Ploypun Narindrarangkura, Siroj Dejhansathit, Uzma Khan, Margaret Day, Suzanne A Boren, Eduardo J Simoes, Min Soon Kim
Background: Older adults with diabetes frequently access their electronic health record (EHR) notes but often report difficulty understanding medical jargon and nonspecific self-care instructions. To address this communication gap, we developed Support-Engage-Empower-Diabetes (SEE-Diabetes), a patient-centered, EHR-integrated diabetes self-management support tool designed to embed tailored educational statements within the assessment and plan section of clinical notes.
Objective: This study aimed to validate the clarity, relevance, and alignment of SEE-Diabetes content with the Association of Diabetes Care & Education Specialists 7 Self-Care Behaviors framework from the perspectives of older adults and clinicians.
Methods: An interdisciplinary team conducted expert reviews and qualitative interviews with 11 older adults with diabetes and 8 clinicians practicing in primary care (family medicine) and specialty diabetes care settings at a Midwestern academic health center. Patients evaluated the readability and relevance of the content, while clinicians assessed clarity, sufficiency, and potential clinical utility. Interview data were analyzed using inductive thematic analysis, and descriptive statistics were used to summarize participant characteristics.
Results: Patients (mean age 72, SD 4.9 y; mean diabetes duration 26, SD 15 y) reported that the SEE-Diabetes statements were clear, relevant, and written in plain language that supported understanding of self-care recommendations. Clinicians (mean 13, SD 9.5 y of diabetes care experience) viewed the content as concise, clinically appropriate, and well aligned with patient self-management goals and the Association of Diabetes Care & Education Specialists 7 Self-Care Behaviors framework. Both groups identified the tool's potential to enhance patient engagement and patient-clinician communication, while noting opportunities to improve the specificity of language, particularly within medication-related content.
Conclusions: SEE-Diabetes demonstrated content validity as a practical, patient-centered digital health tool for supporting diabetes self-management communication within EHR clinical notes. The findings support its use as a complementary approach to reinforce self-care communication in routine clinical practice and highlight areas for refinement to enhance personalization.
{"title":"Content Validation of an Electronic Health Record-Based Diabetes Self-Management Support Tool for Older Adults With Type 2 Diabetes: Qualitative Study.","authors":"Ploypun Narindrarangkura, Siroj Dejhansathit, Uzma Khan, Margaret Day, Suzanne A Boren, Eduardo J Simoes, Min Soon Kim","doi":"10.2196/83448","DOIUrl":"10.2196/83448","url":null,"abstract":"<p><strong>Background: </strong>Older adults with diabetes frequently access their electronic health record (EHR) notes but often report difficulty understanding medical jargon and nonspecific self-care instructions. To address this communication gap, we developed Support-Engage-Empower-Diabetes (SEE-Diabetes), a patient-centered, EHR-integrated diabetes self-management support tool designed to embed tailored educational statements within the assessment and plan section of clinical notes.</p><p><strong>Objective: </strong>This study aimed to validate the clarity, relevance, and alignment of SEE-Diabetes content with the Association of Diabetes Care & Education Specialists 7 Self-Care Behaviors framework from the perspectives of older adults and clinicians.</p><p><strong>Methods: </strong>An interdisciplinary team conducted expert reviews and qualitative interviews with 11 older adults with diabetes and 8 clinicians practicing in primary care (family medicine) and specialty diabetes care settings at a Midwestern academic health center. Patients evaluated the readability and relevance of the content, while clinicians assessed clarity, sufficiency, and potential clinical utility. Interview data were analyzed using inductive thematic analysis, and descriptive statistics were used to summarize participant characteristics.</p><p><strong>Results: </strong>Patients (mean age 72, SD 4.9 y; mean diabetes duration 26, SD 15 y) reported that the SEE-Diabetes statements were clear, relevant, and written in plain language that supported understanding of self-care recommendations. Clinicians (mean 13, SD 9.5 y of diabetes care experience) viewed the content as concise, clinically appropriate, and well aligned with patient self-management goals and the Association of Diabetes Care & Education Specialists 7 Self-Care Behaviors framework. Both groups identified the tool's potential to enhance patient engagement and patient-clinician communication, while noting opportunities to improve the specificity of language, particularly within medication-related content.</p><p><strong>Conclusions: </strong>SEE-Diabetes demonstrated content validity as a practical, patient-centered digital health tool for supporting diabetes self-management communication within EHR clinical notes. The findings support its use as a complementary approach to reinforce self-care communication in routine clinical practice and highlight areas for refinement to enhance personalization.</p>","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e83448"},"PeriodicalIF":2.6,"publicationDate":"2026-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12880848/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146133425","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shilpa Garg, Robert Kitchen, Ramneek Gupta, Emanuele Trucco, Ewan Pearson
