Structured
Approaches to the Evaluation and Treatment of
Inpatient Hyperglycemia and Length of Stay in the
Hospital
Author Block: R. Juneja1, A. Golas1,
N. Kumar2, J. Carroll3, D.
Nelson3, S. Flanders3, C.
Roudebush1; 1Medicine, Indiana
University, Indianapolis, IN, United States, 2Department
of Internal Medicine, Medical College of Wisconsin,
Milwaukee, WI,
United States, 3Medicine,
Clarian Health Partners, Indianapolis, IN, United
States.
Background and Aims:
The Systematic Utilization
of Glucose Assessment and
Response (SUGAR) Program was implemented
in a staged manner to reduce inpatient hyperglycemia
at Clarian Health Partner Hospitals in Indianapolis,
USA.
Materials and
Methods:
The program electronically tracks all in-hospital
Blood Glucose (BG) measurements and triggers
intervention (if BG is >180mg/dl; >110mg/dl in the
ICU). In the ICU, target BG is achieved with IV
insulin (IVI) driven by a computer program prompting
BG monitoring and IVI adjustment to target. In non-ICU
settings, BG control is driven by ongoing programs of
increasing awareness of inpatient hyperglycemia
through education (morning reports/computer training
modules) for physicians/nursing staff by inpatient
diabetes nurses and standardization of s/c insulin
orders.
Results:
We have previously reported significant increase in
the numbers of patients reaching target BG from 75% BG
< 180 mg/dl to 83% (p<0.001) with a simultaneous
increase in the number of BG tests from~ 40,000 to
over 65,000/month. The main impact has been with ICU
targets being achieved through our IVI program
(started Oct 2004) with over 60% now achieving target
BG (up from 26%). In non-ICU patients we have not seen
substantive reduction in overall BG targets, although
there has been a sharp increase in endocrinology
consults for diabetes. This study analyzes the length
of stay(LOS), charges and mortality data from the
hospital database during this time (data from patients
≥18 years of age and with > one BG measurement during
hospitalization).
See Table 1 data for
ALL patients (p<.001:Jan-Jun 2005 vs Jul-Dec 2005 for
LOS and charges; mortality p=0.013.Other time points
mortality p<0.001compared to Jul-Dec 2005).
Table 1: ALL ICU
Patients
|
Months |
# of
Patients |
Mean LOS
(Days) |
Charges ($) |
Mortality
(%) |
| Jan-Jun 04 |
13078 |
8.7 |
34661 |
4.5 |
| Jul-Dec 04 |
13706 |
8.5 |
34900 |
4.4 |
| Jan-Jun 05 |
13882 |
8.4 |
37836 |
4.2 |
| Jul-Dec 05 |
13139 |
7.7 |
35538 |
3.6 |
|
TOTAL |
53805 |
8.3 |
35734 |
4.2 |
See Table 2 data for
patients with at least one day in ICU (p<.001:Jan-Jun
2005 vs Jul-Dec 2005 for LOS and charges; p=ns for
mortality)
Table 2: LOS ICU Patients
|
Months
|
# of
Patients |
Mean LOS
(Days) |
Charges ($) |
Mortality
(%) |
| Jan-Jun 04 |
3209 |
13.2 |
70652 |
11.0 |
| Jul-Dec 04 |
3220 |
13.5 |
72762 |
11.8 |
| Jan-Jun 05 |
3105 |
13.2 |
80478 |
11.6 |
| Jul-Dec 05 |
2865 |
11.5 |
72421 |
10.4 |
|
TOTAL |
12399 |
12.9 |
74078 |
11.2 |
Conclusion:
Our analysis shows that LOS decreased during the
period when tight glycemic control (TGC) measures were
introduced in the hospital accompanied by a reduction
in charges despite a sharp increase in the number of
glucose measurements. However, since overall BG
control has improved only in ICU’s, further studies
are required to evaluate the direct effects of glucose
reduction versus other factors that might be
contributing to this reduction in LOS.
