5 minutes with… Anna Sallis

Behavioural Insights Research Advisor at Public Health England; now with March 2019 update.

A recent trial involving more than 1500 GP practices found that writing to GPs about their antibiotics prescribing resulted in 73,000 fewer prescriptions over six months. The trial (see tinyurl.com/hwoagnr), a collaboration between Chief Medical Officer Dame Sally Davies, Public Health England (PHE), and the Behavioural Insights Team, was part of the government’s plans to slow the growth of antimicrobial resistance. We spoke to Anna Sallis, a health psychologist at the PHE Behavioural Insights Team, for more on the trial and what the future holds for the team.

Why did you have an interest in working in the Public Health England (PHE) Behavioural Insights Team?
The potential to design and conduct robust behaviour change interventions that can have an immediate and widespread impact on supporting healthy choices attracted me.

I trained as a health psychologist whilst working as a government social researcher and later as a Senior Psychologist at the Department for Work and Pensions. My work involved applying health psychology theory and evidence to policy in health and work, sickness absence and welfare reform. Before this I worked at the Maudsley Hospital evaluating staff prevention and management of violence training.   

Around the time I qualified in 2011, the Cabinet Office Behavioural Insights Team was becoming big news across government. This opened up many opportunities to apply behavioural science to policy, and I moved to the Department of Health to help set up their internal Behavioural Insights Team. Soon after this team was established and we had plenty of trials up and running, I moved on to become the expert adviser to the new PHE Behavioural Insights Team. I’m also on the British Psychological Society’s Behaviour Change Advisory Group, and am Policy Officer for the BPS Division of Health Psychology.

Can you give me a little background in how this trial on social norms feedback came about?
PHE leads implementation streams of the cross-government UK Five Year Antimicrobial Resistance Strategy with a remit to facilitate a reduction in total antimicrobial prescribing in primary care to 2009/10 financial year levels. Members of the research team had been involved in numerous policy trials demonstrating the impact of both behaviourally informed letters and social norms feedback on both health and economic outcomes. We know that social norms act as a marker for social comparison against which individuals evaluate the appropriateness of their own behaviour compared to others. Observed discrepancies then motivate the individual to change their behaviour to be in line with their peers.   

Although medical practices have access to their own and others’ prescribing data, we do not know how many practices actively look at this and how much attention is paid to the information. Feeding back this data directly to named prescribers using a high-profile messenger (England’s Chief Medical Officer, Dame Sally Davies) highlights not only how GPs compare with others, but also that others can and do use it to monitor prescribing behaviour. To move the intervention from a passive letter to an active intervention, we included ‘behavioural instruction’ in the form of three simple, concrete actions that the Chief Medical Officer recommends GPs take in order to reduce their prescribing levels (give patients advice on self-care instead, consider offering a delayed prescription instead, and talk to other prescribers to ensure they are also acting).

What’s the future of the antimicrobial resistance work the team is doing?
We have a range of projects under way to deliver aspects of the UK Five Year Antimicrobial Resistance Strategy. These include translating the positive evidence from our social norms feedback randomised controlled trial (RCT) into routine practice, and we have been working with NHS England and the NHS Business Services Authority to send out similar letters in winter 2015/16 to all GPs in practices with high antibiotic prescribing rates.

We are currently implementing a cluster RCT with over 200 GP practices to test the impact of two interventions aimed at reducing patient demand for antibiotics and increasing GP commitment to not prescribing antibiotics when they are not clinically indicated. I am also leading a review of primary care antimicrobial stewardship policies and programmes; the aim of the review is to classify the interventions into the Behaviour Change Wheel set out by Susan Michie and colleagues, to identify gaps and opportunities for policy.

We are also involved in wider PHE work contributing to projects led by others, including the evaluation of evaluating the impact of Antibiotic Guardian (a pledge-based campaign aimed at both public and healthcare professionals to raise awareness of antimicrobial resistance), interviewing community pharmacists about antimicrobial resistance and designing an intervention to reduce inappropriate antibiotic prescribing in out-of-hours services.


In March 2019, our journalist Ella Rhodes caught up with Anna Sallis again to hear about the latest developments.

With regards to reducing AMR in healthcare, in what context does behaviour change need to happen?  

We need a whole systems approach covering general practice, secondary care, pharmacy, nursing homes, out of hours care, dental etc with change needed from patients and healthcare professionals as well as providers and commissioners. Our team work on two key areas for reducing AMR in a healthcare context: infection prevention and control (IPC) and antimicrobial stewardship (AMS). Preventing infections occurring in the first place, in particular, healthcare associated infections, reduces the need for antibiotics. Healthcare professionals (HCP) also play a key role in stewardship of existing antibiotics through responsible prescribing and provision of self-care advice, but patients also need to protect themselves and future generations by managing symptoms of self-limiting infections (such as sore throat) without antibiotics.  

Could you tell me about how your work with GPs progressed after we spoke in the spring of 2016? 

Yes, Natalie Gold is now leading this work for our team and will tell us more. 

“We have sent a letter from the CMO to GPs in practices with high antibiotic prescribing rates every winter since the original trial in 2016 (Hallsworth et al., 2016). We analysed the data from the winter 2016/7 letter using a Regression Discontinuity Design, which can identify the causal effect of the intervention even without a trial, by comparing practices who were just above and just below the threshold for receiving a letter. We found that there was a reduction in prescribing of 3.7%, which equates to 124,952 fewer antibiotic items dispensed over a six-month period and an estimated saving of £220,836 in direct prescription costs alone. This paper is under review. 

