PREVENTION AND MANAGEMENT OF INSULIN-ASSOCIATED HYPOGLYCEMIA IN HOSPITALIZED PATIENTS

Artificial Intelligence for Predicting and Diagnosing Complications of Diabetes

Artificial intelligence can use real-world data to create models capable of making predictions and medical diagnosis for diabetes and its complications. The aim of this commentary article is to provide a general perspective and present recent advances on how artificial intelligence can be applied to improve the prediction and diagnosis of six significant complications of diabetes including (1) gestational diabetes, (2) hypoglycemia in the hospital, (3) diabetic retinopathy, (4) diabetic foot ulcers, (5) diabetic peripheral neuropathy, and (6) diabetic nephropathy.

Machine Learning Models for Inpatient Glucose Prediction

PURPOSE OF REVIEW

Glucose management in the hospital is difficult due to non-static factors such as antihyperglycemic and steroid doses, renal function, infection, surgical status, and diet. Given these complex and dynamic factors, machine learning approaches can be leveraged for prediction of glucose trends in the hospital to mitigate and prevent suboptimal hypoglycemic and hyperglycemic outcomes. Our aim was to review the clinical evidence for the role of machine learning-based models in predicting hospitalized patients' glucose trajectory.

RECENT FINDINGS

The published literature on machine learning algorithms has varied in terms of population studied, outcomes of interest, and validation methods. There have been tools developed that utilize data from both continuous glucose monitors and large electronic health records (EHRs). With increasing sample sizes, inclusion of a greater number of predictor variables, and use of more advanced machine learning algorithms, there has been a trend in recent years towards increasing predictive accuracy for glycemic outcomes in the hospital setting. While current models predicting glucose trajectory offer promising results, they have not been tested prospectively in the clinical setting. Accurate machine learning algorithms have been developed and validated for prediction of hypoglycemia and hyperglycemia in the hospital. Further work is needed in implementation/integration of machine learning models into EHR systems, with prospective studies to evaluate effectiveness and safety of such clinical decision support on glycemic and other clinical outcomes.

Retrospective Quality Improvement Study of Insulin-Induced Hypoglycemia and Implementation of Hospital-Wide Initiatives

BACKGROUND

Hospitalized patients who are receiving antihyperglycemic agents are at increased risk for hypoglycemia. Inpatient hypoglycemia may lead to increased risk for morbidity, mortality, prolonged hospitalization, and readmission within 30 days of discharge, which in turn may lead to increased costs. Hospital-wide initiatives targeting hypoglycemia are known to be beneficial; however, their impact on patient care and economic measures in community nonteaching hospitals are unknown.

METHODS

This retrospective quality improvement study examined the effects of hospital-wide hypoglycemia initiatives on the rates of insulin-induced hypoglycemia in a community hospital setting from January 1, 2016, until September 30, 2019. The potential cost of care savings has been calculated.

RESULTS

Among 49 315 total patient days, 2682 days had an instance of hypoglycemia (5.4%). Mean ± SD hypoglycemic patient days/month was 59.6 ± 16.0. The frequency of hypoglycemia significantly decreased from 7.5% in January 2016 to 3.9% in September 2019 (P = .001). Patients with type 2 diabetes demonstrated a significant decrease in the frequency of hypoglycemia (7.4%-3.8%; P < .0001), while among patients with type 1 diabetes the frequency trended downwards but did not reach statistical significance (18.5%-18.0%; P = 0.08). Based on the reduction of hypoglycemia rates, the hospital had an estimated cost of care savings of $98 635 during the study period.

CONCLUSIONS

In a community hospital setting, implementation of hospital-wide initiatives targeting hypoglycemia resulted in a significant and sustainable decrease in the rate of insulin-induced hypoglycemia. These high-leverage risk reduction strategies may be translated into considerable cost savings and could be implemented at other community hospitals.

Development and Validation of a Machine Learning Model to Predict Near-Term Risk of Iatrogenic Hypoglycemia in Hospitalized Patients

Importance

Accurate clinical decision support tools are needed to identify patients at risk for iatrogenic hypoglycemia, a potentially serious adverse event, throughout hospitalization.

