Q 23What are the optimal methods of managing glucose control in in-patients with stable Type 1 diabetes?

Author/Title/Reference/YrDavies M, Dixon S, Currie CJ, Davis RE, Peters JR 2001 Evaluation of a hospital diabetes specialist nursing service: a randomised controlled trial. Diabetic Medicine 18:301–307
N=300 gave informed consent (intervention group=148, control group=152)
Research DesignRandomised controlled trial
AimTo study the efficacy of nursing services to improve glucose control
PopulationMixed diabetes population. Exclusions: patients unable to complete self reported questionnaires were excluded from secondary outcome assessments, but were entered into the study and primary outcome data recorded
InterventionDiabetes specialist nursing service
Comparisonstandard care
OutcomePrimary measures: Length of stay in hospital, readmission rates (within 12 months), time in days to first readmission.
Secondary measures: community resource use post-discharge, patient satisfaction, diabetes knowledge and disease-specific QOL
CharacteristicsPatient characteristics generally similar in both groups, although slightly more IDDM patients in the intervention groups
ResultsLength of stay
Intervention group: median=8.0 days c.f. 11.0 days in controls (p less than 0.01)
No differences were seen in readmission frequency or time to readmission (p=1.0 and 0.8 respectively)
Both intervention group and those refusing randomisation had a mean LOS of 10.8
Patient satisfaction
Patients in the intervention recorded significantly higher scores for patient satisfaction compared to the controls:
91% were judged as satisfied in the control groups, compared to 59% in the control group (overall p less than 0.001)
Diabetes knowledge
Diabetes knowledge scores were comparable between the two groups at the start of the study, however, this score improved post-admission in the intervention group.
Intervention group knowledge score: baseline=52%, post study=74% (p less than 0.001)
Control group knowledge score: baseline=44%, post study=48% (p greater than 0.05)
Comparison of intervention and control group: baseline p greater than 0.05, post study: p less than 0.05
Quality of life
ADDQOL showed no differences in quality of life between or within groups at base-line or post-discharge (P greater than 0.05 for all comparisons)
Intervention group QOL: baseline=0.65, post study=0.88% (p greater than 0.05)
Control group QOL: baseline=0.88%, post study=0.40 (p greater than 0.05)
Other outcomes
Referral rates to the community DSN service were 38.2% and 28.4% in the control and intervention groups respectively (p=0.09)
Statistically significant lower number of GP contacts were made in the intervention group.
Hierarchy of Evidence GradingIb
CommentsSequential, unselected referrals of in-patients to the DSN service (either IDDM or NIDDM) randomised prior to clinical review.
All patients visited by an independent investigator following randomisation but before any intervention.
Standard Care: any management carried out by health care professionals (medical, general nursing, dietetic) other than the in- patient DSN.
Intervention group received care and advice from DSN in addition to standard care, including individual structured patient education appropriate to need, and practical management advice including verbal and written case-note feedback to ward- based medical and nursing staff. Input began on the day of referral and randomisation and continued until discharge.
Admission and discharge dates taken manually for each patient and verified by the Patient Management System (PMS)
Patient dependency assessed on admission and discharge by a research nurse using a four-point scale
Quality of life measured with the Audit of Diabetes Dependent Quality of Life instrument; diabetes knowledge the Diabetes
Knowledge Scale (modified to provide a version for insulin users and non-users) and patient satisfaction was measured using a modified version of the Diabetes Clinic Satisfaction Questionnaire (DCSQ)
To assess DNS impact on post-discharge events, data on subsequent out-patient attendances, contacts with 1º and social care, and time away from normal activities were recorded at 1 month post-discharge using a postal questionnaire.
Four DSNs were rotated into the single hospital post during the study to reduce possible ‘character bias’
Validation of four-point dependency scale not discussed
Quality of life, diabetes knowledge and patient satisfaction all measured using validated scales.
Power analysis based on previous studies of specialist diabetes care reporting LOS reductions of between 1.3 and 4.6 days (SD:1.7–9.1). Assuming a 2 day reduction in LOS with a SD of 6 days was clinically meaningful, estimated required sample size=140 in each group to achieve a power of 80% to detect the difference with a 5% significance level.
Only small numbers of patients are followed-up for secondary outcomes, therefore data are open to bias
Author/Title/Reference/YrCavan DA, Hamilton P, EverettJ, Kerr D 2001 Reducing hospital inpatient length of stay for patients with diabetes. Diabetic Medicine 18:162–164
N=1611 (792 —prior to nurse introduction; 819 following introduction of specialist nurse)
Research DesignCohort study
AimTo evaluate the effect of nurse care on glucose control
PopulationMixed Diabetes population
InterventionNurse care
Nurse visited all the wards weekly and routinely to identify patients and advise staff about diabetes management problems.
Visited admissions ward, A&E department and coronary care units daily to identify potential problems, facilitate early discharge and provide support and patient education where necessary. Also first point of contact for referrals from other medical teams.
Main objective of post: to ensure patients had good diabetic control during their hospital stay.
Provided education, including basic dietary advice, to patients according to their needs (especially newly diagnosed)
Patients with stable diabetes were assessed and then not seen again unless their condition deteriorated. Patients who posted particular management problems, or whose blood glucose control did not respond to initial changes in regimen, were referred for review by specialist diabetes medical staff.
Specific protocols drawn up for patients requiring attention, including an intensive insulin regimen for patients with MI
Nurse maintained contact for ≤3 months with all patients converted to insulin treatment
Comparisonno nurse care
OutcomeLength of stay in hospital
CharacteristicsInclusion/exclusion criteria not detailed
Results792 and 819 patients with diabetes were admitted to medical and surgical beds in 1997 and 1998 respectively. This represents 3.3 and 3.4% of all patients admitted within that year, respectively.
