Klaus Donsa

Towards Personalization of Diabetes Therapy Using Computerized Decision Support and Machine Learning: Some Open Problems and Challenges

Diabetes mellitus (DM) is a growing global disease which highly affects the individual patient and represents a global health burden with financial impact on national health care systems. Type 1 DM can only be treated with insulin, whereas for patients with type 2 DM a wide range of therapeutic options are available. These options include lifestyle changes such as change of diet and an increase of physical activity, but also administration of oral or injectable antidiabetic drugs. The diabetes therapy, especially with insulin, is complex. Therapy decisions include various medical and life-style related information. Computerized decision support systems (CDSS) aim to improve the treatment process in patient’s self-management but also in institutional care. Therefore, the personalization of the patient’s diabetes treatment is possible at different levels. It can provide medication support and therapy control, which aid to correctly estimate the personal medication requirements and improves the adherence to therapy goals. It also supports long-term disease management, aiming to develop a personalization of care according to the patient’s risk stratification. Personalization of therapy is also facilitated by using new therapy aids like food and activity recognition systems, lifestyle support tools and pattern recognition for insulin therapy optimization. In this work we cover relevant parameters to personalize diabetes therapy, how CDSS can support the therapy process and the role of machine learning in this context. Moreover, we identify open problems and challenges for the personalization of diabetes therapy with focus on decision support systems and machine learning technology.

Standardized Glycemic Management with a Computerized Workflow and Decision Support System for Hospitalized Patients with Type 2 Diabetes on Different Wards.

Abstract Background: This study investigated the efficacy, safety, and usability of standardized glycemic management by a computerized decision support system for non-critically ill hospitalized patients with type 2 diabetes on four different wards. Materials and Methods: In this open, noncontrolled intervention study, glycemic management of 99 patients with type 2 diabetes (62% acute admissions; 41 females; age, 67±11 years; hemoglobin A1c, 65±21 mmol/mol; body mass index, 30.4±6.5 kg/m2) on clinical wards (Cardiology, Endocrinology, Nephrology, Plastic Surgery) of a tertiary-care hospital was guided by GlucoTab® (Joanneum Research GmbH [Graz, Austria] and Medical University of Graz [Graz, Austria]), a mobile decision support system providing automated workflow support and suggestions for insulin dosing to nurses and physicians. Results: Adherence to insulin dosing suggestions was high (96.5% bolus, 96.7% basal). The primary outcome measure, percentage of blood glucose (BG) measurements in the range of 70–140 mg/dL, occurred in 50.2±22.2% of all measurements. The overall mean BG level was 154±35 mg/dL. BG measurements in the ranges of 60–70 mg/dL, 40–60 mg/dL, and <40 mg/dL occurred in 1.4%, 0.5%, and 0.0% of all measurements, respectively. A regression analysis showed that acute admission to the Cardiology Ward (+30 mg/dL) and preexisting home insulin therapy (+26 mg/dL) had the strongest impact on mean BG. Acute admission to other wards had minor effects (+4 mg/dL). Ninety-one percent of the healthcare professionals felt confident with GlucoTab, and 89% believed in its practicality and 80% in its ability to prevent medication errors. Conclusions: An efficacious, safe, and user-accepted implementation of GlucoTab was demonstrated. However, for optimized personalized patient care, further algorithm modifications are required.

Impact of errors in paper-based and computerized diabetes management with decision support for hospitalized patients with type 2 diabetes. A post-hoc analysis of a before and after study