Objective

To predict the risk of iatrogenic hypoglycemia within 24 hours after each blood glucose (BG) measurement during hospitalization using a machine learning model.

Design, Setting, and Participants

This retrospective cohort study, conducted at 5 hospitals within the Johns Hopkins Health System, included 54 978 admissions of 35 147 inpatients who had at least 4 BG measurements and received at least 1 U of insulin during hospitalization between December 1, 2014, and July 31, 2018. Data from the largest hospital were split into a 70% training set and 30% test set. A stochastic gradient boosting machine learning model was developed using the training set and validated on internal and external validation.

Exposures

A total of 43 clinical predictors of iatrogenic hypoglycemia were extracted from the electronic medical record, including demographic characteristics, diagnoses, procedures, laboratory data, medications, orders, anthropomorphometric data, and vital signs.

Main Outcomes and Measures

Iatrogenic hypoglycemia was defined as a BG measurement less than or equal to 70 mg/dL occurring within the pharmacologic duration of action of administered insulin, sulfonylurea, or meglitinide.

Results

This cohort study included 54 978 admissions (35 147 inpatients; median [interquartile range] age, 66.0 [56.0-75.0] years; 27 781 [50.5%] male; 30 429 [55.3%] White) from 5 hospitals. Of 1 612 425 index BG measurements, 50 354 (3.1%) were followed by iatrogenic hypoglycemia in the subsequent 24 hours. On internal validation, the model achieved a C statistic of 0.90 (95% CI, 0.89-0.90), a positive predictive value of 0.09 (95% CI, 0.08-0.09), a positive likelihood ratio of 4.67 (95% CI, 4.59-4.74), a negative predictive value of 1.00 (95% CI, 1.00-1.00), and a negative likelihood ratio of 0.22 (95% CI, 0.21-0.23). On external validation, the model achieved C statistics ranging from 0.86 to 0.88, positive predictive values ranging from 0.12 to 0.13, negative predictive values of 0.99, positive likelihood ratios ranging from 3.09 to 3.89, and negative likelihood ratios ranging from 0.23 to 0.25. Basal insulin dose, coefficient of variation of BG, and previous hypoglycemic episodes were the strongest predictors.

Conclusions and Relevance

These findings suggest that iatrogenic hypoglycemia can be predicted in a short-term prediction horizon after each BG measurement during hospitalization. Further studies are needed to translate this model into a real-time informatics alert and evaluate its effectiveness in reducing the incidence of inpatient iatrogenic hypoglycemia.

The Effect of LM25 and LM50 on Hypoglycemia in Chinese T2DM Patients: Post Hoc Analysis of a Randomized Crossover Trial

INTRODUCTION

To investigate the safety of insulin lispro Mix 25 and 50 (LM25 and LM50) in hypoglycemia in patients with type 2 diabetes mellitus (T2DM).

METHODS

This was a post hoc analysis of a phase IV, randomized, crossover clinical trial in Chinese patients with T2DM switching from premixed human insulin 70/30 (PHI70/30) to LM25 or LM50. Eighty-one subjects received a two-stage crossover protocol of either LM25 or LM50 twice daily for 16 weeks. Habitual diet was taken, and self-monitoring of blood glucose (SMBG) was performed throughout the study period. High-carbohydrate diet (HCD), high-fat diet (HFD) and habitual diet patterns were taken, and 72 h continuous glucose monitoring (CGM) was performed at the last 3 days of each treatment stage.

RESULTS

The frequencies of nocturnal hypoglycemia in LM50 were lower than those in LM25 under a Chinese habitual diet pattern. The related factors of hypoglycemia in patients with T2DM treated with a LM25 or LM50 regimen were the weight-based daily mean insulin dose and the type of combined oral hypoglycemic agents. Under both HCD and habitual diet patterns, the optimal cut point values of bedtime glucose predicting nocturnal hypoglycemia in LM50 were lower than those in LM25.