Median length of stay in 1997 was 11 days (range 1–108) in medicine and 8 (1–109) days in surgery.
Following the introduction of a clinical nurse advisor, these figures were: 8 (1–109) days in medicine and 5 (1–33) in surgery (p less than 0.001 when compared to pre-nurse stats, respectively)
Overall average length of stay in medicine was 8.9 days in 1997 and 10.4 days in 1998, and in surgery 5.1 and 4.9 days respectively.
In 1997 the 3.3% of in patients with diabetes accounted for 6.8% of occupied bed days. In 1998 the 3.4% accounted for only 4% of the occupied bed days—representing a saving of 4171 bed days.
28 Doctors surveyed reported that they regularly called upon the nurse for advice and found her role in patients’ management useful.
Hierarchy of Evidence GradingIIa
CommentsStudies of this kind are very susceptible to bias, due to the possibility that other factors outside of the intervention are influencing the outcome. For example, there could have been other changes of staff or recommended medication over the course of the year which would change the outcome seen in patients with diabetes— vascular surgeons changes their discharge policy for diabetic patients, leading to shorter admissions but the consistency of the reductions across specialities suggests the introduction of the nurse did have a significant role
Study only demonstrates changes in length of stay does not give any clues as to why this happened.
Authors make assumption that shorter length of stay implies improved diabetic control.
Patients entered into the study prospectively but data collected retrospectively. Data collected from official discharge statistics.
No matching between patients, no baseline demographic statistics provided.
Conclusion states: a ward based nurse can positively influence the management of inpatients with diabetes, by adjusting diabetes treatment and providing advice and information to patients carers and staff. This conclusion is not related to the study performed in which the outcome was length of stay in medical and surgical wards
Author/Title/Reference/YrKoproski J, Pretto Z, Poretsky L 1997 Effects on an intervention by a diabetes team in hospitalised patients with diabetes. Diabetes Care 20(10):1553–1555
N=179 (usual care plus diabetes team=85 patients, usual care (controls)=94
Research DesignRandomised Controlled Study
AimAn evaluation of the effect of a diabetes care team
PopulationNot stated
InterventionUsual care supplemented with a diabetes team intervention
Diabetes team visited patients on a daily basis, orders were written by the endocrinologists once discussion took place with the primary care physician. Nutrition and social work consultations were requested by the team based on individual need.
Comparisonusual care.
OutcomeLength of stay, blood glucose control, readmission rates.
Characteristicsaged 15–94 years
Males: 53% in intervention group, 47% in control group
Age (years): intervention group=60.3±15.1, control group 64.7±14.2
Duration of diabetes (years): intervention group=13.1±14.3, control group=16.2±20.7
Insulin treated patients (%): intervention group=48, control group =40
ResultsIntervention and control groups were similar in demographic characteristics (age, gender, duration of diabetes, number with insulin treated diabetes, and co morbid condition diagnoses and frequency)
Mean blood glucose on admission differed between the two groups: intervention group=323.3±192 mg/dl vs. 259.5±143.8 in controls (p less than 0.01)
Mean and median length of stay was not significantly different for patients with a primary diagnosis of diabetes in the intervention or controls (5.5 vs. 7.5 days— 95%CI: 4–8 and 5–11 respectively)
Patients with secondary diagnosis of diabetes, median length of stay was 10.0 days (8–13) in the intervention group and 10.5 (8–13)days in controls.
Compared to the New York State expected values, median length of stay in the control group with was two days less than expected in patients with primary diagnosis of diabetes, and 1 day less than expected in the intervention group for those with secondary diagnosis of diabetes.
Patients with “good” blood glucose levels during the 1st month accounted for 75% in the intervention group and 46% in the control group. This difference was reduced over time.
Documented instructions for blood glucose monitoring were present in 89% of cases in the intervention group and 57% in the control group (p less than 0.01).
Documented instructions for insulin administration were present in 69% of the charts of insulin-treated patients in the intervention group and 57% in the control group (p less than 0.01)
87% of patients had documented education of any kind in the intervention group, compared with 37% in the control group (p less than 0.0001)
76% of patients in the intervention group had nutritional consultation, compared with 40% in the control group (p less than 0.0001)
No significant difference was seen in the percentage of patients seen by the social services.
In the intervention group 15% were readmitted within three months after discharge, compare with 32% in the control group (p less than 0.01). This difference persisted at six months
Hierarchy of Evidence GradingIb
CommentsLength of stay in each group compared with the average expected length of stay based on New York Group Standard days to account for possible variability in severity of illness between groups.
Blood glucose control was considered “good” when 75% of capillary glucose measurements during the last two-thirds of the hospitalisation ranged between 80 and 180 mg/dl
Medical record entries for insulin administration and blood glucose monitoring instructions and nutrition and social service consultations were counted
All patients were tracked for readmissions at 3 and 6 months
Length of stay in each group compared with the average expected length of stay based on New York Group Standard days to account for possible variability in severity of illness between groups.
Blood glucose control was considered “good” when 75% of capillary glucose measurements during the last two-thirds of the hospitalisation ranged between 80 and 180 mg/dl
Medical record entries for insulin administration and blood glucose monitoring instructions and nutrition and social service consultations were counted
All patients were tracked for readmissions at 3 and 6 months

From: Appendix D, Evidence tables

Cover of Type 1 Diabetes in Adults
Type 1 Diabetes in Adults: National Clinical Guideline for Diagnosis and Management in Primary and Secondary Care.
NICE Clinical Guidelines, No. 15.1.
National Collaborating Centre for Chronic Conditions (UK).
Copyright © 2004, Royal College of Physicians of London.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.