Improving the care of diabetes patients at City Hospitals Sunderland - a joint working initiative with Lilly

New collaborative methods of working are required if we are to deliver timely and effective care to patients with diabetes across acute hospitals. Against a backdrop of rising prevalence, greater demand upon secondary care services with increasing costs incurred by both Foundation Trusts and local CCGs due to extended length of stay and higher readmission rates; we embarked upon a joint working agreement with Eli Lily & Co to develop a proactive IT enhanced project to address key aspects of inpatient diabetes care.

The cornerstone of the project was to develop a method that facilitated a proactive clinical review of all adult inpatients (presenting with diabetes as either a primary or secondary reason for admission) admitted to City Hospitals Sunderland NHS Foundation Trust by a dedicated inpatient diabetes team.

Our primary objectives were

  1. To utilise the current overarching IT systems ‘Meditech’ (electronic medical records) and ‘Cobas IT1000’ (web-based blood glucose monitoring system) to develop a method that identified, tracked and allowed remote monitoring of patients from admission to discharge.
  2. Enhance the clinical team and support structure for data acquisition. The extra clinical support needed to implement effective use of the Meditech system was funded by Eli Lilly under a Joint Working Agreement. Eli Lilly also provided project management and data support.


In 2014 City Hospitals Sunderland upgraded their Trust-wide IT system ‘Meditech’ to incorporate Electronic Medical Records (EMR). This upgrade presented the opportunity to develop an admission document for both elective and emergency admissions which allowed the simple identification of those patients with diabetes as a long-term condition, their location within the hospital and importantly treatment (insulin or oral agents) as part of a targeted risk stratification process. Combining this information as a daily diabetes electronic report alongside interrogation of the ‘Cobas IT1000’ glucose monitoring system alerts the diabetes inpatient team whenever a patient with diabetes is admitted into the hospital and triggers a clinical review, either remotely or face-to-face dependent on need or risk.

The clinical reviews were designed to enhance the quality of service, by identifying and addressing gaps in obtaining the eight national key care processes for all patients with diabetes, improving patient satisfaction and diabetes education, ultimately promoting early safe discharge thus improving length of stay (LOS) and reduce readmission rates in particular for a subgroup of patients admitted with Diabetic Ketoacidosis (DKA).

It was predicted that the project would support a recent pathway redesign for those patients with type 1 diabetes admitted with DKA. This pathway recognised the three dimensions of diabetes care (pathological, psychological and social), and attempted to provide a coordinated identification, referral and support service between inpatient specialist nurse, clinical psychology or inpatient mental health liaison team and social worker as clinically indicated with outpatient follow-up by the clinical team at two weeks post discharge. There was an assumption that due to the extra interventions in this patient group, LOS may increase. Therefore the measurable objective with these patients was a percentage reduction in all cause readmissions. In addition, this patient group was asked to complete a series of surveys about the quality of the care they received as an inpatient. A new patient satisfaction questionnaire was developed for this by amending the previously used Friends and Family Test.

The secondary objectives for the project set out in the JWA were as follows:

  • Reduction in hospital length of stay for patients with diabetes
  • Reduction in readmissions for t1 patients with DKA
  • Reduction in hospital errors associated with insulin
  • To improve patient safety linked to medication errors
  • Increase in patient satisfaction with the hospital diabetes service
  • Medicines optimisation to improve blood glucose control leading to a delay in the onset of complications which can impact upon a patient’s quality of life
  • Improved efficiency and capacity within the service
  • To understand patient flow through the system and identify and eradicate blockages to patient care leading to improved efficiency
  • To highlight areas of best practice to share with colleagues locally and nationally.

The expected benefits to Eli Lilly were set out in the JWA as follows:

  • Data associated with this project will enable Eli Lilly to better understand its NHS customers and prescribing patterns within NHS Sunderland CCG. This potentially would highlight an additional need for support around training and education within the account and an opportunity for Eli Lilly to support
  • CCGs are classified nationally against populations within similar groupings. NHS Sunderland CCG falls within the orange CCG classification group. There are 91 CCGs nationally with the same demographic. Due to this alignment there is the potential that this data could be transferable to similar accounts. This potentially could create additional opportunities for Eli Lilly by sharing findings with similar accounts nationally via publication of the project once complete
  • Utilisation of Eli Lilly projects as per local formulary guidelines – right patient, right place, right time
  • Opportunity to utilise an innovative approach to service improvement and share learning in other accounts on both regional and national level

In December 2014 a Joint Working Agreement (JWA) was entered into with Eli Lilly to support the initiative. Eli Lilly provided funding for one additional Diabetes Specialist Nurse (DSN) for a period of 12 months, as well as online tools for data collection and approximately six days per month project management support via two members of its Corporate Affairs team. An additional DSN was recruited to the team in June 2015 which was when the project went live.

