Data-driven curation process for describing the blood glucose management in the intensive care unit.
ABSTRACT: Analysis of real-world glucose and insulin clinical data recorded in electronic medical records can provide insights into tailored approaches to clinical care, yet presents many analytic challenges. This work makes publicly available a dataset that contains the curated entries of blood glucose readings and administered insulin on a per-patient basis during ICU admissions in the Medical Information Mart for Intensive Care (MIMIC-III) database version 1.4. Also, the present study details the data curation process used to extract and match glucose values to insulin therapy. The curation process includes the creation of glucose-insulin pairing rules according to clinical expert-defined physiologic and pharmacologic parameters. Through this approach, it was possible to align nearly 76% of insulin events to a preceding blood glucose reading for nearly 9,600 critically ill patients. This work has the potential to reveal trends in real-world practice for the management of blood glucose. This data extraction and processing serve as a framework for future studies of glucose and insulin in the intensive care unit.
Project description:<h4>Introduction</h4>A single centre has reported that implementation of an intensive insulin protocol, aiming for tight glycaemic control (blood glucose 4.4 to 6.1 mmol/l), resulted in significant reduction in mortality in longer stay medical and surgical critically ill patients. Our aim was to determine the degree to which tight glycaemic control can be maintained using an intensive insulin therapy protocol with computerized decision support and to identify factors that may be associated with the degree of control.<h4>Methods</h4>At a general adult 22-bed intensive care unit, we implemented an intensive insulin therapy protocol in mechanically ventilated patients, aiming for a target glucose range of 4.4 to 6.1 mmol/l. The protocol was integrated into the computerized information management system by way of a decision support program. The time spent in each predefined blood glucose band was estimated, assuming a linear trend between measurements.<h4>Results</h4>Fifty consecutive patients were investigated, involving analysis of 7,209 blood glucose samples, over 9,214 hours. The target tight glycaemic control band (4.4 to 6.1 mmol/l) was achieved for a median of 23.1% of the time that patients were receiving intensive insulin therapy. Nearly half of the time (median 48.5%), blood glucose was within the band 6.2 to 7.99 mmol/l. Univariate analysis revealed that body mass index (BMI), Acute Physiology and Chronic Health Evaluation (APACHE) II score and previous diabetes each explained approximately 10% of the variability in tight glycaemic control. BMI and APACHE II score explained most (27%) of the variability in tight glycaemic control in the multivariate analysis, after adjusting for age and previous diabetes.<h4>Conclusion</h4>Use of the computerized decision supported intensive insulin therapy protocol did result in achievement of tight glycaemic control for a substantial percentage of each patient's stay, although it did deliver 'normoglycaemia' (4.4 to about 8 mmol/l) for nearly 75% of the time. Tight glycaemic control was difficult to achieve in critically ill patients using this protocol. More sophisticated methods such as continuous blood glucose monitoring with automated insulin and glucose infusion adjustment may be a more effective way to achieve tight glycaemic control. Glycaemia in patients with high BMI and APACHE II scores may be more difficult to control using intensive insulin therapy protocols. Trial registration number 05/Q0505/1.
Project description:In this data note we provide the details of a research database of 4831 adult intensive care patients who were treated in the Bristol Royal Infirmary, UK between 2015 and 2019. The purposes of this publication are to describe the dataset for external researchers who may be interested in making use of it, and to detail the methods used to curate the dataset in order to help other intensive care units make secondary use of their routinely collected data. The curation involves linkage between two critical care datasets within our hospital and the accompanying code is available online. For reasons of data privacy the data cannot be shared without researchers obtaining appropriate ethical consents. In the future we hope to obtain a data sharing agreement in order to publicly share the de-identified data, and to link our data with other intensive care units who use a Philips clinical information system.
Project description:Intensive insulin therapy in the critically ill reduces mortality but carries the risk of increased hypoglycemia. Point-of-care blood glucose analysis is standard; however, anemia causes falsely high values and potentially masks hypoglycemia. Permissive anemia is practiced routinely in most intensive care units. We hypothesized that point-of-care glucometer error due to anemia is prevalent, can be corrected mathematically, and correction uncovers occult hypoglycemia during intensive insulin therapy.The study has both retrospective and prospective phases. We reviewed data to verify the presence of systematic error, determine the source of error, and establish the prevalence of anemia. We confirmed our findings by reproducing the error in an in vitro model. Prospective data were used to develop a correction formula validated by the Monte Carlo method. Correction was implemented in a burn intensive care unit and results were evaluated after 9 mos.Burn and trauma intensive care units at a single research institution.Samples for in vitro studies were taken from healthy volunteers. Samples for formula development were from critically ill patients who received intensive insulin therapy.Insulin doses were calculated based on predicted serum glucose values from corrected point-of-care glucometer measurements.Time-matched point-of-care glucose, laboratory glucose, and hematocrit values. We previously found that anemia (hematocrit <34%) produces systematic error in glucometer measurements. The error was correctible with a mathematical formula developed and validated, using prospectively collected data. Error of uncorrected point-of-care glucose ranged from 19% to 29% (p < .001), improving to < or = 5% after mathematical correction of prospective data. Comparison of data pairs before and after correction formula implementation demonstrated a 78% decrease in the prevalence of hypoglycemia in critically ill and anemic patients treated with insulin and tight glucose control (p < .001).A mathematical formula that corrects erroneous point-of-care glucose values due to anemia in intensive care unit patients reduces the prevalence of hypoglycemia during intensive insulin therapy.
