A retrospective audit of insulin infusion management involving a locally developed dynamic insulin infusion guideline in a tertiary ICU.

BACKGROUND: The ideal target blood glucose range for intensive care patients on insulin infusions is controversial. Avoidance of hyperglycaemia and hypoglycaemia are well supported goals. METHODS: An audit of insulin infusion management was conducted following the institution of an insulin infusion guideline in a tertiary adult intensive care unit (ICU). The primary aim was to evaluate this guideline for safety and efficacy. Secondary aims were to compare outcomes such as ICU and hospital mortality, rate of severe hypoglycaemia, length of time within target zones, length of stay in ICU and hospital, ventilator hours and use of renal replacement therapy. Data analysis involved descriptive statistical techniques to allow comparison with other reported outcomes. RESULTS: Thirty-eight (38) patients were included, representing 137 days of insulin infusions and 2537 blood glucose readings. The mean insulin infusion treatment time was 86.4h (sd ± 86.4), median 48 h (IQR 14.4-141.6). The mean insulin dose per day was 97.6 units (sd ± 115.7), with a median of 68.7 (IQR 38.9-108.3). Blood glucose level (BGL) readings were within the desired target (6-9 mmols/L) and/or the buffer zones (4-6 and 9-12 mmols/L), 92.3% of the time. There were no episodes of severe hypoglycaemia (BGL ≤ 2.2 mmols/L). The median length of ICU stay was 5.9 days. Eighty-four (84) % of the cohort received mechanical ventilation and 26% received renal replacement therapy. The mean ventilation and renal replacement duration were days 6.9 and 9.4 days, respectively. The ICU and hospital mortality was 13.2% and 18.4%, respectively. CONCLUSION: The use of this locally developed insulin infusion guideline for hyperglycaemia within this ICU appears safe and effective. When compared to related published randomised controlled trials, the outcomes of this small scale single centre retrospective audit appear congruent. It achieved a severe hypoglycaemic rate of zero, with BGLs within target and buffer zones greater than 90%. It may be worthwhile for intensive care units to consider evaluating their own locally developed insulin infusion guidelines to ensure safety and efficacy.

Medication Error Minimization Scheme (MEMS) in an adult tertiary intensive care unit (ICU) 2009-2011.

INTRODUCTION: The Medication Error Minimisation Scheme (MEMS) is a locally based ongoing multidisciplinary, multifaceted quality improvement (QI) project within an Australian adult tertiary level Intensive Care Unit (ICU). The project commenced in 2009. Its primary aim is to enhance medication safety within this ICU by utilising existing resources. The aim of this paper is to provide a descriptive account of the various activities, interventions and results of this project within the first three years. METHODS: The research design for this project was based upon Plan-Do-Study-Act (PDSA) cycles associated with QI projects. Medication error rates and audits of: intravenous infusions, incompatible intravenous medications and incorrect documentation of withheld medications were analyzed according to simple statistical techniques. Initial and follow up medication safety surveys were compared using basic statistical analysis. Focus groups exploring barriers and enablers of medication incident reporting were analyzed according to qualitative techniques associated with focus group discussions. Other interventions included: regular education sessions; discussions within other departmental meetings such as nursing staff meetings and Morbidity and Mortality meetings; and bedside discussions and demonstrations. Promotion of medication safety occurred within a number of forums; activities and findings were advertised and displayed; a recognizable Logo for MEMS was employed; and incentives were provided for staff. RESULTS: Reported Medication Incidents (MIs) increased from 6.2 to 14.9 MIs per 1000 patient days. Audits and chart reviews confirmed that more MIs are uncovered by employing a variety of techniques in addition to incident reporting. Staff surveys provided a rich source of information regarding medication safety. Audits of intravenous infusions revealed a reduced error rate from 38/331 (11.5%) to 15/468 (3.2%). Chart review of incorrect documentation of omitted medications decreased from 105/347 (30.3%) to 104/486 (21.4%). Focus groups provided information that was able to be used in a number of hospital forums in order to explain the impact of existing systems upon ICU staff. CONCLUSION: This ongoing QI project was able to achieve its targeted goals. The MI reporting rate was increased. This project demonstrated that measurable, "non-incident report" errors can be reduced by focusing upon and promoting medication safety in the ICU. These activities demonstrated a workplace that values medication safety, the discovery of shortfalls and the benefits of ongoing improvement.