How to design safer asthma care from the frontline

Can real-time prescribing intelligence help primary care prevent asthma crises? A neighbourhood programme in East London suggests it can. Academic GP Dr Anna De Simoni, co-author of the research, explains how.

Preventable asthma exacerbations continue to place avoidable pressure on GPs, urgent care and hospitals. In primary care, this presents as repeated requests for reliever inhalers and missed reviews, while in hospitals, it appears as emergency attendances and admissions that could have been prevented upstream.

The publication of the NHS 10-Year Health Plan makes this particularly urgent, as it highlights the gap between the national ambition for preventative neighbourhood care and the reactive systems in which clinicians currently work. 

In East London, hospitalisation for acute asthma is 14% above the average for London, with admissions rising from 1.3 to 7.5 per 100 asthma population as the number of SABA inhalers prescribed rises from one to three to more than 12 a year.

Rates of SABA over-prescribing remain high in this multiethnic, deprived urban population, with more than 30% of patients on average prescribed six or more SABA inhalers in the previous year, and significant variation between practices.

A primary care-led neighbourhood asthma programme was set up to address this.

Aims

This quality improvement programme aimed to enable primary care clinicians to focus management attention on patients with asthma who were at highest risk of hospital admission. It was part of Real Health, a programme funded by Barts Charity. The programme was endorsed by the respiratory lead for Newham and by the ICB, but there were no financial incentives.

The programme was led by academics in the Clinical Effectiveness Group (CEG), part of the Centre for Primary Care at the Wolfson Institute of Population Health at Queen Mary University of London. It was a collaboration between academic GPs, frontline clinicians, data scientists, and a hospital respiratory consultant, with a shared aim of designing a system that would make safer asthma care the default, not the exception. 

The programme was deliberately designed with replication in mind, rather than as a one-off local success.

Approach

The programme involved designing a learning health system embedded within routine general practice. We hoped that by integrating real-time prescribing intelligence into electronic health records (EHRs), we could support proactive, prevention-led asthma management at neighbourhood scale. 

Input from Prof Paul Pfeffer, a hospital consultant and clinical professor of respiratory medicine, ensured that prescribing safety, follow-up after exacerbations and shared accountability across settings were built into the model. He also helped shape webinar content and the clinician guidance we developed as part of the programme.

The initiative, which took place between October 2020 and March 2023, involved all 48 GP practices in 10 PCNs in Newham, East London, with a patient population of approximately 400,000.

The intervention combined three core components.

There were prescribing alerts embedded in GP systems that flagged patients who were prescribed more than 12 SABA inhalers in a year. When initiated by the practice, the system generated lists of patients for review.

Secondly, there was neighbourhood-level collaboration, enabling practices to act collectively rather than in isolation.

And, thirdly, there was a multidisciplinary approach to quality improvement, involving GPs, pharmacists and practice teams. We ran 21 webinars across PCN and practice meetings, and nurse and pharmacist forums, attended by 326 staff in total. We also produced clinician guidance and electronic patient information leaflets.

Prescribing data were collected from electronic health records, and SABA overprescription was evaluated through interrupted time-series analysis.

Outcomes

The pop-up prescribing alerts were associated with a 50% reduction in overprescribing of SABA in the subsequent year among patients who received an ‘active response’, with 221 of 442 dropping to fewer than 12 inhalers. An active response was defined as an action other than cancelling the alert. This included inviting the patient for a review via an automated SMS or email, or completing a medication review. 

This reduction was accompanied by a corresponding increase in appropriate maintenance inhaled corticosteroid (ICS) prescribing, supporting a shift from crisis-driven prescribing towards more precise preventative asthma care.  

We estimate that the approach could result in an absolute reduction of hospital admissions of 11%. This was a projected estimate based on findings from an earlier study in Newham, applied to the local population before and after the intervention. However, during the three years of the project, practices in the study area undertook a range of incentivised activities connected with limiting excessive SABA prescribing. It was not possible to disentangle the contribution of our programme from other concurrent initiatives.

The electronic alerts were only acted upon in 17% of cases. The intervention coincided with significant operational pressures due to COVID-19, which may have been a factor. In addition, 52.2% of alerts were seen by receptionists and admin staff, whose most common recorded action was to invite patients for an asthma review. The tool shows that they can contribute to improving outcomes for asthma patients at highest risk of hospital admission.

One of the clearest findings was lower variation in areas where practices worked more closely together, suggesting they tended to implement change faster and more consistently. Shared learning, peer comparison, and collective problem-solving enabled practical changes - such as removing SABAs from repeat prescription lists - and embedding pharmacist-led asthma reviews. 

Qualitative feedback showed that the intervention was well received by primary care teams. As it was co-designed by clinicians and embedded in existing workflows, it was experienced as supportive rather than as an additional administrative burden. 

Future

Our programme demonstrates that when primary care is equipped with real-time data, integrated digital tools, and collaborative structures - and when it is designed in partnership with secondary care - we can reduce avoidable harm, improve patient safety, and relieve pressure on hospitals. 

The approach was adopted nationally in a modified form by health innovation organisation UCLPartners in its Proactive Care Frameworks initiative and the software tools have been used in other regions. In Newham, it was left to individual practices to decide whether to continue with the initiative. 

Our experience suggests that scaling this model would require access to high-quality, near-real-time prescribing data and seamless EHR integration, so that it operates within routine consultations. It also needs neighbourhood-level data-sharing agreements to support population health management. Protected time and support for multidisciplinary quality improvement are also requirements, as well as alignment with national respiratory priorities (see box). 

Without these enablers, the benefits will remain confined to small pockets of innovation rather than becoming standard NHS infrastructure. 

Dr Anna De Simoni is a clinical reader in Primary Care Research at Queen Mary University of London, and co-author of Reducing short-acting beta-agonist overprescribing in general practice: Evaluation of a quality improvement programme in East London.