<p><strong>Background: </strong>Sulfonylureas are commonly prescribed for managing type 2 diabetes, yet treatment responses vary significantly among individuals. Although advances in machine learning (ML) may enhance predictive capabilities compared to traditional statistical methods, their practical utility in real-world clinical environments remains uncertain.</p><p><strong>Objective: </strong>This study aimed to evaluate and compare the predictive performance of linear regression models with several ML approaches for predicting glycemic response to sulfonylurea therapy using routine clinical data, and to assess model interpretability using Shapley Additive Explanations (SHAP) analysis as a secondary analysis.</p><p><strong>Methods: </strong>A cohort of 7557 individuals with type 2 diabetes who initiated sulfonylurea therapy was analyzed, with all patients followed for 1 year. Linear and logistic regression models were used as baseline comparisons. A range of ML models was trained to predict the continuous change in hemoglobin A1c (HbA1c) levels and the achievement of HbA1c <58 mmol/mol at follow-up. These models included random forest, extreme gradient boosting, support vector machines, a conventional feedforward neural network, and Bayesian additive regression trees. Model performance was assessed using standard metrics including R² and root mean squared error for regression tasks and area under the receiver operating characteristic for classification. In a subset of 2361 patients, nonfasting connecting peptide (C-peptide) was analyzed as a proxy for β-cell function. SHAP analysis was performed to identify and compare key predictors driving model performance across methods.</p><p><strong>Results: </strong>All models exhibited similar performance, with no significant advantages of ML techniques over linear regression. For continuous outcomes, Bayesian additive regression trees demonstrated the highest R² (0.445) and lowest root mean squared error (0.105), though the differences among models were minimal. For the binary outcome, extreme gradient boosting achieved the highest area under the receiver operating characteristic curve (0.712), with CIs overlapping those of other models. Across all models, baseline HbA1c was consistently the primary predictor, explaining the majority of the variance. SHAP analyses confirmed that baseline HbA1c, age, BMI, and sex were the most influential predictors. Sensitivity analyses and hyperparameter tuning did not significantly improve model performance. In the C-peptide subset, higher C-peptide levels were associated with greater glycemic improvement (β=-3.2 mmol/mol per log(C-peptide); P<.001).</p><p><strong>Conclusions: </strong>In this large, population-based cohort, ML models did not outperform traditional regression for predicting glycemic response to sulfonylureas. These findings suggest that limited gains from ML likely reflect an absence of strong nonlinear or high-order interactions in routine clinical
{"title":"Predictors of Glycemic Response to Sulfonylurea Therapy in Type 2 Diabetes Over 12 Months: Comparative Analysis of Linear Regression and Machine Learning Models.","authors":"Shilpa Garg, Robert Kitchen, Ramneek Gupta, Emanuele Trucco, Ewan Pearson","doi":"10.2196/82635","DOIUrl":"10.2196/82635","url":null,"abstract":"<p><strong>Background: </strong>Sulfonylureas are commonly prescribed for managing type 2 diabetes, yet treatment responses vary significantly among individuals. Although advances in machine learning (ML) may enhance predictive capabilities compared to traditional statistical methods, their practical utility in real-world clinical environments remains uncertain.</p><p><strong>Objective: </strong>This study aimed to evaluate and compare the predictive performance of linear regression models with several ML approaches for predicting glycemic response to sulfonylurea therapy using routine clinical data, and to assess model interpretability using Shapley Additive Explanations (SHAP) analysis as a secondary analysis.</p><p><strong>Methods: </strong>A cohort of 7557 individuals with type 2 diabetes who initiated sulfonylurea therapy was analyzed, with all patients followed for 1 year. Linear and logistic regression models were used as baseline comparisons. A range of ML models was trained to predict the continuous change in hemoglobin A1c (HbA1c) levels and the achievement of HbA1c <58 mmol/mol at follow-up. These models included random forest, extreme gradient boosting, support vector machines, a conventional feedforward neural network, and Bayesian additive regression trees. Model performance was assessed using standard metrics including R² and root mean squared error for regression tasks and area under the receiver operating characteristic for classification. In a subset of 2361 patients, nonfasting connecting peptide (C-peptide) was analyzed as a proxy for β-cell function. SHAP analysis was performed to identify and compare key predictors driving model performance across methods.