Project description:Although a device is needed to continuously measure blood glucose levels within an intensive care setting, and several large-scale prospective studies have shown that patients might benefit from intensive insulin, potassium, or glucose therapy during intensive care, no devices are currently available to continuously assess blood glucose levels in critically ill patients. We conceived the study described here to evaluate the clinical use of the Continuous Glucose Monitor (CGM) performed via a central vein, and to determine the impact of phenomena, such as drift and shift, on the agreement between the CGM and a RAPIDLab® 1265 blood gas analyser (BGA).In the CONTinuous ASSessment of blood GLUcose (CONTASSGLU) study, up to 130 patients under intensive care will be fitted with the CGM, an ex vivo device that continuously measures blood glucose and lactate levels. Readings from the device taken 8?h after initial placement and calibration will be compared with values measured by a BGA. For this study, we chose the BGA as it is an established standard point-of-care device, instead of the devices used in certified central laboratories. Nevertheless, we will also independently compare the results from the point-of-care BGA with those determined by a central laboratory-based device. Blood samples will be collected from each patient from the same site in which the CGM will measure blood glucose. Consequently, each participant will serve as their own control, and no randomisation is necessary. The 95% limits of agreement and the corresponding confidence intervals will be calculated and compared with a prespecified clinically acceptable relative difference of 20%.Several attempts have been made to develop a device to continuously measure blood glucose levels within an intensive care setting or to use the devices that were originally designed for diabetes management, as several of these devices are already available. However, none of these devices were successful in intensive care settings. CONTASSGLU may well bridge this gap by confirming the ability of the CGM to continuously measure blood glucose levels in intensive care settings.ClinicalTrials.gov NCT01580176.
Project description:Most currently available glycan structure databases use their own proprietary structure representation schema and contain numerous annotation errors. These cause problems when glycan databases are used for the annotation or mining of data generated in the laboratory. Due to the complexity of glycan structures, curating these databases is often a tedious and labor-intensive process. However, rigorously validating glycan structures can be made easier with a curation workflow that incorporates a structure-matching algorithm that compares candidate glycans to a canonical tree that embodies structural features consistent with established mechanisms for the biosynthesis of a particular class of glycans. To this end, we have implemented Qrator, a web-based application that uses a combination of external literature and database references, user annotations and canonical trees to assist and guide researchers in making informed decisions while curating glycans. Using this application, we have started the curation of large numbers of N-glycans, O-glycans and glycosphingolipids. Our curation workflow allows creating and extending canonical trees for these classes of glycans, which have subsequently been used to improve the curation workflow.
Project description:For the nearly 75% of patients living with type 2 diabetes (T2DM) that do not use insulin, decisions regarding self-monitoring of blood glucose (SMBG) can be especially problematic. While in theory SMBG holds great promise for sparking favorable behavior change, it is a resource intensive activity without firmly established patient benefits. This study describes our study protocol to assess the impact of three different SMBG testing approaches on patient-centered outcomes in patients with non-insulin treated T2DM within a community-based, clinic setting.Using stakeholder engagement approach, we developed and implemented a pragmatic trial of patient with non-insulin treated T2DM patients from five primary care practices randomized to one of three SMBG regimens: 1) no testing; 2) once daily testing with standard feedback consisting of glucose values being immediately reported to the patient through the glucose meter; and 3) once daily testing with enhanced patient feedback consisting of glucose values being immediately reported to the patient PLUS automated, tailored feedback messaging delivered to the patient through the glucose meter following each testing. Main outcomes assessed at 52 weeks include quality of life and glycemic control.This pragmatic trial seeks to better understand the value of SMBG in non-insulin treated patients with T2DM. This paper outlines the protocol used to implement this study in fifteen community-based primary care practices and highlights the impact of stakeholder involvement from the earliest stages of project conception and implementation. Plans for stakeholder involvement for result dissemination are also discussed.ClinicalTrials.gov NCT02033499 , January 9, 2014.
Project description:Acute hyperglycemia is common in critically ill patients. Strict control of blood glucose (BG) concentration has been considered important because hyperglycemia is associated independently with increased intensive care unit mortality. After intensive insulin therapy was reported to reduce mortality in selected surgical critically ill patients, lowering of BG levels was recommended as a means of improving patient outcomes. However, a large multicenter multination study has found that intensive insulin therapy increased mortality significantly. A difference in variability of BG control may be one possible explanation why the effect of intensive insulin therapy varied from beneficial to harmful. Several studies have confirmed significant associations between variability of BG levels and patient outcomes. Decreasing the variability of the BG concentration may be an important dimension of glucose management. If reducing swings in the BG concentration is a major biologic mechanism behind the putative benefits of glucose control, it may not be necessary to pursue lower glucose levels with their attendant risk of hypoglycemia.
