scholarly journals Effective Treatment Recommendations for Type 2 Diabetes Management Using Reinforcement Learning: Treatment Recommendation Model Development and Validation (Preprint)

2021 ◽  
Author(s):  
Xingzhi Sun ◽  
Yong Mong Bee ◽  
Shao Wei Lam ◽  
Zhuo Liu ◽  
Wei Zhao ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Type 2 Diabetes ◽  
Reinforcement Learning ◽  
Blood Lipids ◽  
Diabetes Complications ◽  
Data Driven ◽  
Lipid Lowering ◽  
Treatment Recommendation

BACKGROUND Type 2 diabetes mellitus (T2DM) and its related complications represent a growing economic burden for many countries and health systems. Diabetes complications can be prevented through better disease control, but there is a large gap between the recommended treatment and the treatment that patients actually receive. The treatment of T2DM can be challenging because of different comprehensive therapeutic targets and individual variability of the patients, leading to the need for precise, personalized treatment. OBJECTIVE The aim of this study was to develop treatment recommendation models for T2DM based on deep reinforcement learning. A retrospective analysis was then performed to evaluate the reliability and effectiveness of the models. METHODS The data used in our study were collected from the Singapore Health Services Diabetes Registry, encompassing 189,520 patients with T2DM, including 6,407,958 outpatient visits from 2013 to 2018. The treatment recommendation model was built based on 80% of the dataset and its effectiveness was evaluated with the remaining 20% of data. Three treatment recommendation models were developed for antiglycemic, antihypertensive, and lipid-lowering treatments by combining a knowledge-driven model and a data-driven model. The knowledge-driven model, based on clinical guidelines and expert experiences, was first applied to select the candidate medications. The data-driven model, based on deep reinforcement learning, was used to rank the candidates according to the expected clinical outcomes. To evaluate the models, short-term outcomes were compared between the model-concordant treatments and the model-nonconcordant treatments with confounder adjustment by stratification, propensity score weighting, and multivariate regression. For long-term outcomes, model-concordant rates were included as independent variables to evaluate if the combined antiglycemic, antihypertensive, and lipid-lowering treatments had a positive impact on reduction of long-term complication occurrence or death at the patient level via multivariate logistic regression. RESULTS The test data consisted of 36,993 patients for evaluating the effectiveness of the three treatment recommendation models. In 43.3% of patient visits, the antiglycemic medications recommended by the model were concordant with the actual prescriptions of the physicians. The concordant rates for antihypertensive medications and lipid-lowering medications were 51.3% and 58.9%, respectively. The evaluation results also showed that model-concordant treatments were associated with better glycemic control (odds ratio [OR] 1.73, 95% CI 1.69-1.76), blood pressure control (OR 1.26, 95% CI, 1.23-1.29), and blood lipids control (OR 1.28, 95% CI 1.22-1.35). We also found that patients with more model-concordant treatments were associated with a lower risk of diabetes complications (including 3 macrovascular and 2 microvascular complications) and death, suggesting that the models have the potential of achieving better outcomes in the long term. CONCLUSIONS Comprehensive management by combining knowledge-driven and data-driven models has good potential to help physicians improve the clinical outcomes of patients with T2DM; achieving good control on blood glucose, blood pressure, and blood lipids; and reducing the risk of diabetes complications in the long term.

scholarly journals Effective Treatment Recommendations for Type 2 Diabetes Management Using Reinforcement Learning: Treatment Recommendation Model Development and Validation

10.2196/27858 ◽  
2021 ◽  
Vol 23 (7) ◽  
pp. e27858
Author(s):  
Xingzhi Sun ◽  
Yong Mong Bee ◽  
Shao Wei Lam ◽  
Zhuo Liu ◽  
Wei Zhao ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Type 2 Diabetes ◽  
Reinforcement Learning ◽  
Blood Lipids ◽  
Diabetes Complications ◽  
Data Driven ◽  
Lipid Lowering ◽  
Treatment Recommendation

