Background
Internationally, an opioid epidemic has evolved during the two first decades of the twenty-first century. Especially, the use of opioids has increased in North America, Western Europe, and Oceania [
1]. Increased prescribed analgesic opioid use has been followed by a dramatic increase in addiction disorders, use of illegal opioids, overdose deaths and even in the number of suicides [
2‐
4]. Long-term opioid treatment is also problematic due to serious adverse effects such as sedation, cognitive impairment and tolerance development, with potentially devastating effects on functional capacity and quality of life [
5‐
7].
The short-term use of opioids for acute and terminal conditions has traditionally been widely accepted. Conversely, there is no strong evidence to support the use of opioids in chronic non-cancer pain conditions, while risks and adverse effects are well established. Long-term use of opioids should therefore usually be avoided for chronic non-cancer pain [
5,
8]. Consequently, updated and knowledge-based practice for opioid prescription is important to avoid further development of an opioid epidemic.
One measure to support optimal opioid prescription is academic detailing (AD). The term academic detailing was first introduced by Avorn and Soumerai in 1983 [
9], when they showed that prescribers receiving personal educational visits (AD visits) reduced the prescription of target drugs significantly compared to groups that received either specified printed information on the matter only or no specific information beside the standard.
AD visits have shown to be a useful and cost-effective way to improve the quality of decisions made about prescribing drugs, as well as to reduce unnecessary expenses [
9]. A systematic review found that AD can be effective, either as a single intervention or as part of a multiple intervention to change the prescriber’s practice [
10]. Moreover, AD has been proposed as a potentially effective intervention to address the epidemic of opioid overuse [
11]. A recent publication showed considerable alignment between self-reported practice change intentions following academic detailing and actual changes in subsequent opioid prescribing [
12], whereas a British study on the effects of an evidence- and theory-informed feedback intervention on opioid prescribing for non-cancer pain in primary care, found that prescribing of strong opioids, total opioid prescriptions, and prescribing in high-risk groups generally fell, although effects lessened after the feedback stopped [
13].
In summary, there are several studies that have done assessments on the changes resulting from AD [
14,
15]. Most studies have been reported from USA [
16], but only a few of them concern opioid prescription. Hence, few studies have reported on usefulness of AD regarding opioid prescription outside USA. Furthermore, there is a lack of knowledge on what professionals who perform AD-visits (academic detailers) experience as widespread challenges in general practitioners’ (GPs’) opioid prescribing for chronic non-cancer pain.
Hence, the aim of the current study was to investigate the usefulness of AD visits on GPs’ opioid prescribing patterns in Norway, and academic detailers’ experiences from AD visits to GPs on opioid prescription.
Discussion
The AD campaign on opioid prescription had been well received by the visited GPs. The non-intervention municipalities in Central Norway showed a slight increase in prevalence of opioids use, similar to what have been shown in the national totals for many years [
23,
26,
27]. In the intervention municipalities, there was a decrease in the total number of prevalent and incident users of opioids in Central Norway, but not in Northern Norway. One of the reasons probably being that the number of opioid users, and particularly those using reimbursed opioids for chronic non-cancer pain, before the intervention were lower in Northern Norway [
28].
We did only study a one-year period after the intervention. Tapering of opioid analgesic treatment might be a slow process that takes time and changes might not have been discovered during the first year.
The findings are in line with a British study [
13], which found that prescribing of strong opioids, total opioid prescriptions, and prescribing in high-risk groups generally fell in intervention practices and rose in control practices. The intervention in the British study included feedback to the prescribers on their prescribing of opioids. Although this separates the British intervention from the AD-visits, the similarities of the content in the messages to the prescribers in the interventions gives support to that updated evidence-and theory-based information on opioid prescription is useful for prescribers in primary care.
Midboe et. al [
16] highlight one-on-one sessions and provider networking as two of five key lessons important to gain success performing AD interventions addressing the opioid epidemic. The AD-visits in the current study were performed one-on-one, and most of the providers visited had been part of previous AD campaigns and were thus familiar with both KUPP and the academic detailers. Midboe et. al [
16] also suggest that training detailers in motivational interviewing (MI) in general is helpful. The academic detailers in the current study were not specifically trained in MI, but they all had taken part in a one-day training course specially designed for the campaign. Our study can add that academic detailers might need more knowledge on the various treatment options that are actually available for the different GPs they visit, including MI. Notably, further research should explore the GPs experiences and views on this as there is a lack of knowledge on this matter from their perspective.
