Prescription drug monitoring programs in Australia: A call for a comprehensive evaluation

Much has been written on the multiple ‘opioid crises’ that have emerged over the past few decades in North America [1-3]. The first ‘wave’ began in the late 1990s, with a key driver being the widespread availability of prescription opioids. In response to identifying this putative causal factor in rising opioid overdose deaths, a range of interventions were introduced, including prescription drug monitoring programs (PDMP). PDMPs are electronic databases that monitor the prescribing and dispensing of controlled medicines including opioids, with the potential to identify patient-level medication-related risks [4]. The purpose of PDMPs is to curtail inappropriate opioid prescribing, identify and prevent opioid misuse and diversion [5] and ultimately reduce opioid overdose deaths [6]. Although basic, paper-based PDMPs were first introduced in the USA over a century ago [7], it is recent technological advances that have seen these programs evolve significantly and, as of 2021, all 50 states in the USA have PDMPs [8]. They are also widely adopted in Canada, parts of Europe and more recently Australia [9, 10].

Despite their widespread adoption, the evidence base around the effectiveness of PDMPs is inconclusive and often contradictory. Of greatest concern locally is that PDMPs may be causing rather than reducing harm. International literature has shown PDMPs have been associated with a range of unintended consequences including abrupt discontinuation of opioids, ‘substitution effects’ including prescribing less effective pain medications, fractured provider–patient relationships and stigmatising behaviours and perceptions by health-care providers towards people who may engage in non-medical prescription medication use [4, 11]. Furthermore, several US studies suggest that restricting access to opioids via PDMPs caused patients to seek illicit drugs, particularly heroin [12-14]. The consequences of this shift are such that, although the USA has observed a 50% reduction in the volume of prescription opioids supplied in the past decade, no equivalent reduction in prescription opioid-related deaths has been observed [15]. Meanwhile, heroin-related deaths have risen. To the extent that this pattern can be attributed to the introduction of PDMPs, it begs the question of whether these programs are having their intended effect of reducing opioid-related deaths.

Australia has one of the highest rates of prescription opioid utilisation globally [16], and opioids were the most common drug class associated with unintended drug-induced deaths in 2020 [17]. To address these harms, in 2018, all Australian states and territories agreed to implement a ‘real-time’ PDMP, which allows prescribers and pharmacists to access real-time prescribing and dispensing information for monitored medicines. Individual jurisdictions are responsible for the operation and implementation of their PDMP [18], with the overall aims of these programs being to identify patients diverting medicines and who are at risk of harm due to dependence or non-medical use of controlled medicines, and to limit multiple provider episodes [18].

Given Australia's recent commitment to PDMP implementation, it is imperative that unintended consequences observed in North America are avoided locally. Central to this aim, a commitment to a robust, dedicated evaluation of agreed key outcome indicators is essential. This will provide policymakers with a detailed understanding of whether PDMPs are having the desired impact on specific outcomes such as reductions in opioid prescribing patterns, diversion, fatal and non-fatal overdoses and rates of opioid use disorder, dependence and non-medical use [9, 19]. Importantly, rigorous evaluation efforts can result in findings being provided to decision-makers, allowing them to adapt these programs in light of evidence-based outcomes.

Secondary to evaluating key outcome measures, efforts should also focus on identifying specific nuances of PDMPs that are associated with positive or negative outcomes and the reasons for such outcomes [19]. For example, PDMPs commonly vary in the risks they identify (e.g. high dose, multiple prescriber episodes, high-risk drug combinations), the medications they monitor, who can access the PDMP (e.g. prescribers, dispensers, law enforcement) and the data captured [20].

Several US studies have sought to identify whether specific PDMP features or characteristics are associated with improved outcomes, including reduced opioid-related overdoses and deaths. For example, more robust PDMPs, often defined as monitoring greater number of controlled medicines, having more frequent or real-time reporting and having mandates to check the PDMP, were associated with reduced prescription-opioid-related deaths [21-23]. Consistent with mandates reducing prescription-opioid-related harms, several other studies have reported these mandates are associated with increased heroin-related deaths [13, 24]. Understanding the impacts of specific features can be used to inform and improve the efficacy of both current and future PDMPs [22] and should be a priority given the heterogeneity of Australian PDMPs and their specific features.

Study design is another important consideration. A US academic review of state policy and systems-level interventions on prescription opioid overdoses identified the need for more rigorous study designs, including natural experiments, quasi-experimental designs with comparison groups, and time-series analyses [25]. Australian PDMP features vary by jurisdiction, for example, prescribers and pharmacists are mandated to review the PDMP in Victoria, Queensland and South Australia, while in other jurisdictions it is ‘recommended’. These distinctions provide ideal conditions to undertake studies adopting these rigorous designs, which can draw stronger conclusions about the impacts of such features. Such evidence is key to ensuring the significant investment in these interventions is warranted and that they are having the required, tangible effects.

Early signs of possible unintended consequences have already been observed in Victoria, which in April 2020 was the first state to mandate PDMP use. Amongst a representative sample of community pharmacists, initial evidence of automation bias, the process by which clinicians attribute greater importance to PDMP alerts than other possible risk factors, was observed [26]. More specifically, results revealed a considerable proportion of pharmacists made decisions based on PDMP alerts alone, while other critical clinical information did not influence decisions to supply opioids [27]. Another Victorian study amongst a cohort of people who inject drugs revealed one in five people reported being refused a prescription for a PDMP-monitored medication by a general practitioner, while approximately 3% of the sample reported being refused dispensing by a pharmacist, since the PDMP was introduced [10]. Many of these medications were for mental health-related conditions and of concern is the impact of these refusals, yet it is unclear whether these refusals were in fact clinically appropriate, and what other care was offered, given the absence of any further published monitoring or evaluation. Despite PDMPs being implemented to improve patient outcomes, there is a lack of research on individual patient outcomes, including pain management, quality of life and functioning [28]. Evaluation efforts should therefore also determine the subsequent, longer-term clinical and other outcomes of PDMP use and related clinical decisions.

For complex problems, there are rarely simple solutions. PDMPs are an example of a technological solution that has been proposed to solve a complex problem, yet research has demonstrated that outcomes are often inconclusive and, more concerning, can even cause harm. PDMPs have enormous potential to reduce opioid-related risks; however, it is essential that known unintended consequences of PDMPs, observed in the USA, are at a minimum measured so that they can be mitigated. Given Australia's substantial national investment in PDMPs, we must understand the impact of these interventions, so harms are minimised, utilisation optimised and any possible benefits maximised.

FUNDING INFORMATION

LP is the recipient of the National Health and Medical Research Council Investigator Grant (#2016909). AR is the recipient of the National Health and Medical Research Council Investigator Grant (#2016695).

CONFLICT OF INTEREST STATEMENT

SN has received untied educational grants to study pharmaceutical-opioid-related harm from Seqirus, and is a named investigator on an implementation trial of buprenorphine depot funded by Indivior (no funding received by SN personally or through her institution).