Evaluation of newly-diagnosed patients from 2018 at Royal United Hospital NHS Foundation Trust demonstrated glycaemic outcomes deteriorated significantly between three months and one year post diagnosis. The team joined Wave 3 of the Royal College of Paediatrics and Child Health (RCPCH) National Diabetes Quality Improvement Collaborative in November 2018 and focused on improving median and mean HbA1c in this period post diagnosis in newly-diagnosed children with Type 1 diabetes. The new care pathway emphasised intensive management and consistency of team message from all staff involved. Carbohydrate counting and Expert meters were introduced on the ward at diagnosis and patients home downloaded. Initial results showed improvement in average blood glucose (BG) levels at 28 days post diagnosis and HbA1c at three months. Patient feedback was positive and all patients were downloading BG meters regularly at home.
Get the latest updates
A whole-team approach was adopted to implement change. Quality Improvement (QI) methodology was incorporated into daily work via weekly QI meetings. Meetings were short and focused on reviewing project progress, task allocation and identifying QI bright spots and dark spots. Actions and task allocation were recorded in a QI action log and reviewed at each meeting. Maintaining run charts of average BG data and HbA1c for new patients was a powerful team motivator and identified patients requiring additional support. A QI board in the office displayed these results.
Process mapping of the existing newly-diagnosed patient pathway highlighted variation in the service, particularly around timing of carbohydrate counting and home download setup for new patients. Additionally, three months seemed too long to wait for a first multidisciplinary team (MDT) clinic appointment, so an additional appointment was introduced at six weeks post diagnosis. A goal was set to improve both mean and median HbA1c outcomes for newly-diagnosed patients in the first year. Fishbone analysis identified key project elements. The ward care plan for newly diagnosed patients was updated, a daily dietary record sheet was created for recording carbohydrate content of food consumed and insulin delivered as per Expert meter advice. The clinical guideline for medical staff was updated, including a step-by-step Expert meter setup guide with starting insulin carbohydrate ratios and insulin sensitivity factors for patients of different age ranges. The first new patient was diagnosed at the end of November during diabetes team working hours and the changes were piloted on this patient. Subsequently the ward diabetes drawers were re-labelled and packs were introduced for newly-diagnosed patients, containing an Expert meter, Expert meter cable, plus a Carbs and Cals book. This helped make change easier to adopt in a busy ward environment and has been well received, particularly by junior doctors on the ward. Nursing staff were already trained to count carbohydrates so dieticians organised refresher training. This was facilitated in a busy ward by organising ‘tea trolley’ training sessions where a training station was set up and the diabetes nurse relieved the ward nurses of their clinic duties to allow them to take part. Medical staff practised with training Expert meters and the doctors set up the meters using the guideline. An Expert meter quick reference guide is now supplied in each ward bed space. Diasend download was used to measure average BG level on the day of discharge and at day 28 post diagnosis (using one week of data prior to this). Process measurements included recording whether patients were downloading from diagnosis and also keeping records of the numbers of ward staff trained in Expert meter use. Measures such as length of stay were used to ensure that the changes were not negatively impacting patients. The psychologist was concerned that giving overnight BG corrections in hospital would lead to parents checking blood glucose levels overnight, so this was investigated on day 28 via Diasend download. Episodes of hypoglycaemia were measured and recorded in hospital as well as rates of hypoglycaemia on Diasend download at day 28.
Between November 2018 and May 2019, ten new patients were diagnosed with Type 1 diabetes. All experienced the new patient pathway and had a first MDT clinic appointment at six weeks post diagnosis. They were all discharged from hospital with an average BG level near or in target. Average length of stay was 4.3 working days (DKA days excluded from analysis). Nine out of ten had their first Diasend download completed and a home Diasend account set up at discharge. Hypoglycaemic episodes in hospital only occurred in two patients. Two patients had evidence of ‘routine’ overnight checks at day 28 on Diasend but this was discussed and the overnight checks ceased. Percentage hypoglycaemia rates at day 28 were as high as 31% in one patient but these were mild and none experienced severe hypoglycaemia. Average blood glucose at day 28 showed improvement, with mean 5.9mmol at day 28 in the post-QI cohort, compared with a mean of 8.9mmol pre-QI (median 6.8mmol and 5.7mmol, respectively). Three-month HbA1c results showed an improvement in mean HbA1c at three months, although median was unchanged. However, variation in terms of both standard deviation and inter-quartile ranges for patients post QI improved. Now, new patients expect to achieve a BG level in target. They are downloading more regularly at home and the team and patients have a better understanding of the relationship between HbA1c and average BG levels. Setting high standards from the outset will have a long-term positive impact for patients.
Sustainability and Spread
This project is sustainable and will be continued with all newly-diagnosed patients. Materials have been shared with other teams. Any enthusiastic and motivated paediatric diabetes team can implement the changes. The challenge now is to sustain the improvement in glycaemic outcomes over and beyond the first year of diagnosis. QI methodology is being used to achieve this and weekly download clinics are being implemented.
Back to the top