Background Type 2 diabetes mellitus (T2DM) and its related complications represent a growing economic burden for many countries and health systems. Diabetes complications can be prevented through better disease control, but there is a large gap between the recommended treatment and the treatment that patients actually receive. The treatment of T2DM can be challenging because of different comprehensive therapeutic targets and individual variability of the patients, leading to the need for precise, personalized treatment. Objective The aim of this study was to develop treatment recommendation models for T2DM based on deep reinforcement learning. A retrospective analysis was then performed to evaluate the reliability and effectiveness of the models. Methods The data used in our study were collected from the Singapore Health Services Diabetes Registry, encompassing 189,520 patients with T2DM, including 6,407,958 outpatient visits from 2013 to 2018. The treatment recommendation model was built based on 80% of the dataset and its effectiveness was evaluated with the remaining 20% of data. Three treatment recommendation models were developed for antiglycemic, antihypertensive, and lipid-lowering treatments by combining a knowledge-driven model and a data-driven model. The knowledge-driven model, based on clinical guidelines and expert experiences, was first applied to select the candidate medications. The data-driven model, based on deep reinforcement learning, was used to rank the candidates according to the expected clinical outcomes. To evaluate the models, short-term outcomes were compared between the model-concordant treatments and the model-nonconcordant treatments with confounder adjustment by stratification, propensity score weighting, and multivariate regression. For long-term outcomes, model-concordant rates were included as independent variables to evaluate if the combined antiglycemic, antihypertensive, and lipid-lowering treatments had a positive impact on reduction of long-term complication occurrence or death at the patient level via multivariate logistic regression. Results The test data consisted of 36,993 patients for evaluating the effectiveness of the three treatment recommendation models. In 43.3% of patient visits, the antiglycemic medications recommended by the model were concordant with the actual prescriptions of the physicians. The concordant rates for antihypertensive medications and lipid-lowering medications were 51.3% and 58.9%, respectively. The evaluation results also showed that model-concordant treatments were associated with better glycemic control (odds ratio [OR] 1.73, 95% CI 1.69-1.76), blood pressure control (OR 1.26, 95% CI, 1.23-1.29), and blood lipids control (OR 1.28, 95% CI 1.22-1.35). We also found that patients with more model-concordant treatments were associated with a lower risk of diabetes complications (including 3 macrovascular and 2 microvascular complications) and death, suggesting that the models have the potential of achieving better outcomes in the long term. Conclusions Comprehensive management by combining knowledge-driven and data-driven models has good potential to help physicians improve the clinical outcomes of patients with T2DM; achieving good control on blood glucose, blood pressure, and blood lipids; and reducing the risk of diabetes complications in the long term.

scholarly journals Low-cost exercise interventions improve long-term cardiometabolic health independently of a family history of type 2 diabetes: a randomized parallel group trial

2020 ◽  
Vol 8 (2) ◽  
pp. e001377
Author(s):  
Niko S Wasenius ◽  
Bo A Isomaa ◽  
Bjarne Östman ◽  
Johan Söderström ◽  
Björn Forsén ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Type 2 Diabetes ◽  
Exercise Prescription ◽  
Low Cost ◽  
Density Lipoprotein ◽  
Cardiometabolic Health ◽  
Supervised Exercise ◽  
History Of

IntroductionTo investigate the effect of an exercise prescription and a 1-year supervised exercise intervention, and the modifying effect of the family history of type 2 diabetes (FH), on long-term cardiometabolic health.Research design and methodsFor this prospective randomized trial, we recruited non-diabetic participants with poor fitness (n=1072, 30–70 years). Participants were randomly assigned with stratification for FH either in the exercise prescription group (PG, n=144) or the supervised exercise group (EG, n=146) group and compared with a matched control group from the same population study (CON, n=782). The PG and EG received exercise prescriptions. In addition, the EG attended supervised exercise sessions two times a week for 60 min for 12 months. Cardiometabolic risk factors were measured at baseline, 1 year, 5 years, and 6 years. The CON group received no intervention and was measured at baseline and 6 years.ResultsThe EG reduced their body weight, waist circumference, diastolic blood pressure, and low-density lipoprotein-cholesterol (LDL-C) but not physical fitness (p=0.074) or insulin or glucose regulation (p>0.1) compared with the PG at 1 year and 5 years (p≤0.011). The observed differences were attenuated at 6 years; however, participants in the both intervention groups significantly improved their blood pressure, high-density lipoprotein-cholesterol, and insulin sensitivity compared with the population controls (p≤0.003). FH modified LDL-C and waist circumference responses to exercise at 1 year and 5 years.ConclusionsLow-cost physical activity programs have long-term beneficial effects on cardiometabolic health regardless of the FH of diabetes. Given the feasibility and low cost of these programs, they should be advocated to promote cardiometabolic health.Trial registration numberClinicalTrials.gov identifier NCT02131701.

