Background:
Counties Manukau Health (CMH) is located in South Auckland with a single hospital, serving a region with a population of 520,000. The community has the largest population of people living with diabetes in New Zealand (42,000) of whom over 8,000 have poor control (HbA1c >75mmol/mol). 20-25 percent of Middlemore Hospital’s inpatients have diagnosed diabetes and these people can be identified by information support embedded in patient management software. In patients with diabetes have longer average length of stay (LOS) and higher rates of readmission than patients without diabetes. The diabetes service provides a hospital-wide consult service to these patients, with few dedicated inpatient beds, and has increased its coverage from less than 20% to 30% of all patients. As part of an earlier quality improvement project, we have identified that inpatients are often not well prepared for discharge and ongoing self-management of their diabetes when they leave hospital. CMH undertook a hospital wide program with primary aims to reduce length of inpatient stay (LOS) and enhance care and reduce readmissions to hospital. As part of this clinical quality improvement process we have implemented a new integrated care pathway (ICP) for people with diabetes (PWD), which is user-friendly to general clinical ward staff.

Aims: 
Successful introduction of a new ICP for people with diabetes (PWD) which: 
a. Provides the best possible co-ordinated service for PWD during their inpatient stay 
b. Guides health professionals to prepare the inpatient PWD for safe discharge, with a plan of ongoing care to manage health and patient self-management. 
c. To reduce length of stay 
d. To reduce 30 day hospital readmission rates 
e. To reduce HbA1c at follow-up. 

Method:
Standard care (SC) in two medical wards was observed December 2013 to January 2014. The ICP was then developed by EH and CO’B with ward staff. The ICP was implemented in June 2014 for use by the multidisciplinary team (Doctors, Nurses, Pharmacists, and Dieticians). All PWD admitted to the two medical wards who consented were included. Data was collected by surveys of patient self-assessment of the discharge process and diabetes knowledge, health professional self-assessment of knowledge and competency, and diabetes control by HbA1c after discharge, 30 day readmission rates, and hospital LOS. 

Results:
31 PWD were reviewed under standard care and. 16 were observed in ICP. Results are expressed as SC then ICP. The ICP was successfully implemented as judged by the ward clinical staff and diabetes service. In both groups HbA1c improved after discharge with a favourable trend in the ICP group (8.1 mmol/mol vs 9.3 mmol/mol, ns). Average LOS was similar in both groups (5.48 days vs 5.31 days, ns). Readmissions were similar for the two groups with a trend for fewer in the ICP group (32.2% vs 25% ns). Hospital readmissions were not primarily for diabetes in either group. Patients on insulin were more likely to be prescribed and dispensed insulin, needles and test strips at discharge in the ICP group (4.5% vs 40.0%, P=0.01), and both groups were discharged with glucose meters (95.5% vs 80.0%, ns). Patient knowledge of self-care and hypoglycaemia symptoms was not different in the two groups, but surprisingly there was a trend to fewer patients feeling confident in self-management after the ICP (90.3% vs 68.8% p=0.062). Staff knowledge about management showed a favourable trend (28.6% vs 50.0%, ns), but there was a trend towards lower confidence caring for people with diabetes (100.0% vs 87.5%, ns) and reduced confidence to provide patients with education about their glycaemic medications (100% vs 81.3%, p<0.04).

Discussion:
- The ICP did not result in reduced LOS, readmission rates or HbA1c 3 months post admission compared to historical controls. However trends were favourable in this small group 
- ICP was associated with increased dispensing of insulin and self-management resources after discharge, a marker of enhanced transition management. 
- There were trends towards increased staff knowledge in diabetes management but no evidence of enhanced patient knowledge of self-management with the ICP. However confidence in glycaemic management appeared lower in both clinicians and patients in the ICP group. This is likely due to overconfident self-assessment (100% and 90.3 % in the respective groups) prior to the intervention, and suggests unrecognised practical knowledge gaps in general staff and patients.

Integrated care pathway (ICP) for discharge preparation of patients with diabetes