Using Oregon Medicaid administrative data from 1 January 2014 to 30 September 2016, a retrospective cohort study was conducted with propensity score–matched Medicaid eligibility groups (newly and previously eligible). Outcome measures included total per-member per-month (PMPM) Medicaid expenditures and PMPM expenditures in the following 12 categories: inpatient visits, emergency department visits, primary care physician visits, specialist visits, prescription drugs, transportation services, tests, imaging and echography, procedures, durable medical equipment, evaluation and management, and other or unknown services.

Total PMPM Medicaid expenditures for newly eligible enrollees with diabetes were initially considerably lower compared with PMPM expenditures for matched previously eligible enrollees during the first postexpansion quarter (mean values $561 vs. $793 PMPM, P = 0.018). Within the first three postexpansion quarters, PMPM expenditures of the newly eligible increased to a similar but slightly lower level. Afterward, PMPM expenditures of both groups continued to increase steadily. Most of the overall PMPM expenditure increase among the newly eligible was due to rapidly increasing prescription drug expenditures.

Newly eligible Medicaid enrollees with diabetes had slightly lower PMPM expenditures than previously eligible Medicaid enrollees. The increase in PMPM prescription drug expenditures suggests greater access to treatment over time.

We included 17,026 adults (8,346 men and 8,680 women) who were part of the China Health and Nutrition Survey (1991–2015) prospective cohort. Dietary intake was measured by three consecutive 24-h dietary recalls combined with a household food inventory. Diabetes cases were identified through a questionnaire. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% CIs.

A total of 547 men and 577 women developed diabetes during 202,138 person-years of follow-up. For men, the adjusted HRs (95% CIs) for quintiles of nonheme iron intake were 1.00, 0.77 (0.58–1.02), 0.72 (0.54–0.97), 0.63 (0.46–0.85), and 0.87 (0.64–1.19) (P-nonlinearity = 0.0015). The corresponding HRs (95% CIs) for women were 1.00, 0.63 (0.48–0.84), 0.57 (0.43–0.76), 0.58 (0.43–0.77), and 0.67 (0.49–0.91) (P-nonlinearity < 0.0001). The dose-response curves for the association between nonheme iron and total iron intake and diabetes followed a reverse J shape in men and an L shape in women. No significant associations were observed between heme iron intake and diabetes risk.

Total iron and nonheme iron intake was associated with diabetes risk, following a reverse J-shaped curve in men and an L-shaped curve in women. Sufficient intake of nonheme or total iron might be protective against diabetes, while excessive iron intake might increase the risk of diabetes among men.

Data from 2,662 individuals with self-reported diabetes who participated in the Health and Retirement Study (HRS) were used. Participants were followed from 2006 through 2014. Financial hardship, psychosocial, and neighborhood-level social determinant factors were based on validated surveys from the biennial core interview and RAND data sets. All social determinant factors and measurements of HbA_1c from the time period were used and treated as time varying in analyses. SAS PROC GLIMMIX was used to fit a series of hierarchical linear mixed models. Models controlled for nonindependence among the repeated observations using a random intercept and treating each individual participant as a random factor. Survey methods were used to apply HRS weighting.

Before adjustment for demographics, difficulty paying bills (β = 0.18 [95% CI 0.02, 0.24]) and medication cost nonadherence (0.15 [0.01, 0.29]) were independently associated with increasing HbA_1c over time, and social cohesion (–0.05 [–0.10, –0.001]) was independently associated with decreasing HbA_1c over time. After adjusting for both demographics and comorbidity count, difficulty paying bills (0.13 [0.03, 0.24]) and religiosity (0.04 [0.001, 0.08]) were independently associated with increasing HbA_1c over time.

Using a longitudinal cohort of older adults with diabetes, this study found that financial hardship factors, such as difficulty paying bills, were more consistently associated with worsening glycemic control over time than psychosocial and neighborhood factors.

Data on health care resource utilization from 100,391 people with type 2 diabetes were extracted from the electronic database used at the Catalan Health Institute. Multivariate regression models were carried out to test the impact of glycemic control (HbA_1c) on total health care, hospital admission, and medication costs; model 1 adjusted for a variety of covariates, and model 2 also included micro- and macrovascular complications. Glycemic control was classified as good for HbA_1c <7%, fair for ≥7% to <8%, poor for ≥8% to <10%, and very poor for ≥10%.

