism and prescription type (new or repeat prescription). A prescription was defined as ‘new’ if this was the first time the patient received this ATC code in the past 12 months. The total number of opioid users per year was available from the publicly available part of the GIP database.
SES was provided as one of five categories, ranging from low to high. The Dutch National Healthcare Institute used zip code data from Statistics Netherlands (Centraal Bureau voor de Statistiek) to obtain the neighbourhood SES of the underlying patients. Neighbourhood SES is based on mean income, education and employment history.25
We used the Oxford Government Response Score tracker to define lockdown periods.26
Patient and public involvement statement
Patients and/or the public were not involved in the design, conduct, reporting, or dissemination plans of this research.
Timing
We included data from the first 2 years of the COVID-19 pandemic and 1 year before the COVID-19 pandemic (reference year). We defined the first year of the COVID-19 pandemic to start at week 11 of 2020 until week 10 of 2021 (from 9 March 2020 to 14 March 2021) and the second year of the pandemic from week 11 of 2021 until week 10 of 2022 (from 15 March 2021 to 13 March 2022). We compared these data to a reference year: week 11 of 2019 to week 10 of 2020 (from 11 March 2019 to 8 March 2020).
To explore variations in opioid prescribing due to lockdown effects, we defined three COVID-19 lockdown periods in the Netherlands. A period was considered a lockdown if the Oxford Government Response Score was consistently 50 out of 100 points for 6 weeks or longer (online supplemental figure 1).26 For the Netherlands, these correspond to 9 March until 5 July 2020 (lockdown 1); 17 August 2020 until 27 June 2021 (lockdown 2) and 22 November 2021 until 30 January 2022 (lockdown 3). The included time periods are depicted in figure 1.
Definition of two COVID-19 years, the reference year and three lockdowns.
Statistical analysis
Descriptive analyses
We conducted descriptive analyses for the total number of opioids prescribed per week, with separate analyses split by prescription type (first or repeat prescription), sex, age and SES.
Opioid usage per 100 000 inhabitants was calculated using population data from Statistics Netherlands by sex, age and SES.27
Interrupted time series analyses
Interrupted time series analyses (ITSA) were conducted to evaluate long-term prescribing trends before and after the onset of COVID-19. An autoregressive integrated moving average model (1,0,0) was used for this purpose. We compared the 52 weeks pre-COVID-19 with the 105 weeks during the COVID-19 pandemic. The p values were Bonferroni corrected.
Risk ratios (RRs)
To account for recurring seasonal effects in prescribing, we calculated a weekly RR for receiving a prescription during the pandemic compared with the same week’s prepandemic. We calculated the weekly prevalence of receiving an opioid prescription by dividing all prescriptions in a week by the number of inhabitants that month. Both the weekly prevalence of prescriptions during the first and second years of the pandemic were compared to the weekly prevalence in the reference year prepandemic (ie, prescription data from week 14 of 2020 and week 14 of 2021 were both compared with data from week 14 of 2019). 2020 had 53 weeks; both the prescribing during weeks 52 and 53 of 2020 were compared with week 52 of 2019. Additionally, we calculated RRs per lockdown (comparing the whole time period of a lockdown to the same weeks in the reference year). Again, analyses were split by prescription type, patient sex, patient age and patient SES.
A 95% CI for the RR was created for each data point.28 Given the large dataset and small CIs, we defined an RR as relevant rather than statistically significant. A difference in prescribing was considered relevant if the increase or decrease in the number of prescriptions changed by more than 10% (RR with CI under 0.9 or over 1.1).
As we included both N02 (analgesia) and N07 (addictive disorders) opioids, we conducted a sensitivity analysis to assess the stability of our findings when excluding N07 opioids. Descriptive analyses and RRs were analysed in R V.2022.02.01,29 and the ITSA was conducted in SPSS V.29.0.30
Results
Prescription and patient characteristics
Per 100 000 inhabitants, there were 5703 individual opioid users in the Netherlands, and an individual patient received on average 5.3 prescriptions. The average size of a prescription was 10.1 DDD (10.8 for men and 9.7 for women). Oxycodone, tramadol and fentanyl were most often prescribed (respectively, 36.5%, 30.5% and 14.3% of all prescriptions). Prescriber specialism was not registered for half of the prescriptions, precluding analysis on the prescriber level. Descriptive data can be found in online supplemental table 2: figures 2–6.