Project description:BACKGROUND: Sustained hyperglycemia is a known risk factor for adverse outcomes in critically ill patients. The specific aim was to determine if a nurse initiated insulin infusion protocol (IIP) was effective in maintaining blood glucose values (BG) within a target goal of 100-150 mg/dL across different intensive care units (ICUs) and to describe glycemic control during the 48 hours after protocol discontinuation. METHODS: A descriptive, retrospective review of 366 patients having 28,192 blood glucose values in three intensive care units, Surgical Trauma Intensive Care Unit (STICU), Medical (MICU) and Coronary Care Unit (CCU) in a quaternary care hospital was conducted. Patients were > 15 years of age, admitted to STICU (n = 162), MICU (n = 110) or CCU (n = 94) over 8 months; October 2003-June 2004 and who had an initial blood glucose level > 150 mg/dL. We summarized the effectiveness and safety of a nurse initiated IIP, and compared these endpoints among STICU, MICU and CCU patients. RESULTS: The median blood glucose values (mg/dL) at initiation of insulin infusion protocol were lower in STICU (188; IQR, 162-217) than in MICU, (201; IQR, 170-268) and CCU (227; IQR, 178-313); p < 0.0001. Mean time to achieving a target glucose level (100-150 mg/dL) was similar between the three units: 4.6 hours in STICU, 4.7 hours in MICU and 4.9 hours in CCU (p = 0.27). Hypoglycemia (BG < 60 mg/dL) occurred in 7% of STICU, 5% of MICU, and 5% of CCU patients (p = 0.85). Protocol violations were uncommon in all three ICUs. Mean blood glucose 48 hours following IIP discontinuation was significantly different for each population: 142 mg/dL in STICU, 167 mg/dL in MICU, and 160 mg/dL in CCU (p < 0.0001). CONCLUSION: The safety and effectiveness of nurse initiated IIP was similar across different ICUs in our hospital. Marked variability in glucose control after the protocol discontinuation suggests the need for further research regarding glucose control in patients transitioning out of the ICU.
Project description:BACKGROUND: Studies in adult intensive care have highlighted the importance of insulin and improved glucose control on survival, with 32% reduction in mortality, 22% reduction in intensive care stay and halving of the incidence of bacteraemia. Very low birth weight infants requiring intensive care also have relative insulin deficiency often leading to hyperglycaemia during the first week of life. The physiological influences on insulin secretion and sensitivity, and the potential importance of glucose control at this time are not well established. However there is increasing evidence that the early postnatal period is critical for pancreatic development. At this time a complex set of signals appears to influence pancreatic development and beta cell survival. This has implications both in terms of acute glucose control but also relative insulin deficiency is likely to play a role in poor postnatal growth, which has been associated with later motor and cognitive impairment, and fewer beta cells are linked to risk of type 2 diabetes later in life. METHODS: A multi-centre, randomised controlled trial of early insulin replacement in very low birth weight babies (VLBW, birth weight < 1500 g). 500 infants will be recruited from 10 centres in the UK and Europe. Babies will be randomised to receive a continuous insulin infusion (0.05 units/kg/h) or to receive standard neonatal care from the first day of life and for the next 7 days. If blood glucose (BG) levels fall infants will receive 20% dextrose titrated to maintain normoglycaemia (4-8 mmol/l). If BG is consistently above 10 mmol/l babies will receive standard treatment with additional insulin infusion. The primary end point will be mortality on or before expected date of delivery, secondary end points will be markers of morbidity and include episodes of sepsis, severity of retinopathy, chronic lung disease and growth.
Project description:This pilot, randomized, open-label controlled study compared the basal-bolus regimens of insulin glargine (IG) and neutral protamine Hagedorn (NPH) insulin in stroke patients with hyperglycemia receiving intensive care. The study recruited acute stroke patients requiring intensive care within 72 h (h) of onset and had blood glucose > 200 mg/dL. 50 patients received IG (n = 26) or NPH (n = 24) with added short-acting prandial regular insulin over a 72-h period. The primary end point was the percentage of glucose within 80-180 mg/dL assessed through continuous glucose monitoring. The baseline characteristics were comparable, except the IG had higher glucose pre-randomization than the NPH (290.69 ± 82.31 vs. 246.04 ± 41.76 mg/dL, P = 0.021). The percentage of time with glucose between 80 and 180 mg/dL was 45.88 ± 27.04% in the IG and 53.56 ± 22.89% in the NPH (P = 0.341) and the percentage of glucose reduction was 31.47 ± 17.52% in the IG and 27.28 ± 14.56% in the NPH (P = 0.374). The percentage of time with glucose < 60 mg/dL was 0.14 ± 0.49% in the IG and 0.47 ± 1.74% in the NPH. Poststroke outcomes were not significantly different. In conclusion, IG is safe and equally effective as an NPH-based basal-bolus regimen for acute stroke patients with hyperglycemia receiving intensive care.Trial registration ClinicalTrials.gov, NCT02607943. Registered 18/11/2015, https://clinicaltrials.gov/ct2/show/NCT02607943 .