OBJECTIVE Most preventable adverse drug events and medication errors occur during medication ordering. Medication order entry and clinical decision support are available on paper or as computerized systems. In this post-hoc analysis we investigated frequency and clinical impact of blood glucose (BG) documentation- and user-related calculation errors as well as workflow deviations in diabetes management. We aimed to compare a paper-based protocol to a computerized medication management system combined with clinical workflow and decision support. METHODS Seventy-nine hospitalized patients with type 2 diabetes mellitus were treated with an algorithm driven basal-bolus insulin regimen. BG measurements, which were the basis for insulin dose calculations, were manually entered either into the paper-based workflow protocol (PaperG: 37 patients) or into GlucoTab(®)-a mobile tablet PC based system (CompG: 42 patients). We used BG values from the laboratory information system as a reference. A workflow simulator was used to determine user calculation errors as well as workflow deviations and to estimate the effect of errors on insulin doses. The clinical impact of insulin dosing errors and workflow deviations on hypo- and hyperglycemia was investigated. RESULTS The BG documentation error rate was similar for PaperG (4.9%) and CompG group (4.0%). In PaperG group, 11.1% of manual insulin dose calculations were erroneous and the odds ratio (OR) of a hypoglycemic event following an insulin dosing error was 3.1 (95% CI: 1.4-6.8). The number of BG values influenced by insulin dosing errors was eightfold higher than in the CompG group. In the CompG group, workflow deviations occurred in 5.0% of the tasks which led to an increased likelihood of hyperglycemia, OR 2.2 (95% CI: 1.1-4.6). DISCUSSION Manual insulin dose calculations were the major source of error and had a particularly strong influence on hypoglycemia. By using GlucoTab(®), user calculation errors were entirely excluded. The immediate availability and automated handling of BG values from medical devices directly at the point of care has a high potential to reduce errors. Computerized systems facilitate the safe use of more complex insulin dosing algorithms without compromising usability. In CompG group, missed or delayed tasks had a significant effect on hyperglycemia, while in PaperG group insufficient precision of documentation times limited analysis. The use of old BG measurements was clinically less relevant. CONCLUSION Insulin dosing errors and workflow deviations led to measurable changes in clinical outcome. Diabetes management systems including decision support should address nurses as well as physicians in a computerized way. Our analysis shows that such systems reduce the frequency of errors and therefore decrease the probability of hypo- and hyperglycemia.

A Mobile Computerized Decision Support System to Prevent Hypoglycemia in Hospitalized Patients With Type 2 Diabetes Mellitus: Lessons Learned From a Clinical Feasibility Study

Background: Diabetes management requires complex and interdisciplinary cooperation of health care professionals (HCPs). To support this complex process, IT-support is recommended by clinical guidelines. The aim of this article is to report on results from a clinical feasibility study testing the prototype of a mobile, tablet-based client-server system for computerized decision and workflow support (GlucoTab®) and to discuss its impact on hypoglycemia prevention. Methods: The system was tested in a monocentric, open, noncontrolled intervention study in 30 patients with type 2 diabetes mellitus (T2DM). The system supports HCPs in performing a basal-bolus insulin therapy. Diabetes therapy, adverse events, software errors and user feedback were documented. Safety, efficacy and user acceptance of the system were investigated. Results: Only 1.3% of blood glucose (BG) measurements were <70 mg/dl and only 2.6% were >300 mg/dl. The availability of the system (97.3%) and the rate of treatment activities documented with the system (>93.5%) were high. Only few suggestions from the system were overruled by the users (>95.7% adherence). Evaluation of the 3 anonymous questionnaires showed that confidence in the system increased over time. The majority of users believed that treatment errors could be prevented by using this system. Conclusions: Data from our feasibility study show a significant reduction of hypoglycemia by implementing a computerized system for workflow and decision support for diabetes management, compared to a paper-based process. The system was well accepted by HCPs, which is shown in the user acceptance analysis and that users adhered to the insulin dose suggestions made by the system.

A Toolbox to Improve Algorithms for Insulin-Dosing Decision Support

BACKGROUND Standardized insulin order sets for subcutaneous basal-bolus insulin therapy are recommended by clinical guidelines for the inpatient management of diabetes. The algorithm based GlucoTab system electronically assists health care personnel by supporting clinical workflow and providing insulin-dose suggestions. OBJECTIVE To develop a toolbox for improving clinical decision-support algorithms. METHODS The toolbox has three main components. 1) Data preparation: Data from several heterogeneous sources is extracted, cleaned and stored in a uniform data format. 2) Simulation: The effects of algorithm modifications are estimated by simulating treatment workflows based on real data from clinical trials. 3) ANALYSIS: Algorithm performance is measured, analyzed and simulated by using data from three clinical trials with a total of 166 patients. RESULTS Use of the toolbox led to algorithm improvements as well as the detection of potential individualized subgroup-specific algorithms. CONCLUSION These results are a first step towards individualized algorithm modifications for specific patient subgroups.

GlucoTab guided insulin therapy using insulin glargine U300 enables glycaemic control with low risk of hypoglycaemia in hospitalised patients with type 2 diabetes

To investigate efficacy, safety and usability of the GlucoTab system for glycaemic management using insulin glargine U300 in non‐critically ill hospitalized patients with type 2 diabetes (T2D).