Mean per patient annual direct medical costs were 3,039 ± SD 6,581. Worse glycemic control was associated with higher total health care costs: compared with good glycemic control, health care costs increased by 18% (509.82) and 23% (661.35) in patients with very poor and poor glycemic control, respectively, when unadjusted and by 428.3 and 395.1, respectively, in model 2. Medication costs increased by 12% in patients with fair control and by 28% in those with very poor control (model 2). Patients with poor control had a higher probability of hospitalization than those with good control (5% in model 2) and a greater average cost when hospitalization occurred (811).

Poor glycemic control was directly related to higher total health care, hospitalization, and medication costs. Preventive strategies and good glycemic control in people with type 2 diabetes could reduce the economic impact associated with this disease.

We measured inpatient (n = 816) health care utilization for Bronx A1C participants using an administrative data set containing all hospital discharges for New York State from 2006 to 2014. Multilevel mixed modeling was used to assess changes in health care utilization and costs between the telephonic diabetes intervention (Tele/Pr) arm and print-only (PrO) control arm.

During follow-up, excess relative reductions in all-cause hospitalizations for the Tele/Pr arm compared with PrO arm were statistically significant for odds of hospital use (odds ratio [OR] 0.89; 95% CI 0.82, 0.97; P < 0.01), number of hospital stays (rate ratio [RR] 0.90; 95% CI 0.81, 0.99; P = 0.04), and hospital costs (RR 0.90; 95% CI 0.84, 0.98; P = 0.01). Reductions in hospital use and costs were even stronger for diabetes-related hospitalizations. These outcomes were not significantly related to changes observed in hemoglobin A_1c during individuals’ participation in the 1-year intervention.

These results indicate that the impact of the Bronx A1C intervention was not just on short-term improvements in glycemic control but also on long-term health care utilization. This finding is important because it suggests the benefits of the intervention were long-lasting with the potential to not only reduce hospitalizations but also to lower hospital-associated costs.

We conducted a retrospective, cross-sectional analysis of 2003–2016 National Health and Nutrition Examination Survey (NHANES) data. We included people ≥18 years who had ever been told they had diabetes, had an HbA_1C >6.4%, or had a fasting plasma glucose >125 mg/dL. Pregnant women, and those aged <20 years receiving only insulin were excluded. We assessed trends in use of ADA’s seven preferred classes from 2003–2004 to 2015–2016. We also examined use by hypoglycemia risk (sulfonylureas, insulin, and meglitinides), weight effect (sulfonylureas, thiazolidinediones [TZDs], insulin, and meglitinides), cardiovascular benefit (canagliflozin, empagliflozin, and liraglutide), and cost (brand-name medications and insulin analogs).

The final sample included 6,323 patients. The proportion taking any medication increased from 58% in 2003–2004 to 67% in 2015–2016 (P < 0.001). Use of metformin and insulin analogs increased, while use of sulfonylureas, TZDs, and human insulin decreased. Following the 2012 ADA recommendation, the choice of drug did not vary significantly by older age, weight, or presence of cardiovascular disease. Patients with low HbA_1C, or HbA_1C <6%, and age ≥65 years were less likely to receive hypoglycemia-inducing medications, while older patients with comorbidities were more likely. Insurance, but not income, was associated with the use of higher-cost medications.

Following ADA recommendations, the use of metformin increased, but physicians generally did not individualize treatment according to patients’ characteristics. Substantial opportunities exist to improve pharmacologic management of diabetes.

We included 2,236 adults aged ≥65 years without known diabetes from the 2005–2016 National Health and Nutrition Examination Survey. Diabetes was defined using: 1) the Endocrine Society approach (HbA_1c ≥6.5%, FPG ≥126 mg/dL, or 2-h plasma glucose ≥200 mg/dL among those with HbA_1c 5.7–6.4% or FPG 100–125 mg/dL); and 2) a standard approach (HbA_1c ≥6.5% or FPG ≥126 mg/dL). Treatment eligibility was defined using HbA_1c cut points (≥7 to ≥9%). OGTT screening costs were estimated using Medicare fee schedules.

Diabetes prevalence was 15.7% (~5.0 million) using the Endocrine Society’s approach and 7.3% (~2.3 million) using the standard approach. Treatment eligibility ranged from 5.4 to 0.06% and 11.8–1.3% for diabetes cases identified through the Endocrine Society or standard approach, respectively. By definition, diabetes identified exclusively through the Endocrine Society approach had HbA1_1c <6.5% and would not be recommended for glucose-lowering treatment. Screening all older adults with prediabetic HbA_1c/FPG (~18.3 million) with OGTT could cost between $737 million and $1.7 billion.