Number of opioid prescriptions dispensed to all patients
Opioid prescribing rose by 0.75% (from 30 206 to 30 432 prescriptions per 100 000 inhabitants) in the first year of the COVID-19 pandemic compared with the prepandemic. 2020 had 53 weeks instead of 52. In the second year of the pandemic, opioid prescribing declined again (30 165 prescriptions per 100 000 inhabitants, −0.88% compared with the first year of the pandemic and −0.14% compared with prepandemic). Overall, opioid prescribing increased by 0.31% in the two COVID-19 years compared with the prepandemic year.
During the first lockdown, the RR compared to prepandemic was 0.96 (95% CI 0.96 to 0.96). For the second lockdown, it was 1.03 (95% CI 1.02 to 1.03) and for the third lockdown, 1.02 (95% CI 1.01 to 1.02). See also online supplemental file 5.
For the total population, there were 4 weeks (13–19 April, 4–10 May, 18–24 May and 1–7 June 2020) during the first lockdown in which the RR of receiving any opioid prescription was reduced by over 10%. During the second lockdown, there were 1 week (21 December 2020–10 January 2021) in which opioid prescribing increased by over 10% and 2 weeks in which it decreased by more than 10% (10–16 May and 24–30 May 2021). During the third lockdown, there was 1 week in which opioid prescriptions increased by over 10% (27 December 2021–2 January 2022) (see online supplemental figure 7). See online supplemental file 7 for the exact RRs. Yet, the overall number of opioids dispensed during the course of the pandemic remained stable; the ITSA showed a slope change of 0.68 prescriptions per 100 000 inhabitants per week during COVID-19 years. See also table 1.
Interrupted time series analysis, including the p-value for the slope change
Number of opioid prescriptions dispensed, split by first and repeated prescriptions
83.8% of all opioid prescriptions were repeat prescriptions for existing patients, which remained relatively stable over time. However, during the first lockdown, there were fewer first-time prescriptions: RR 0.88 (95% CI 0.88 to 0.89). Using weekly data, there were 9 weeks in which the prevalence of a first opioid prescription was reduced by 10% or more compared with the same week prepandemic. During the second lockdown, there were 4 weeks in which the prevalence of receiving a first opioid prescription was reduced (11–17 January, 5–11 April, 10–16 May, 24–30 May 2021) and one (4–10 January 2021) in which it was increased. During the third lockdown, prescriptions were increased for 2 weeks (27 December 2021–9 January 2022). These fluctuations with an RR are depicted in figure 2; the grey areas represent lockdown periods.
RR for receiving an opioid prescription during the pandemic (mid-March 2020–mid-March 2022), compared to receiving an opioid prescription the same week in the period prepandemic (mid-March 2019–mid-March 2020), split by prescription type. The dashed line represents an RR that deviates −0.1 or +0.1 from the RR of 1. The grey areas represent lockdown periods. RR, risk ratio.
Number of opioid prescriptions dispensed, split by patient sex
During both the COVID-19 pandemic and the reference year, women received more opioid prescriptions than men—nearly 60% of all prescriptions. This proportion remained stable during the COVID-19 pandemic. Throughout the pandemic, there were fluctuations in opioid prescribing, but these were not related to sex (online supplemental figure 8).
Number of opioid prescriptions dispensed, split by patient age
People aged 80 and over received the most opioids during the pandemic, with 2301 prescriptions per week per 100 000 inhabitants. Those aged between 60 and 79 received 1017 prescriptions, respectively; those aged 40–59 received 674 prescriptions and those aged between 20 and 39 received 223 prescriptions. Young people (aged 19 and under) received fewer opioids throughout the whole pandemic, decreasing from 12.5 weekly prescriptions prepandemic to 11.5 prescriptions during the pandemic. This effect was most pronounced during the first lockdown, with an RR of 0.73 (95% CI 0.70 to 0.75) for receiving a prescription compared with the same period of prepandemic. This reduction was most pronounced during the first lockdown, with a reduction of over 50% compared with the prepandemic year (figure 3).