CONCLUSIONS

The risk of nocturnal hypoglycemia in the LM50 regimen was lower than that in the LM25 regimen under the HCD pattern, and the safety range of bedtime glucose for the LM50 regimen was wider than that of the LM25 regimen in Chinese T2DM patients. Premixed insulin analogs combined with acarbose were more helpful to reduce the incidence of hypoglycemia.

TRIAL REGISTRATION

http://www.chictr.org.cn #ChiCTR-TTRCC-12002516.

Development and Validation of a Hypoglycemia Risk Model for Intensive Insulin Therapy in Patients with Type 2 Diabetes

AIMS

To develop a simple hypoglycemic prediction model to evaluate the risk of hypoglycemia during hospitalization in patients with type 2 diabetes treated with intensive insulin therapy.

METHODS

We performed a cross-sectional chart review study utilizing the electronic database of the Third Affiliated Hospital of Sun Yat-sen University, and included 257 patients with type 2 diabetes undergoing intensive insulin therapy in the Department of Endocrinology and Metabolism. Logistic regression analysis was used to derive the clinical prediction rule with hypoglycemia (blood glucose ≤ 3.9 mmol/L) as the main result, and internal verification was performed.

RESULTS

In the derivation cohort, the incidence of hypoglycemia was 51%. The final model selected included three variables: fasting insulin, fasting blood glucose, and total treatment time. The area under the curve (AUC) of this model was 0.666 (95% CI: 0.594-0.738, P < 0.001).

CONCLUSIONS

The model's hypoglycemia prediction and the actual occurrence are in good agreement. The variable data was easy to obtain and the evaluation method was simple, which could provide a reference for the prevention and treatment of hypoglycemia and screen patients with a high risk of hypoglycemia.

Inpatient Glycemic Management of Non-cardiac CVD: Focus on Stroke and PVD

PURPOSE OF REVIEW

Hyperglycemia occurs frequently in hospitalized patients with stroke and peripheral vascular disease (PVD). Guidelines for inpatient glycemic management are not well established for this patient population. We will review the clinical impact of hyperglycemia in this acute setting and review the evidence for glycemic control.

RECENT FINDINGS

Hyperglycemia in acute stroke is associated with poor short and long-term outcomes, and perioperative hyperglycemia in those undergoing revascularization for PVD is linked to increased post-surgical complications. Studies evaluating tight glucose control do not demonstrate improvement in clinical outcomes, although the risk for hypoglycemia increases substantially. Additional studies are needed to evaluate tight glucose goals relative to our current standard of care and the role of permissive hyperglycemia. Given the limited data to guide glycemic management in these patient populations, it is recommended that general guidelines for inpatient glycemic control be followed. Special considerations should be made to address factors that may impact glucose management, including neurological deficits and clinical changes that occur in the postoperative state.

Internal Experts Collaborate to Reduce Critical Hypoglycemia and Insulin Errors and Improve Insulin Administration Timing

IN BRIEF "Quality Improvement Success Stories" are published by the American Diabetes Association in collaboration with the American College of Physicians, Inc., and the National Diabetes Education Program. This series is intended to highlight best practices and strategies from programs and clinics that have successfully improved the quality of care for people with diabetes or related conditions. Each article in the series is reviewed and follows a standard format developed by the editors of Clinical Diabetes. The following article describes a project aimed at reducing inpatient critical hypoglycemia episodes in a community hospital setting.

Development and validation of a prediction model for insulin-associated hypoglycemia in non-critically ill hospitalized adults

OBJECTIVE

To develop and validate a multivariable prediction model for insulin-associated hypoglycemia in non-critically ill hospitalized adults.

RESEARCH DESIGN AND METHODS

We collected pharmacologic, demographic, laboratory, and diagnostic data from 128 657 inpatient days in which at least 1 unit of subcutaneous insulin was administered in the absence of intravenous insulin, total parenteral nutrition, or insulin pump use (index days). These data were used to develop multivariable prediction models for biochemical and clinically significant hypoglycemia (blood glucose (BG) of ≤70 mg/dL and <54 mg/dL, respectively) occurring within 24 hours of the index day. Split-sample internal validation was performed, with 70% and 30% of index days used for model development and validation, respectively.