According to the academic detailers not all suggested non-pharmacological treatment options presented in the campaign were considered as relevant or available by the GPs they had visited. This implies that, even though the suggested alternatives to opioids were considered well suited for chronic non-cancer pain patients, they were not necessarily perceived as available for the GPs and their patients. With regards to the finding in the registry study, it might be that a lack of available non-pharmacological treatment methods could be a possible explanation for the difference in changes between Central and Northern Norway, understood as a difference in available options to opioid prescription between Central and Northern parts of Norway.
A relevant reason for that the decrease in total numbers of users were less marked than the fall in new (incident) users, is that, due to pharmacodynamical effects of opioids, it will be easier not to recruit new users than to discontinue treatment in established opioid users [
29,
30]. This hypothesis is supported by the qualitative finding on the academic detailers’ practice of emphasising that a main message in the campaign was to avoid new opioid users. In addition, the number of patients with chronic non-cancer pain diagnoses has increased steadily the last years, so achieving a non-rise in the number of opioid prescriptions may well represent an improvement when compared to the steady rise the last years [
26,
27].
Because a patient’s opioid use includes prescriptions not only from the patient’s GP but also from hospital doctors, a concern among the academic detailers was about having only GPs as the target group for the campaign. Hence, a question raised was whether a similar campaign also should have been offered to hospital prescribers. The effect pain centres and hospital doctors may have on opioid prescriptions has recently been addressed in the new Norwegian National guidelines on prescribing restricted drugs [
31], advising only to prescribe small amounts of opioid analgesics before the patients are referred back to their GP for further treatment. From the current study, it can be added that when planning future AD campaigns, one should consider how to bring the same message across to other suited receivers, such as hospital prescribers as well as to patients.
Strengths and limitations
A major strength of this study is the combined approach with qualitative and quantitative design. Moreover, utilization of registry data excludes possibility of recall bias, and the nature of the registry allow for identification of control municipalities in the quantitative part of the study.
There are some noteworthy limitations. By interviewing the academic detailers, we got focus group participants that had gathered experience from several visits, whereas GPs could have just accounted for their own single visit.
The sampling strategy for the focus groups could have led to a biased sample as the informants were initially identified by the KUPP management. Nevertheless, the sample showed variations as planned. Transcribing only the most important parts from the audio recordings of the focus group interviews can potentially lead to selection bias and influence the results. However, the audio recordings and notes written after each focus group interview were actively used throughout the analysis process to minimize the bias.
Furthermore, data from the NorPD only contains filled prescriptions, i.e. prescriptions where the patients have had their medications dispensed from a pharmacy. Hence, prescribed, non-filled prescriptions are not a part of the NorPD-data. The number of non-filled prescriptions, where a change would reflect a change in patients’ behavior and not in the GPs’ prescription behavior, are most likely to be the same before and after the intervention, and will thus not have influenced our results.
The municipalities were not randomly assigned to receive AD visits or not. The booking process were likely to select GPs that were positive to receive visits. We do not know if the GPs in the intervention municipalities differ from those in the non-intervention municipalities. Since we are looking at change within the different municipalities, we consider this to be acceptable, but if the GPs who were positive to receive visits were more interested and willing to make changes in their prescription practice, this could have influenced the results.
Due to regulatory limitations in the NorPD, the results do not include any changes in prescriptions for other analgesics (e.g., NSAIDs or paracetamol (acetaminophen)) or individual changes in prescribing as we used municipalities as the study unit. Hence, we are not able to compare individual self-reported intention of practice change with the actual change in the individual prescriber’s practice as done by Saffore et al. [
12].
Many studies on the use of opioids benchmark their data using morphine/opioid milligram equivalents (MME or OME) as a measure to standardize opioid doses and quantities across agents. However, that was not a possibility within the datasets used in this study.
Using DDDs, as we have done in this study, could be a limitation if the distribution of strong and weak opioids varied in the two time periods that were compared and between the intervention and the control groups. This is because the DDD value of a weak opioid represents a lower analgesic effect than a DDD of a strong opioid since the DDD is a technical value assigned according to indication. Weak opioids are indicated for weak and moderate pain while strong opioids are indicated for strong pain. Since the time period studied was relatively short, and all municipalities were in Norway, it is not probable that there should be large changes in the distribution of prescriptions of weak and strong opioids. Moreover, as this is not a randomized controlled study, other campaigns or messages released during the same period might have influenced the results.
Conclusions
In the intervention municipalities, we found a reduction in the number of prevalent and incident opioid analgesic users, and most prominent in incident users who received reimbursed opioids for chronic non-cancer pain in Central Norway.
The GPs’ length of working experience and familiarity with the topic gave different presumptions for making use of the information presented in the AD-visits. When planning future AD campaigns, one should consider how to bring the same message across to other suited receivers, such as hospital prescribers, and also to patients.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.