Clinical Prediction Tool To Identify Adults With Type 2 Diabetes at Risk for Persistent Adverse Glycemia in Hospital

Risk factors for hypoglycaemia in non-critical hospitalised diabetic patients

OBJECTIVE

To determine the risk factors for hypoglycaemia in patients with diabetes on general hospital wards based on a systematic review of the literature since 2013 and meta-analysis.

METHODS

Systematic review of the literature focused on the conceptual and methodological aspects of the PRISMA Declaration. The search carried out in Pub Med, Web of Science, Medline, Scielo, Lilacs, OVID, grey literature and Google Academic focused on risk factors for hypoglycaemia in patients with diabetes on general hospital wards. The CASPe (Critical Appraisal Skills Programme Spanish) tool was applied for quality control.

RESULTS

From 805 references, 70 potentially eligible articles were identified for review of abstracts and full text. Finally, according to inclusion and exclusion criteria, seven studies with 554,601 patients of Asian, European and North American ethnicity were selected. A meta-analysis performed using the random effects model found an association between the presence of hypoglycaemia and: the use of insulin (OR 2.89 [95% CI: 1.8-4.5]); the use of long-acting insulin (OR 2.27 [95% CI: 1.8-2.8]) or fast-acting insulin (OR 1.4 [95% CI: 1.18-1.85]); nasogastric tube feeding (OR 1.75 [95% CI: 1.33-2.3]); chronic kidney disease (OR 1.65 [95% CI: 1.14-2.38]); congestive heart failure (OR 1.36 [95% CI: 1.10-1.68]); and elevated levels of glycosylated haemoglobin (OR 1.59 [95% CI: 1.32-1.91]).

CONCLUSION

The factors associated with the risk of hypoglycaemia in non-critically ill hospitalised patients with type 2 diabetes were: use of any insulin; nasogastric tube feeding; elevated glycosylated haemoglobin levels; history of congestive heart failure; and chronic kidney disease.

Longitudinal Digital Glucometric Benchmarking to Evaluate the Impact of Institutional Diabetes Care Initiatives in Adults With Diabetes Mellitus Over the 2016-2020 Period

BACKGROUND

While glucometric benchmarking has been used to compare glucose management between institutions, the value of longitudinal intra-institution benchmarking to assess quality improvement changes is not established.

METHODS

A prospective six-month observational study (October 2019-March 2020 inclusive) of inpatients with diabetes or newly detected hyperglycemia admitted to eight medical and surgical wards at the Royal Melbourne Hospital. Networked blood glucose (BG) meters were used to collect capillary BG levels. Outcomes were measures of glycemic control assessed by mean and threshold glucometric measures and comparison with published glucometric benchmarks. Intra-institution comparison was over the 2016-2020 period.

RESULTS

In all, 620 admissions (588 unique individuals) met the inclusion criteria, contributing 15 164 BG results over 4023 admission-days. Compared with the 2016 cohort from the same institution, there was increased BG testing (3.8 [SD = 2.2) vs 3.3 [SD = 1.7] BG measurements per patient-day, P < .001), lower mean patient-day mean glucose (PDMG; 8.9 mmol/L [SD = 3.2] vs 9.5 mmol/L [SD = 3.3], P < .001), and reduced mean and threshold measures of hyperglycemia (P < .001 for all). Comparison with institutions across the United States revealed lower incidence of mean PDMG >13.9 or >16.7 mmol/L, and reduced hypoglycemia (<3.9, <2.8, and <2.2 mmol/L), when compared with published benchmarks from an earlier period (2009-2014).

CONCLUSIONS

Comprehensive digital-based glucometric benchmarking confirmed institutional quality improvement changes were followed by reduced hyperglycemia and hypoglycemia in a five-year comparison. Longitudinal glucometric benchmarking enables evaluation and validation of changes to institutional diabetes care management practices.

Glycaemic outcomes in hospital with IDegAsp versus BIAsp30 premixed insulins

BACKGROUND AND AIMS

IDegAsp (Ryzodeg 70/30), a unique premixed formulation of long-acting insulin degludec and rapid-acting insulin aspart, is increasing in use. Management of IDegAsp during hospitalisation is challenging because of degludec's ultra-long duration of action. We investigated inpatient glycaemia in patients treated with IDegAsp compared to biphasic insulin aspart (BIAsp30; Novomix30).

METHODS

We performed a retrospective observational study at two hospitals assessing inpatients with type 2 diabetes treated with IDegAsp or BIAsp30 prior to and during hospital admission. Standard inpatient glycaemic outcomes were analysed based on capillary blood glucose (BG) measurements.