Long-term blood pressure variability, incidence of hypertension and changes in renal function in type 2 diabetes

2020 ◽  
Vol 38 (11) ◽  
pp. 2279-2286
Author(s):  
Francesca Viazzi ◽  
Elisa Russo ◽  
Antonio Mirijello ◽  
Paola Fioretto ◽  
Carlo Giorda ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Type 2 Diabetes ◽  
Renal Function ◽  
Blood Pressure Variability

scholarly journals Roles of Cardiometabolic Factors in Mediating the Causal Effect of Type 2 Diabetes on Cardiovascular Diseases: A Two-Step, Two-Sample Multivariable Mendelian Randomization Study

2021 ◽  
Author(s):  
Ken Chen ◽  
Zhenhuang Zhuang ◽  
Chunli Shao ◽  
Jilin Zheng ◽  
Qing Zhou ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Type 2 Diabetes ◽  
Blood Lipids ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Density Lipoprotein ◽  
Mediation Effect ◽  
Mediation Effects ◽  
Cardiometabolic Factors

Abstract ObjectivesTo investigate the roles of cardiometabolic factors (including blood pressure, blood lipids, thyroid function, body mass, and insulin sensitivity) in mediating the causal effect of type 2 diabetes (T2DM) on cardiovascular disease (CVD) outcomes. DesignTwo-step, two-sample multivariable Mendelian randomization (MVMR) study.SettingInternational genome-wide association study (GWAS) consortia data.ExposureT2DM, blood pressure: systolic blood pressure (SBP), diastolic blood pressure (DBP); blood lipids: low-density lipoprotein (LDL), high-density lipoprotein (HDL), total cholesterol (TC), triglycerides (TG); thyroid function: hyperthyroidism, hypothyroidism; body mass index (BMI), waist-hip-ratio (WHR), and insulin sensitivity. Main outcomesCVD including coronary heart disease (CHD), myocardial infarction (MI) and stroke.MethodsSummary-level data for exposures and main outcomes were extracted from GWAS consortia. We used two-sample MR to illustrate the causal effect of T2DM on CVD subtypes and regression-based MVMR to quantify the possible mediation effects of cardiometabolic factors on CVD.ResultsEach additional unit of log odds of T2DM increased 16% risk of CHD [OR: 1.16, 95% confidence interval (CI): 1.12-1.21], 15% risk of MI (OR: 1.15, 95%CI: 1.10-1.20), and 10% risk of stroke (OR: 1.10, 95%CI: 1.06-1.13). In mediation analysis, SBP, DBP and TG were found as main mediators, while the mediation effects of other cardiometabolic factors were not significant. The proportion of total effect of T2DM on CHD mediated by SBP, DBP and TG was 16% (95%CI: 8%-24%), 7% (95%CI: 1%-13%) and 10% (95%CI: 2%-18%), respectively. Mediation effect of SBP and DBP on MI and stroke, TG on MI was also prominent, while mediation effect of TG on stroke was not significant. Combined mediation effect of all three mediators accounted for 29%, 26% and 13% of total effect of T2DM on CHD, MI and stroke, respectively.ConclusionSBP, DBP and TG mediate a substantial proportion of the causal effect of T2DM on CVD and thus interventions on these factors might reduce considerable excess risk of CVD among T2DM patients.

Sign in / Sign up

Export Citation Format

Share Document