Nudging: can letters to GPs reduce antipsychotics in aged care?

JUST before Christmas 2019, GPs received a letter from the Department of Health that caused concern.

The Interim Report of the Royal Commission into Aged Care Quality and Safety, titled Neglect, identified the overuse of chemical restraint as requiring urgent action. In December 2019, the Department of Health announced that GPs who are high prescribers of antipsychotics and benzodiazepines in residential aged care facilities (RACFs) would receive a letter containing data on the individual GP’s prescribing patterns in comparison to their peers.

The announcement of this strategy, known as a “social norm feedback nudge”, quickly prompted a negative response from GPs and people advocating for better care in RACFs. Previous nudge letters have also been met with scepticism from GPs.

Can this approach work, and why is it causing so much controversy? In this article we explain the theory, evidence and controversy behind the use of social norm feedback nudges, and consider whether targeting GPs can change practices in RACFs.

What is a nudge?

Social norm feedback influences our behaviour by using our inherent desire to conform to societal norms. Essentially, underperformers are notified of their performance relative to their peers. Social norm feedback belongs to a suite of behavioural economics strategies that aim to exploit our cognitive biases in decision making, in order to guide us to the optimal decision – called nudges. Collectively, nudges are attractive as they are usually low cost, simple interventions that do not involve legislation or financial incentives.

Can nudges change prescribing behaviour?

Nudges have mostly been applied in areas other than health care, but their use is growing (here and here). They have been used in trials aiming to reduce antibiotic prescribing in primary care in Australia and England. GPs who were high prescribers of antibiotics were notified via letter that they prescribed antibiotics at rates higher than their peers (eg, higher than 80% of their peers). Both trials showed a reduction in antibiotic prescribing.

However, some key questions about social norm feedback remain unanswered. These include how long the intervention effect persists and whether the effect diminishes with repeat messages. Furthermore, in complex care settings such as RACFs, it is unclear who the ideal targets of the nudge should be and if such interventions can cause unintended harm.

What are the criticisms of nudges in primary care?

After the perceived success of the antibiotic prescribing letters, the Department of Health applied this strategy to reduce opioid prescribing in primary care. Feedback from GPs provided through the Royal Australian College of General Practitioners identified several concerns regarding the letters. First, the data used were not risk-adjusted to take into account the types of patients in the GP’s care, which can be a key driver of prescribing. For example, GPs caring for patients in palliative care would have appropriately higher opioid prescribing rates. The data need to be thoughtfully adjusted for factors that explain warranted variation and this should be transparently communicated to enhance trust and perceptions of fairness.

The second criticism was that the letters had a threatening tone. Social norm feedback nudges influence our behaviour by harnessing our desire for our practices to align with that of our peers, not by serving as “sticks”. Use of negative personal consequences or threats is not necessary and may be undermining. Instead, there is evidence that providing information on alternative options may augment the intervention.

Third, the opioid letters may have had a negative impact on access to services and the limited evidence has not established if this was the case. GP visits to RACFs have been in decline, and monitoring potential negative consequences is crucial to avoid undermining care delivery in RACFs.

All of the above factors illustrate the need to engage GPs in the intervention design to ensure acceptability.

Can nudges improve prescribing in RACFs?

Targeting only GPs may not change practice in RACFs, as facility staff influence decisions on the use of chemical restraint. Nudges have not yet been employed in RACFs. We do not know if nudges in RACFs need to target GPs, facility staff or both. Medication management in RACFs is a complex process involving many stakeholders in addition to GPs, and thus it is a different prescribing context than outpatient primary care. Furthermore, fundamental changes to the current model of care are needed to address system factors that lead to high rates of chemical restraint, including staffing levels, resources to implement non-pharmacological management for residents living with dementia, and GP visits to facilities.

Conclusion

Nudges are a promising approach to optimising prescribing practices. Care must be taken to ensure that the strategy is applied with risk-adjusted data and following nudge principles. Many important questions about the optimal design of this approach in complex care settings such as RACFs remain. What is clear is that the engagement of GPs and facility staff in the design process can only be positive.

Dr Magdalena Raban is a Senior Research Fellow with the Australian Institute of Health Innovation at Macquarie University. She holds a PhD and Masters in Public Health, and a Bachelor of Pharmacy, from the University of Sydney. Magda practiced as a community and hospital pharmacist in Australia and the UK for approximately 10 years, before working in research. Dr Raban’s recent research has combined her quantitative, epidemiology and clinical experience in focusing on medication safety, evaluation of health system interventions, and the use of information technology to improve the quality use of medicines.

Dr Kimberly Lind is a Research Fellow with the Australian Institute of Health Innovation at Macquarie University, where she specialises in aged care research. She holds a PhD in Health Services Research & Policy, MPH in Epidemiology and BS in Psychology, Ecology & Evolutionary Biology. Her expertise is in observational study design and secondary data analysis of claims and electronic health record data. Her research interests broadly include health outcomes, health care utilisation and health equity among older adults. She has worked on a wide range of research studies from randomised controlled trials to policy evaluation in the US and Australia.

The statements or opinions expressed in this article reflect the views of the authors and do not represent the official policy of the AMA, the MJA or InSight+ unless so stated.