RESULTS

Using predictors of age, weight, admitting service, insulin doses, mean BG, nadir BG, BG coefficient of variation (CVBG), diet status, type 1 diabetes, type 2 diabetes, acute kidney injury, chronic kidney disease (CKD), liver disease, and digestive disease, our model achieved a c-statistic of 0.77 (95% CI 0.75 to 0.78), positive likelihood ratio (+LR) of 3.5 (95% CI 3.4 to 3.6) and negative likelihood ratio (-LR) of 0.32 (95% CI 0.30 to 0.35) for prediction of biochemical hypoglycemia. Using predictors of sex, weight, insulin doses, mean BG, nadir BG, CVBG, diet status, type 1 diabetes, type 2 diabetes, CKD stage, and steroid use, our model achieved a c-statistic of 0.80 (95% CI 0.78 to 0.82), +LR of 3.8 (95% CI 3.7 to 4.0) and -LR of 0.2 (95% CI 0.2 to 0.3) for prediction of clinically significant hypoglycemia.

CONCLUSIONS

Hospitalized patients at risk of insulin-associated hypoglycemia can be identified using validated prediction models, which may support the development of real-time preventive interventions.

Derivation and validation model for hospital hypoglycemia

BACKGROUND

An objective and simple prognostic model for hospitalized patients with hypoglycemia could be helpful in guiding initial intensity of treatment.

METHODS

We carried out a derivation rule for hypoglycemia using data from a nationwide retrospective cohort study of patients with diabetes or hyperglycemia carried out in 2014 (n=839 patients). The rule for hypoglycemia was validated using a second data set from a nationwide retrospective cohort study carried out in 2016 (n=561 patients). We derived our prediction rule using logistic regression with hypoglycemia (glucose less than 70mg/dL) as the primary outcome.

RESULTS

The incidence of hypoglycemia in the derivation cohort was 10.3%. Patient's characteristics independently associated with hypoglycemia included episodes of hypoglycemia during the previous three months (odds ratio [OR]: 6.29, 95% confidence interval [95%CI]: 3.37-11.79, p<0.001) estimated glomerular filtration rate lower than 30mL/min/1.73m² (OR: 2.32, 95%CI: 1.23-4.35, p=0.009), daily insulin dose greater than 0.3units per Kg (OR: 1.74, 95%CI: 1.06-2.85, p=0.028), and days of hospitalization (OR: 1.03, 95%CI: 1.01-1.04, p=0.001). The model showed an area under the curve (AUC): 0.72 (95%CI: 0.66-0.78, p<0.001). The AUC in the validation cohort was: 0.71 (95%CI: 0.63-0.79, p<0.001).

CONCLUSIONS

The rule showed fair accuracy to predict hypoglycemia. Implementation of the rule into computer systems could be used in guiding initial insulin therapy.

How Low Can You Go? Reducing Rates of Hypoglycemia in the Non-critical Care Hospital Setting

PURPOSE OF REVIEW

The purpose of this review is to discuss strategies to reduce rates of hypoglycemia in the non-critical care setting.

RECENT FINDINGS

Strategies to reduce hypoglycemia rates should focus on the most common causes of iatrogenic hypoglycemia. Creating a standardized insulin order set with built-in clinical decision support can help reduce rates of hypoglycemia. Coordination of blood glucose monitoring, meal tray delivery, and insulin administration is an important and challenging task. Protocols and processes should be in place to deal with interruptions in nutrition to minimize risk of hypoglycemia. A glucose management page that has all the pertinent information summarized in one page allows for active surveillance and quick identification of patients who may be at risk of hypoglycemia. Finally, education of prescribers, nurses, food and nutrition services, and patients is important so that every member of the healthcare team can work together to prevent hypoglycemia. By implementing strategies to reduce hypoglycemia, we hope to lower rates of adverse events and improve quality of care while also reducing hospital costs. Future research should focus on the impact of an overall reduction in hypoglycemia to determine whether the expected benefits are achieved.