In winter 2017/8, as well as sending the standard letter to high prescribers, we ran a trial, sending letters to GPs in practices, not in the top 20% of prescribers, but whose prescribing rate increased by more than 4% in the past year. We expect the results of this trial in the coming months. 

In November 2018, we ran another set of trials, to determine the effect of: (i) adding a bar chart and the practice’s specific prescribing percentile to the standard letter sent to the top 20% of practices; (ii) targeting broad spectrum prescribing instead of overall prescribing for practices who are amongst the highest prescribers both overall and of broad spectrum items; (iii) sending letters to practices with moderate prescribing and high broad spectrum prescribing, including a bar chart and feedback on the percentile the practice is on for broad spectrum prescribing. Results are pending.”

Are there other behaviour change or behaviourally focused projects going on in the team? 

Yes. Our team are involved in, or leading, several projects on AMR and there are multiple other projects ongoing across PHE. We are working on two strategic behavioural analysis and implementation projects – one on reducing inappropriate antibiotic prescribing in primary care and one on HCP behaviours related to the prevention of Catheter-Associated Urinary Tract Infections (CAUTI). We are working with the UCL Centre for Behaviour Change (CBC) and the Nuffield Department of Primary Care Health Sciences, University of Oxford. 

In these two projects, the Theoretical Domains Framework (TDF) (Cane, O’Connor, & Michie, 2012) is used to classify influences on behaviour (derived from the literature) and the Behaviour Change Wheel (BCW) (Michie, Van Stralen, & West, 2011) and Behaviour Change Technique Taxonomy V1 (Michie et al., 2013) are used to describe the content of national interventions aimed at addressing these behaviours. Using mapping matrices, we can see whether existing national, or widely implemented local, interventions are addressing influences on the behaviours they are trying to change with the aim of improving their behaviour-change content. Using this methodology, we can get a good picture of what we are doing nationally to address a problem and understand how best to distribute or redistribute resources to tackle it. 

The next stage for both projects, known as ENACT – ‘Exploring the implementation of iNterventions to reduce Antibiotic use and Catheter associated urinary Tract infections’ – is to consider how best to optimise the suite of national interventions by proposing intervention components that address BCTs and intervention functions which are currently not being targeted. This stage of the project engages with intervention users and commissioners to ensure the relevance of BCTs in context using the APEASE criteria. 

We have also been collaborating with the PHE Primary Care Unit on (i) a qualitative TDF study on the experiences and perceptions of community pharmacists and their teams around AMS activities (Leah Ffion Jones et al., 2018) (ii)  diagnostic tools for Urinary Tract Infections (UTI) and their management, and a UTI leaflet for older adults to be used by care staff, clinicians, patients and their relatives (Jones, Cooper, & McNulty, 2018) and, with Lead Pharmacist Dr Diane Ashiru-Oredope, we have been working on a cluster RCT assessing the impact of training pharmacy teams in antimicrobial stewardship (AMS) and use of the TARGET leaflet to manage self-limiting infections.

Where can psychology/psychologists be particularly useful in tackling large societal problems such as AMR? 
Application of the BCW and related tools using the Strategic Behavioural Analysis methodology is a great example of how we can use psychological methods and thinking to make sense of complex behaviour systems. In the process of conducting this work we have identified a number of research gaps which psychologists could fill for example on HCP CAUTI prevention behaviours in primary and community care. There are also still many under-researched areas around the use of diagnostics in antibiotic prescribing practice – particularly as new diagnostics are developed. These raise questions of fidelity to guidelines for their use and public perception and reaction. 

Although the trend in prescribing of antibiotics is moving in the right direction, unnecessary prescriptions are still commonplace and there is huge variation in practice that is not well characterised. There is also a paucity of strongly evidence-based interventions that can be affordably delivered at scale with impact. This provides a challenge in the protection of medical care for future generations that psychologists are well-placed to tackle collaboratively with other disciplines. 

- See also the recent article on the social dilemma of antibiotic use.


Cane, J., O’Connor, D., & Michie, S. (2012). Validation of the theoretical domains framework for use in behaviour change and implementation research. Implementation science, 7(1), 37. 

Hallsworth, M., Chadborn, T., Sallis, A., Sanders, M., Berry, D., Greaves, F., . . . Davies, S. C. (2016). Provision of social norm feedback to high prescribers of antibiotics in general practice: a pragmatic national randomised controlled trial. Lancet, 387(10029), 1743–1752. doi:10.1016/S0140-6736(16)00215-4

Jones, L. F., Cooper, E., & McNulty, C. (2018). Urinary tract infections (UTIs); a leaflet for older adults, and carers: the development of a UTI leaflet for older adults and their carers. . British Journal of General Practice,, 68(suppl 1), bjgp18X69683. doi:doi:10.3399/bjgp18X696833

Jones, L. F., Owens, R., Sallis, A., Ashiru-Oredope, D., Thornley, T., Francis, N. A., . . . McNulty, C. A. (2018). Qualitative study using interviews and focus groups to explore the current and potential for antimicrobial stewardship in community pharmacy informed by the Theoretical Domains Framework. BMJ open, 8(12), e025101. 

Michie, S., Richardson, M., Johnston, M., Abraham, C., Francis, J., Hardeman, W., . . . Wood, C. E. (2013). The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behavior change interventions. Annals of Behavioral Medicine, 46(1), 81-95. 

Michie, S., Van Stralen, M. M., & West, R. (2011). The behaviour change wheel: a new method for characterising and designing behaviour change interventions. Implementation science, 6(1), 42.

BPS Members can discuss this article

Already a member? Or Create an account

Not a member? Find out about becoming a member or subscriber