Does a feedback intervention reduce potentially harmful gabapentinoid painkiller prescribing in general practice?

Problem

Many doctors and professional bodies are concerned about the rise in potentially harmful gabapentinoid prescribing, with primary care prescribing doubling between 2007-2017. There is little evidence for the effectiveness of these drugs in long-term chronic neuropathic pain and accumulating research indicating that harms, such as addiction, increased hospitalisations and mortality, outweigh the benefits. Audit and feedback is a frequently used, moderately effective, population-level quality improvement approach that aims to improve patient care by reviewing health care performance against explicit standards. We investigated whether an enhanced feedback intervention, comprising comparative and personalised practice feedback, embedded with evidence-based behaviour change techniques, reduced gabapentinoid prescribing in primary care.

Approach

We delivered a feedback intervention to all 310 practices within West Yorkshire, bimonthly from January-December 2018. 373 practices across the wider Yorkshire and Humber area form our control group. We compared gabapentinoid prescription rates in intervention and control practices for five pre-intervention, one intervention and two post-intervention years using publicly available prescribing data and a controlled interupted time-series analysis. We also assessed potential wider impacts on primary care prescribing for pain (opioid and non-steroidal anti-inflammatory drugs (NSAIDS)), whether practice or patient characteristics are associated with any intervention effects and the impact of reclassification of gabapentinoids as class C drugs introduced after the intervention (April 2019).

Findings

Intervention and control practices differed significantly pre-intervention in deprivation scores, percentage of patients employed, percentage of patients with long-term conditions and Quality Outcomes Framework achievement. Rates of gabapentinoid, opioid and NSAID prescribing were lower in the intervention group compared to the control group at all time points. There was no difference between rates of gabapentinoid prescribing between intervention and control group during or post- intervention (p=0.71). Full analysis to assess the impact of regulatory changes and associations between practice or patient characteristics is awaited.

Consequences

This feedback intervention did not change the underlying trend of gabapentinoid prescribing compared to control. We suggest five potential explanations: i) Prescriber awareness of upcoming changes in drug classification reducing intervention impact; ii) The intervention began six months after a similar intervention to reduce opioid prescribing, leading to competing priorities; iii) Relatively small numbers of patients taking gabapentinoid medications (compared to opioids) making it more challenging to reduce prescribing; iv) Publicly available data includes ‘noise’ (e.g. gabapentinoids prescribed for other indications); and v) Significant differences between intervention and control practices pre-intervention.

Ideally, head-to-head comparisons of different ways of delivering feedback within randomised designs to increase effectiveness of feedback interventions are needed. External contexts, competing demands and patient influences remain important challenges to quality improvement in prescribing. Addressing the rise of gabapentinoid prescribing is likely to require sustained, coordinated efforts across all levels of healthcare systems and target organisational, clinical and patient behaviours.