RESULTS

We assessed 88 individuals treated with IDegAsp and 88 HbA1c-matched individuals treated with BIAsp30. Patient characteristics, including insulin dose at admission, were well matched, but the IDegAsp group had less frequent twice-daily insulin dosing than the BIAsp30 group (49% vs 87%, P < 0.001). Patient-days with BG <4 mmol/L were not different (10.6% vs 9.9%, P = 0.7); however, the IDegAsp group had a higher patient-day mean BG (10.4 (SD 3.4) vs 10.0 (3.4) mmol/L, P < 0.001), and more patient-days with mean BG >10 mmol/L (48% vs 38%, P < 0.001) compared to the BIAsp30 group. Glucose was higher in the IDegAsp group in the evening (4 PM to midnight) (11.6 (SD 4.0) vs 10.9 (4.6) mmol/L, P = 0.004), but not different at other times during the day.

CONCLUSIONS

Inpatients treated with IDegAsp compared to BIAsp30 had similar hypoglycaemia incidence, but higher hyperglycaemia incidence, potentially related to less frequent twice-daily dosing. With the increasing use of IDegAsp in the community, development of hospital management guidelines for this insulin formulation is needed.

Data-based modeling for hypoglycemia prediction: Importance, trends, and implications for clinical practice

Background and objective

Hypoglycemia is a key barrier to achieving optimal glycemic control in people with diabetes, which has been proven to cause a set of deleterious outcomes, such as impaired cognition, increased cardiovascular disease, and mortality. Hypoglycemia prediction has come to play a role in diabetes management as big data analysis and machine learning (ML) approaches have become increasingly prevalent in recent years. As a result, a review is needed to summarize the existing prediction algorithms and models to guide better clinical practice in hypoglycemia prevention.

Materials and methods

PubMed, EMBASE, and the Cochrane Library were searched for relevant studies published between 1 January 2015 and 8 December 2022. Five hypoglycemia prediction aspects were covered: real-time hypoglycemia, mild and severe hypoglycemia, nocturnal hypoglycemia, inpatient hypoglycemia, and other hypoglycemia (postprandial, exercise-related).

Results

From the 5,042 records retrieved, we included 79 studies in our analysis. Two major categories of prediction models are identified by an overview of the chosen studies: simple or logistic regression models based on clinical data and data-based ML models (continuous glucose monitoring data is most commonly used). Models utilizing clinical data have identified a variety of risk factors that can lead to hypoglycemic events. Data-driven models based on various techniques such as neural networks, autoregressive, ensemble learning, supervised learning, and mathematical formulas have also revealed suggestive features in cases of hypoglycemia prediction.

Conclusion

In this study, we looked deep into the currently established hypoglycemia prediction models and identified hypoglycemia risk factors from various perspectives, which may provide readers with a better understanding of future trends in this topic.

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.

Safety and Efficacy of Inpatient Diabetes Management with Non-insulin Agents: an Overview of International Practices

PURPOSE OF REVIEW

The field of inpatient diabetes has advanced significantly over the last 20 years, leading to the development of personalized treatment approaches. However, outdated guidelines still recommend the use of basal-bolus insulin therapy as the preferred treatment approach, and against the use of non-insulin anti-hyperglycemic agents.

RECENT FINDINGS

Several observational and prospective randomized controlled studies have demonstrated that oral anti-hyperglycemic agents are widely used in the hospital, including studies of DPP-4 agents and GLP-1 agonists. With advances in the field of inpatient diabetes management, a paradigm shift has occurred, from an approach of recommending "basal-bolus regimens" for all patients to a more precision medicine option for hospitalized non-critically ill patients with type 2 diabetes.

Development and validation of a machine learning model for classification of next glucose measurement in hospitalized patients

BACKGROUND

Inpatient glucose management can be challenging due to evolving factors that influence a patient's blood glucose (BG) throughout hospital admission. The purpose of our study was to predict the category of a patient's next BG measurement based on electronic medical record (EMR) data.

METHODS

EMR data from 184,361 admissions containing 4,538,418 BG measurements from five hospitals in the Johns Hopkins Health System were collected from patients who were discharged between January 1, 2015 and May 31, 2019. Index BGs used for prediction included the 5th to penultimate BG measurements (N = 2,740,539). The outcome was category of next BG measurement: hypoglycemic (BG  ≤  70 mg/dl), controlled (BG 71-180 mg/dl), or hyperglycemic (BG > 180 mg/dl). A random forest algorithm that included a broad range of clinical covariates predicted the outcome and was validated internally and externally.

FINDINGS

In our internal validation test set, 72·8%, 25·7%, and 1·5% of BG measurements occurring after the index BG were controlled, hyperglycemic, and hypoglycemic respectively. The sensitivity/specificity for prediction of controlled, hyperglycemic, and hypoglycemic were 0·77/0·81, 0·77/0·89, and 0·73/0·91, respectively. On external validation in four hospitals, the ranges of sensitivity/specificity for prediction of controlled, hyperglycemic, and hypoglycemic were 0·64-0·70/0·80-0·87, 0·75-0·80/0·82-0·84, and 0·76-0·78/0·87-0·90, respectively.

INTERPRETATION

A machine learning algorithm using EMR data can accurately predict the category of a hospitalized patient's next BG measurement. Further studies should determine the effectiveness of integration of this model into the EMR in reducing rates of hypoglycemia and hyperglycemia.