RR for receiving an opioid prescription during the pandemic (mid-March 2020–mid-March 2022), compared with receiving an opioid prescription the same week in the period prepandemic (mid-March 2019–mid-March 2020), split by patient age. The dashed line represents an RR that deviates −0.1 or +0.1 from the RR of 1. The grey areas represent lockdown periods. RR, risk ratio.
Number of opioid prescriptions dispensed, split by patient SES
Patients living in lower SES zip code areas were prescribed more opioids over the whole study period. When controlling for the population size of the respective zip code areas, patients living in the lowest category SES codes received nearly twice the number of opioids (738 per week/per 100 000) than those living in the highest category SES codes (340 per week/per 100 000). Patients in the second, third and fourth quantiles (low, medium and high SES) received 628, 525 and 445 opioid prescriptions per week/per 100 000, respectively. Fluctuations in opioid prescribing due to lockdowns were similar for all SES groups (online supplemental figure 9).
Sensitivity analysis
When excluding N07 opioids in the sensitivity analysis, we observed a 0.33% increase in the number of prescriptions during the COVID-19 years, compared with 0.31% for all opioids. For the ITSA, the slope change with the onset of COVID-19 for all opioids was 0.68 (p=0.01), compared with 0.61 (p=0.02) when excluding N07 opioids. Regarding RRs, when including N07 opioids, we identified 9 weeks with RRs outside the 0.9–1.1 range. When excluding N07 opioids, we identified 10 relevant weeks. These findings did not change the overall results. See also online supplemental file 8.
Discussion
This study shows that in the Netherlands, the number of opioid prescriptions to existing patients remained stable during the COVID-19 pandemic and that there was an overall brief reduction in first-time prescriptions. This was mostly driven by reduced prescribing to young patients during the first lockdown. Overall, there was no meaningful change in the number of opioids prescribed during the pandemic. Women, older patients and patients living in lower SES areas received more opioids per capita, but these differences were not amplified by the pandemic.
Our findings are in line with UK data, where overall prescription opioid use remained stable during the COVID-19 pandemic.22 The reduction we observed in prescriptions to young people could be explained by a reduction in traffic and sports incidents as well as the postponement of elective surgeries during the first lockdown.1 5 31 Other studies show similar results; in the USA, the number of patients receiving their first opioid prescription was also reduced during their first lockdown.13 32 The differences in opioid prescribing per capita by sex and SES have previously been described, and it has also been shown that pain is more prevalent in women and patients living in lower SES areas.33–38
In contrast with our data, there was an overall decline in opioid prescribing during the COVID-19 pandemic in the USA, Canada and Australia.9 13–18 At the same time, these countries also witnessed a significant increase in opioid-related harms and overdose deaths.11 12 39 40 Fortunately, in the Netherlands, we did not observe a worsening of opioid-related harms during the pandemic; there were 124 opioid-related deaths in 2019, 164 in 2020 and 150 deaths in 2021, on a population of 17.5 million people.41
In the USA, the cuts in opioid prescribing may have forced patients to seek opioids on the illicit market.20 The lower number of overdose deaths in the Netherlands may be associated with better access to emergency and addiction care. Prior research has indeed shown that opioid-related deaths are higher in countries where the utilisation of addiction care is lower.42
It is important to acknowledge that the chronic use of opioids is still high and requires constant attention. Since 2018, significant focus has been directed towards reducing opioid utilisation around the globe, also in the Netherlands. Before 2018, there had been a notable upward trend in opioid consumption for 10 consecutive years,43 driven by increased prescribing to patients with chronic non-cancer pain.44 A minor decline in usage was observed in 2019, and this slight reduction persisted through 2020, after which the number of individual opioid users slightly increased.45 Initiatives to reduce opioid initiation promote short-term rather than long-term opioid use (since long-term opioid use is not effective46) and support tapering for existing patients.47
The COVID-19 pandemic may have also affected the utilisation of other pain medications in the Netherlands, such as NSAIDs. GPs prescribed fewer NSAIDs to new patients during the first lockdown.23 24 However, NSAIDS and acetaminophen are readily available over-the-counter in pharmacies, druggists and supermarkets, which all remained open during COVID lockdowns. Medication that is bought without a prescription is not monitored in the Netherlands. Therefore, we cannot measure a potential increase in the use of over-the-counter acetaminophen or NSAIDs.
