The Diabetes care in Elderly, Frail, ENd-of-life and Dementia (DEFEND) project


Project DEFEND was established to improve the quality of diabetes care and safety in Brent’s frail adult community. It aligns local pathways and resources to create a data-driven collaborative care model. The databases of five GP practices highlighted that, out of 1,000 patients with frailty, 695 had an HbA1c lower than 58 mmol/mol. Of these, 100 patient records were reviewed in a virtual clinic with primary care clinicians and grouped into high, medium or low risk of diabetes emergencies. More than 40% were in the medium or high-risk group, over 60% needed a medication change and nearly 20% needed referral to supported care.

Get the latest updates


Searches were conducted in five GP practice EMIS systems to find people with diabetes aged over 65yrs, including HbA1c, frailty, medication and dementia. This resulted in 1,000 patient records, which were grouped in to high, medium or low risk categories, based on the following criteria. The group termed High Risk were on the following medication: Sulphonylurea/glinides/Insulin/polypharmacy. They had HbA1c above 100 mmols/mol and less than 53 mmols/mol, severe frailty, an eGFR <30 and were on metformin; they had dementia/end of life/other severe comorbidity and a recent ambulance or 111 call out (due to hypoglycaemia or hyperglycaemia), a recent admission for Hypoglycaemia, frequent telephone consultations, DNA appointments, homeless or living alone, housebound and with safeguarding concerns. The Moderate Risk group were on: Sulphonylurea/Insulin/ Polypharmacy. They had a HbA1c of 53 to 99 mmols/mol and moderate frailty. Their eGFR was <30 and they were not on metformin. They had DNA appointments, were homeless or lived alone, were house bound – with District Nurse (DN) visits and had had no review in the past year. The Low Risk group’s medication criteria was that they were on dual oral therapy, not including medications causing hypoglycaemia, and not on insulin. Their HbA1c was 58 to 70 mmols and frailty was mild. They did not have the other conditions included in the moderate and high-risk groups. Next, a virtual clinic at a GP practice risk stratified 100 patients into low, medium or high risk based on HbA1c (lower than 58 or higher than 75) and medications that can cause hypoglycaemia. Relaxed HbA1c targets were recommended with suitable medication changes. The GP EMIS system was used to set up and run searches for Frailty, Medication, HbA1c, Dementia, Heart Failure. The 100 records were reviewed in a GP practice via a virtual clinic session and the data were analysed and risk stratified. The data were reviewed and discussed with the team. Meetings were held with DNs and the Complex Patient Management Group (CPMG) to review patients and develop a seamless pathway to manage this vulnerable group.


Of the 100 records reviewed, 22% were high risk, 22% were medium risk and 57% were low risk. The number of people with HbA1c was lower than 58: 81%. The medication review and changes resulted in 62% having a metformin dose adjustment due to low eGFR, 19% had their sulphonylurea dose reduced or stopped, and 19% were referred to DNs or the CPMG for supported care. These results highlight the gap in the care for this vulnerable adult group who are exposed to direct risks from diabetes-related complications or inappropriate use of medications, and to indirect risks because of age-associated frailty, dementia or other chronic health conditions. Just over 80% had HbA1c lower than the recommended range for this group and more than half of them needed an intervention, either in the form of medication change or supported care. These practice-changing results led to incremental implementation of targeted searches and interventions in GP practices across Brent by BIDS nurses. In addition, wider dissemination of the results to primary care colleagues resulted in proactive searches and medication changes by the clinicians working in primary care.

Data collection and analysis was paused because of COVID-19, but the learnings from the above data have been applied in the day-to-day virtual consultation model.

Sustainability and Spread

Learnings from this project have already been incorporated into daily clinical practice in the form of a focused approach to the high-risk adult group with frailty, dementia or at end-of-life. Dissemination of these results by nurses, via GP communication channels and one-to-one visits (face to face or virtual), have led to proactive implementation of risk stratification, safer HbA1c targets, mitigation of hypoglycaemia risks and safer diabetes medication use in people with lower eGFR in primary care. This data-driven model can be implemented by primary care with some support from specialist diabetes teams working in the community, which is becoming a standard model in the UK. Therefore, the learning from this project should be scalable and sustainable, as well as cost-effective in the shorter and longer term. Expansion plans were limited by the pandemic, but plans are in place to review the data and share it with peers in due course. The WHAT next plan is to: Work with other stakeholders like CPMG, Integrated Care Practice and Primary Care Networks on virtual clinics, care planning, and user-friendly functional integration pathways; Have planned PDSA cycles to improve services; Audit to learn from dashboard and survey data for individual and population level care planning; and develop a better Type of education with learning modules for patients, carers and HCPs.

Back to the top

QiC Diabetes Finalist 2020
Patient Care Pathway, Secondary and Community
The Diabetes Care in Elderly, Frail, ENd-of-Life and Dementia(DEFEND) Project
by London Northwest University Healthcare NHS Trust


Quality In Care Diabetes

QiC Diabetes is partnered by:
Supported by:
  • 2022 KEY DATES
  • Open for entry:
    Monday 28 March 2022
  • Entry Deadline:
    Friday 1 July 2022
  • Judging day:
    Tuesday 6 September 2022
  • Awards ceremony:
    Thursday 13 October 2022