The cohort examined comprised 227,864 people who were dispensed a total of 3,441,851 prescriptions in the PBS 10% sample, and 4,871 unique pregabalin overdose calls were recorded to the New South Wales Poisons Centre. Use peaked in 2018 before plateauing and then declining. Investigators detected three distinct changepoints in dispensing data at December 2018, May 2020, and December 2021. The combined effect across these breakpoints produced a dispensing rate ratio of 0.748 (95% confidence interval 0.700 to 0.800). Compared with a counterfactual projected from historical trends, the authors estimated about 118,390 (93,960 to 140,940) fewer national pregabalin prescriptions by December 2021 and an ongoing monthly decline of 0.7% (0.5% to 0.9%).
Overdose calls mirrored the dispensing trend. Each breakpoint aligned with reductions in overdose rates, producing a combined overdose rate ratio of 0.532 (95% CI 0.406 to 0.695). The reduction corresponded to an estimated 41.3 (26.9 to 52.4) fewer pregabalin overdose calls in December 2021 and a continued monthly reduction of 1.6% (1.0% to 2.3%).
The authors examined temporal relationships between the observed breakpoints and three major external signals. First, late 2018 national media coverage highlighted harms associated with pregabalin; a changepoint was observed in December 2018. Second, prescribing and dispensing restrictions occurred in April–May 2020 at state and national levels; the changepoint in May 2020 coincided with these measures. Third, boxed warnings and manufacturer information updates were implemented in February 2021; a December 2021 breakpoint may represent a delayed effect of those warnings, as reported by the study team. The investigators did not claim definitive causation but reported consistent temporal correspondence between interventions, communications, and declines in both dispensing and overdose calls.
Latent class analysis of 57,744 people dispensed pregabalin in the 12 months after the final breakpoint identified four clusters. One cluster was characterised by features the study team defined as high risk. This high risk group included 7,790 people (13.5% of the analysed subgroup). They were more likely to be male and younger than other clusters, received larger quantities of pregabalin with shorter refill intervals, and had higher rates of concomitant opioid and benzodiazepine use—6,120 (65.7%) had opioid co-use and 2,455 (31.5%) had benzodiazepine co-use within that group. The article reported these proportions directly; further demographic breakdowns and clinical comorbidities were not detailed in the summary provided.
The study used two linked, population-level sources. Dispensing was measured using Australia’s Pharmaceutical Benefits Scheme (PBS) 10% sample, a national dataset that contains a random 10% sample of people ever dispensed a PBS-subsidised drug; PBS package codes for pregabalin strengths (25 mg, 75 mg, 150 mg, 300 mg) were included. Overdose was measured using calls to the New South Wales Poisons Centre, which handles roughly half of Australia’s poisoning calls and codes pregabalin with a specific substance identifier. The Poisons Centre dataset included intentional pregabalin exposures—recreational use, self-harm overdoses, and other intentional non-medical uses. Unique calls were counted; follow-up calls for the same person were excluded from overdose counts.
The study team reported using changepoint analysis as the primary approach to identify timing of shifts in monthly population-adjusted dispensing rates. They supplemented this with interrupted time series and latent class analysis to estimate rate ratios, counterfactual differences, ongoing monthly trends, and to classify people who use pregabalin into distinct clusters based on dispensing and drug use characteristics. Key outcomes reported in the article excerpt were population-adjusted rates of dispensing and overdose, timing of breakpoints, rate ratios with 95% confidence intervals, estimated differences from counterfactual trends (number of prescriptions and overdose calls avoided), and proportion and characteristics of a high risk user cluster.
The investigators reported observational, retrospective analyses of administrative and poisons-call data; temporal associations were described between media, regulatory, and prescribing changes and observed declines in pregabalin dispensing and overdoses. The authors did not, in the provided text, report causal proof that the interventions produced the declines, nor did they present certain model parameter details in the summary. Unanswered questions documented by the study team include the extent to which each individual policy or communication action directly influenced clinician or patient behaviour, and how durable the reductions will be beyond the study window through December 2024. The study therefore ends on concrete, measurable follow-up priorities rather than a generalised wrap-up: disentangling the relative contributions of media versus regulation, monitoring persistence of reduced prescribing and harms, and tracking outcomes in high risk subgroups identified by the latent class analysis.