Purpose To evaluate the accuracy of various claims‐based definitions of diabetes‐related complications (coronary artery disease [CAD], heart failure, cerebrovascular disease and dialysis). Methods We evaluated data on 1379 inpatients who received care at the Niigata University Medical & Dental Hospital in September 2018. Manual electronic medical chart reviews were conducted for all patients with regard to diabetes‐related complications and were used as the gold standard. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of each claims‐based definition associated with diabetes‐related complications based on Diagnosis Procedure Combination (DPC), International Classification of Diseases, Tenth Revision (ICD‐10) codes, procedure codes and medication codes were calculated. Results DPC‐based definitions had higher sensitivity, specificity, and PPV than ICD‐10 code definitions for CAD and cerebrovascular disease, with sensitivity of 0.963–1.000 and 0.905–0.952, specificity of 1.000 and 1.000, and PPV of 1.000 and 1.000, respectively. Sensitivity, specificity, and PPV were high using procedure codes for CAD and dialysis, with sensitivity of 0.963 and 1.000, specificity of 1.000 and 1.000, and PPV of 1.000 and 1.000, respectively. DPC and/or ICD‐10 codes + medication were better for heart failure than the ICD‐10 code definition, with sensitivity of 0.933, specificity of 1.000, and PPV of 1.000. The PPVs were lower than 60% for all diabetes‐related complications using ICD‐10 codes only. Conclusion The DPC‐based definitions for CAD and cerebrovascular disease, procedure codes for CAD and dialysis, and DPC or ICD‐10 codes with medication codes for heart failure could accurately identify these diabetes‐related complications from claims databases.

Background Machine learning (ML) algorithms have been widely introduced to diabetes research including those for the identification of hypoglycemia. Objective The objective of this meta-analysis is to assess the current ability of ML algorithms to detect hypoglycemia (ie, alert to hypoglycemia coinciding with its symptoms) or predict hypoglycemia (ie, alert to hypoglycemia before its symptoms have occurred). Methods Electronic literature searches (from January 1, 1950, to September 14, 2020) were conducted using the Dialog platform that covers 96 databases of peer-reviewed literature. Included studies had to train the ML algorithm in order to build a model to detect or predict hypoglycemia and test its performance. The set of 2 × 2 data (ie, number of true positives, false positives, true negatives, and false negatives) was pooled with a hierarchical summary receiver operating characteristic model. Results A total of 33 studies (14 studies for detecting hypoglycemia and 19 studies for predicting hypoglycemia) were eligible. For detection of hypoglycemia, pooled estimates (95% CI) of sensitivity, specificity, positive likelihood ratio (PLR), and negative likelihood ratio (NLR) were 0.79 (0.75-0.83), 0.80 (0.64-0.91), 8.05 (4.79-13.51), and 0.18 (0.12-0.27), respectively. For prediction of hypoglycemia, pooled estimates (95% CI) were 0.80 (0.72-0.86) for sensitivity, 0.92 (0.87-0.96) for specificity, 10.42 (5.82-18.65) for PLR, and 0.22 (0.15-0.31) for NLR. Conclusions Current ML algorithms have insufficient ability to detect ongoing hypoglycemia and considerate ability to predict impeding hypoglycemia in patients with diabetes mellitus using hypoglycemic drugs with regard to diagnostic tests in accordance with the Users’ Guide to Medical Literature (PLR should be ≥5 and NLR should be ≤0.2 for moderate reliability). However, it should be emphasized that the clinical applicability of these ML algorithms should be evaluated according to patients’ risk profiles such as for hypoglycemia and its associated complications (eg, arrhythmia, neuroglycopenia) as well as the average ability of the ML algorithms. Continued research is required to develop more accurate ML algorithms than those that currently exist and to enhance the feasibility of applying ML in clinical settings. Trial Registration PROSPERO International Prospective Register of Systematic Reviews CRD42020163682; http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42020163682

Background Efficacy of programs for patients with diabetes mellitus (DM) that have promoted family members to help with patients’ self-care activities has been largely inconsistent. This meta-analysis aims to assess the effect of family-oriented diabetes programs for glycemic control (GC). Methods Electronic literature searches were conducted for clinical trials with a parallel design wherein there were two groups according to whether family members were included (intervention group) or not included (control group) and changes in glycohemoglobin A1C (A1C) were assessed as a study outcome. Each effect size (i.e. difference in A1C change between the intervention and control group) was pooled with a random-effects model. Results There were 31 eligible trials consisting of 1466 and 1415 patients in the intervention and control groups, respectively. Pooled A1C change [95% confidence interval (CI)] was −0.45% (−0.64% to −0.26%). Limiting analyses to 21 trials targeted at patients with type 1 DM or 9 trials targeted at patients with type 2 DM, the pooled A1C changes (95% CI) were −0.35% (−0.55% to −0.14%) and −0.71% (−1.09% to −0.33%), respectively. Conclusion This meta-analysis suggests that focusing on the family as well as the individual patient in self-management diabetes programs to improve the performance of self-care activities of patients with DM is effective in terms of proper GC.

Aims/Introduction Recently, an increasing number of cohort studies have suggested using machine learning (ML) to predict type 2 diabetes mellitus. However, its predictive ability remains inconclusive. This meta‐analysis evaluated the current ability of ML algorithms for predicting incident type 2 diabetes mellitus. Materials and Methods We systematically searched longitudinal studies published from 1 January 1950 to 17 May 2020 using MEDLINE and EMBASE. Included studies had to compare ML’s classification with the actual incidence of type 2 diabetes mellitus, and present data on the number of true positives, false positives, true negatives and false negatives. The dataset for these four values was pooled with a hierarchical summary receiver operating characteristic and a bivariate random effects model. Results There were 12 eligible studies. The pooled sensitivity, specificity, positive likelihood ratio and negative likelihood ratio were 0.81 (95% confidence interval [CI] 0.67–0.90), 0.82 [95% CI 0.74–0.88], 4.55 [95% CI 3.07–6.75] and 0.23 [95% CI 0.13–0.42], respectively. The area under the summarized receiver operating characteristic curve was 0.88 (95% CI 0.85–0.91). Conclusions Current ML algorithms have sufficient ability to help clinicians determine whether individuals will develop type 2 diabetes mellitus in the future. However, persons should be cautious before changing their attitude toward future diabetes risk after learning the result of the diabetes prediction test using ML algorithms.

Purpose To determine the degree of control of multiple risk factors under real-world conditions for coronary artery disease (CAD) according to the presence or absence of diabetes mellitus (DM) and to determine whether reaching multifactorial targets for blood pressure (BP), low-density lipoprotein-cholesterol (LDL-C), HbA1c, and current smoking is associated with lower risks for CAD. Methods We investigated the effects on subsequent CAD of the number of controlled risk factors among BP, LDL-C, HbA1c, and current smoking in a prospective cohort study using a nationwide claims database of 220,894 individuals in Japan. Cox regression examined risks over a 4.8-year follow-up. Results The largest percentage of participants had two risk factors at target in patients with DM (39.6%) and subjects without DM (36.4%). Compared with those who had two targets achieved, the risks of CAD among those who had any one and no target achieved were two and four times greater, respectively, regardless of the presence of DM. The effect of composite control was sufficient to bring CAD risk in patients with DM below that for subjects without DM with any two targets achieved, whereas the risk of CAD in the DM group with all four risk factors uncontrolled was 9.4 times more than in the non-DM group who had achieved two targets. Conclusions These findings show that composite control of modifiable risk factors has a large effect in patients with and without DM. The effect was sufficient to bring CAD risk in patients with DM below that in the non-DM group who had two targets achieved.