Baseline inpatient activity, as measured by ICD-10 coding, was collected by City Hospital Sunderland, and cross referenced with data provided by Eli Lilly via a Health Intelligence Tool under licence from Harvey Walsh. Data for all patients with diabetes admitted with diabetes as a secondary code for admission was collected, along with the eight most common diabetes primary admission codes. Codes for primary admissions were chosen by maximum patient number, the cut off being less than 10 admissions per 12 month period. Codes for foot- or eye-related complications were also omitted as out of scope of the project. The eight codes that data was collected for were as follows: E101, E111, E115, E116, E119, E160, E161, E162.

A stakeholder group was established at project conception to scope, design, run and measure the results of the project utilising Prince 2 principles and methodology. The group included clinical, nursing and management staff from CHS, and two quality improvement consultants from Eli Lilly.

Twelve month data showed that for patients with diabetes from Sunderland CCG, admitted to a non-specialty ward of Sunderland Royal Hospital the tariff cost per patient decreased from £1996 (June 14 – May 15) to £1876 (June 15 – May 16). This corresponds to a reduction in average LOS from 4.23 days (range from 3.67 – 4.72) to 3.68 days (range 3.08 – 4.37) This is an average reduction of 0.55 days per patient. This corresponded to a reduction in average monthly bed days from 2590 to 2490, a saving of 100 bed days per month.

Data from the National Diabetes Audit2 confirms high attainment of the eight key care processes for both patients with type 1 (61 percent CHS vs 37.5 percent national average), and type 2 (65.2 percent CHS vs 53.9 percent national average) diabetes.

Data associated with ‘Harm resulting from inpatient stays’ were derived from the National Audit of Inpatient Diabetes care (NaDIA3). Medication Errors: 25.8 percent of patients with diabetes experienced one or more medication error at CHS vs 37.8 percent national average. Medication management errors: 19.1 percent of patients with diabetes experienced at least one medication management error at CHS vs 24.1 percent national average. Insulin Errors: Of the patients on insulin, 19.1 percent experienced one or more insulin (prescription or medication management) error at CHS vs 22.7 percent national average. Prescription Errors: 9 percent of those patients admitted to CHS experience one or more prescription error vs 21.1 percent national average.

In addition to this, results from a subgroup of patients admitted with DKA were as follows: Baseline a total of 108 admissions with 20 percent re-admission rate over a 12 month period. Post-intervention a total of 86 admissions with 15 percent readmission rate over 12 months. Key care processes attainment: Baseline 38 percent of patients admitted with DKA had data collected. Post-intervention 65 percent. Attendance to outpatient clinic 2 weeks post discharge. Baseline 0 percent, post intervention 50 percent. Treatment satisfaction questionnaires and Friends and Family tests confirm a high level of satisfaction with the new method of working.

When grouped together, data from the pre-agreed primary admission codes (E101, E111, E115, E116, E119, E160, E161, E162) showed an average LOS reduction from 6.29 days per admission to 5.6 days per admission. A reduction of 0.69 days, per admission.

Following initial results, a business case was submitted to the Hospital Trust to secure ongoing funding for the DSN post. This application was successful.

Patient privacy

No patient-identifiable data was provided by or shared with Eli Lilly. All patient satisfaction questionnaires were anonymised before sharing. ICD-10 data provided by Eli Lilly was done so via a tool commissioned from Harvey Walsh with a commercial reuse licence for this data.

Lessons Learned

A proactive opportunistic approach to enhancing care is achievable and requires a change in team mindset along with a supportive internal and external management structure.

Specific patient questionnaires should be piloted in the general and specific subset of patients with and clear direction on how, where, and when these should be completed. Ideally patients should complete these where possible without the support of nursing staff as this can alter results.

Errors related to insulin (national average rate 22.7 percent vs local 19.1 percent) and prescribing (national average rate 21.1 percent vs 9 percent local rate) although common, were not easily captured by this work.


This project model is reproducible across other Foundation Trusts who have, or wish to develop, an Inpatient Diabetes service which proactively reviews patients, seeing each admission as an opportunity to improve glycaemic control, whilst actively looking for gaps in care, data collection and patient knowledge.

Key Points

Regardless of reason of admission, patients with diabetes admitted to hospital benefit from an intervention by a diabetes clinical specialist team.

  • Investment in extra clinical support for inpatients with diabetes leads to savings in LOS and associated tariff for the CCG
  • Investment in a coordinated multidisciplinary fast track pathway for patients admitted with DKA improves outcomes and reduces readmissions for this group of patients
  • Joint-working with pharma provided valuable project management and monitoring as well as financial input.

Conflict of interest

Dr Nayar has received honoraria from Eli Lilly for work unrelated to this project.

Acknowledgements: Thanks to Stephanie Edgar – quality improvement consultant, Eli Lilly, Sarah Landells – quality improvement consultant, Eli Lilly, Melissa Robinson – diabetes specialist nurse, City Hospitals Sunderland, and Jackie Burlinson – ex-directorate manager, City Hospitals Sunderland. This article was unsolicited.