Adopting the 2019 Endocrine Society guidelines would substantially increase the number of older adults classified as having diabetes, require significant financial resources, but likely offer limited benefits.

We included 84,285 postmenopausal women without a history of diabetes from the national Women’s Health Initiative Observational Study (WHI-OS). Replication analysis was then conducted among 62,338 women who participated in the WHI-Clinical Trial (WHI-CT). Additionally, data from a case-control study of 3,749 women nested in the WHI-OS with information on biomarkers of inflammation and endothelial dysfunction were examined using mediation analysis to determine the relative contributions of these known biomarkers by which manganese affects type 2 diabetes risk.

Compared with the lowest quintile of energy-adjusted dietary manganese, WHI-OS participants in the highest quintile had a 30% lower risk of type 2 diabetes (hazard ratio [HR] 0.70 [95% CI 0.65, 0.76]). A consistent association was also confirmed in the WHI-CT (HR 0.79 [95% CI 0.73, 0.85]). In the nested case-control study, higher energy-adjusted dietary manganese was associated with lower circulating levels of inflammatory biomarkers that significantly mediated the association between dietary manganese and type 2 diabetes risk. Specifically, 19% and 12% of type 2 diabetes risk due to manganese were mediated through interleukin 6 and hs-CRP, respectively.

Higher intake of manganese was directly associated with a lower type 2 diabetes risk independent of known risk factors. This association may be partially mediated by inflammatory biomarkers.

Four years of data were extracted from a hospital electronic health record system. This included laboratory and point-of-care blood glucose (BG) values to identify biochemical and clinically significant hypoglycemic episodes (BG ≤3.9 and ≤2.9 mmol/L, respectively). We used patient demographics, administered medications, vital signs, laboratory results, and procedures performed during the hospital stays to inform the model. Two iterations of the data set included the doses of insulin administered and the past history of inpatient hypoglycemia. Eighteen different prediction models were compared using the area under the receiver operating characteristic curve (AUROC) through a 10-fold cross validation.

We analyzed data obtained from 17,658 inpatients with diabetes who underwent 32,758 admissions between July 2014 and August 2018. The predictive factors from the logistic regression model included people undergoing procedures, weight, type of diabetes, oxygen saturation level, use of medications (insulin, sulfonylurea, and metformin), and albumin levels. The machine learning model with the best performance was the XGBoost model (AUROC 0.96). This outperformed the logistic regression model, which had an AUROC of 0.75 for the estimation of the risk of clinically significant hypoglycemia.

Advanced machine learning models are superior to logistic regression models in predicting the risk of hypoglycemia in inpatients with diabetes. Trials of such models should be conducted in real time to evaluate their utility to reduce inpatient hypoglycemia.

A total of 4,699 people with diabetes who enrolled in the National Health and Nutrition Examination Survey from 2003 to 2014 were recruited for this study. Energy and macronutrient intake was measured by a 24-h dietary recall. The differences () in energy and macronutrient intake between dinner and breakfast ( = dinner – breakfast) were categorized into quintiles. Death information was obtained from the National Death Index until 2015. Cox proportional hazards regression models were developed to evaluate the survival relationship between and diabetes, cardiovascular disease (CVD), and all-cause mortality.

Among the 4,699 participants, 913 deaths, including 269 deaths due to diabetes and 314 deaths due to CVD, were documented. After adjustment for potential confounders, compared with participants in the lowest quintile of in terms of total energy and protein, participants in the highest quintile were more likely to die due to diabetes (hazard ratio [HR]_energy 1.92, 99% CI 1.08–3.42; HR_protein 1.92, 99% CI 1.06–3.49) and CVD (HR_energy 1.69, 99% CI 1.02–2.80; HR_protein 1.96, 99% CI 1.14–3.39). The highest quintile of total fat was related to CVD mortality (HR 1.67, 99% CI 1.01–2.76). Isocalorically replacing 5% of total energy at dinner with breakfast was associated with 4% and 5% lower risk of diabetes (HR 0.96, 95% CI 0.94–0.98) and CVD (HR 0.95, 95% CI 0.93–0.97) mortality, respectively.

Higher intake of energy, total fat, and protein from dinner than breakfast was associated with greater diabetes, CVD, and all-cause mortality in people with diabetes.