Our study has few limitations. As we only had access to aggregated data, we could not track individual patients over time, potentially missing shifts in opioid use in individual patients or mistakenly identifying a prescription as ‘new’ if a patient switched to another opioid. We only had access to data on extramural prescriptions, so it remains unclear how the pandemic has affected opioid use within the inpatient hospital setting. However, hospital stays after surgery are generally very short in the Netherlands, and medication prescribed at discharge is part of extramural use. Furthermore, codeine was excluded because the combination paracetamol/codeine (N02AJ06) is not reimbursed in the Netherlands. Codeine is only prescribed with the R05 ATC code and is primarily used to treat cough.48 Also, the prescriber type (GP vs hospital specialist) was unknown for over half of all prescriptions, preventing reliable analysis on the prescriber level. Furthermore, while we assessed differences in prescribing by patient sex, age and SES, we could not split prescriptions by the ethnic or migration background of the patient. Dutch research indicates that patients with a migrant background suffered more healthcare reductions during the pandemic,19 and international studies show that fewer opioids were prescribed to ethnic minorities during the pandemic.49 50
Despite these limitations, we present data that cover nearly the entire Dutch population (96%) and the whole COVID-19 period. At the end of January 2022, the most severe restrictions were lifted and by March 2022 (the end of our inclusion period), most COVID-19 restrictions were lifted.51 52 Furthermore, using weekly data, we monitored the immediate effects of the lockdown and changes over time. By using a risk ratio to compare prepandemic prescription patterns with those during the pandemic, we filtered out recurring seasonal variations while having an overview of overall trends.
Future research could assess the aforementioned inequalities based on migration background in prescribing in the Netherlands. Furthermore, insight needs to be created into opioid prescribing for the whole of Europe, especially in countries with the highest opioid-related harm, such as Scotland, Ireland and Estonia.53 Both opioid prescribing and opioid-related harms in European countries should be continuously monitored so that appropriate policy responses may be taken. Although there was no increase in opioid-related deaths in the Netherlands, it is still unknown how the COVID-19 pandemic affected other opioid-related harms, such as opioid-related hospital admissions and the number of opioid use disorder treatments, which may also be a direction for future research.
Conclusion
The overall number of opioid prescriptions remained stable throughout the pandemic; however, there was a temporary reduction in first-time prescriptions for young people, especially during the first lockdown. In the Netherlands, the pandemic’s influence on opioid prescribing appears limited. Yet, chronic use of opioids remains an important public health issue.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Acknowledgments
The authors thank the Dutch National Healthcare Institute (Zorginstituut Nederland), specifically Saskia Knies and Ilo Boukes, for providing and preparing the data analysed in this study and for their technical support.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
AContributors Concept and design: HE, AS, GAK, DAV, MB, ED and SK. Acquisition, analysis or interpretation of data: HE, AS, GAK, DAV, CK, AD, SASvdH, MB and EvD. Drafting of the manuscript: HE, AS and ED. Statistical analysis: HE, AS, DAV and EvD. Critical revision of the manuscript for important intellectual content: HE, AS, GAK, DAV, CK, AD, SASvdH, MB and EvD. Approval of the final version: HE, AS, GAK, DAV, CK, AD, SASvdH, MB and EvD. Data preparation and technical support: Saskia Knies and Ilo Boukes. HE is responsible for the overall content [as guarantor].
Funding This work is part of the research project Tackling and Preventing the Opioid Epidemic (TAPTOE) and received funding from the Dutch Research Council (Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO)) in the framework of the NWA-ORC Call (NWA.1160.18.300). The views expressed in this paper are those of the authors and do not necessarily reflect the views or policies of the NWO. The NWO is not liable for any use that may be made of the information presented. The Dutch National Healthcare Institute (Zorginstituut Nederland) provided data and technical support as an in-kind contribution for the study.
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.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.