</p><p><strong>Results: </strong>All models exhibited similar performance, with no significant advantages of ML techniques over linear regression. For continuous outcomes, Bayesian additive regression trees demonstrated the highest R² (0.445) and lowest root mean squared error (0.105), though the differences among models were minimal. For the binary outcome, extreme gradient boosting achieved the highest area under the receiver operating characteristic curve (0.712), with CIs overlapping those of other models. Across all models, baseline HbA1c was consistently the primary predictor, explaining the majority of the variance. SHAP analyses confirmed that baseline HbA1c, age, BMI, and sex were the most influential predictors. Sensitivity analyses and hyperparameter tuning did not significantly improve model performance. In the C-peptide subset, higher C-peptide levels were associated with greater glycemic improvement (β=-3.2 mmol/mol per log(C-peptide); P<.001).</p><p><strong>Conclusions: </strong>In this large, population-based cohort, ML models did not outperform traditional regression for predicting glycemic response to sulfonylureas. These findings suggest that limited gains from ML likely reflect an absence of strong nonlinear or high-order interactions in routine clinical","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e82635"},"PeriodicalIF":2.6,"publicationDate":"2026-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12880802/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146133414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Background: In the past decade, telehealth has transformed health care delivery by allowing patients more rapid and convenient access to necessary care without the cost and logistical challenges of traveling to a health care facility. Telehealth services can benefit patients with type 2 diabetes mellitus (T2DM) amid a growing epidemic of T2DM in the United States that affects people of all ages and races. In 2020, 33 million people were diagnosed with this chronic disease, with the number expected to rise by 50% by 2040. Telehealth facilitates regular contact between patients and their providers, especially when there are geographic barriers and time constraints prohibiting physical interaction, at little or no added cost to the patient and at their convenience.
Objective: This study examines cultural and technological barriers affecting telehealth adoption among Asian American people with T2DM.
Methods: A qualitative case study approach was employed, utilizing semistructured interviews with 30 Asian American individuals in Missouri. Thematic analysis was used to identify key barriers.
Results: Four major barriers emerged: (1) language and cultural barriers-limited availability of translated materials and interpreters; (2) limited digital literacy and access-older adults and individuals with low technological exposure struggled with telehealth platforms; (3) limited provider recommendations-health care providers did not actively endorse telehealth, reducing patient awareness of telehealth as an option; and (4) technology access and infrastructure disparities-low-income participants faced challenges with the costs of and access to broadband and telehealth-compatible devices.
Conclusions: Addressing cultural and technological barriers is crucial to increasing telehealth adoption among Asian American people with T2DM. Culturally tailored interventions, provider engagement, and digital literacy programs should be prioritized. Policy efforts must focus on expanding broadband access and providing multilingual telehealth resources.
{"title":"Cultural and Technological Barriers to Telehealth Adoption for Type 2 Diabetes Management Among Asian American Patients: Qualitative Case Study.","authors":"Devi Gurung States","doi":"10.2196/75689","DOIUrl":"10.2196/75689","url":null,"abstract":"<p><strong>Background: </strong>In the past decade, telehealth has transformed health care delivery by allowing patients more rapid and convenient access to necessary care without the cost and logistical challenges of traveling to a health care facility. Telehealth services can benefit patients with type 2 diabetes mellitus (T2DM) amid a growing epidemic of T2DM in the United States that affects people of all ages and races. In 2020, 33 million people were diagnosed with this chronic disease, with the number expected to rise by 50% by 2040. Telehealth facilitates regular contact between patients and their providers, especially when there are geographic barriers and time constraints prohibiting physical interaction, at little or no added cost to the patient and at their convenience.</p><p><strong>Objective: </strong>This study examines cultural and technological barriers affecting telehealth adoption among Asian American people with T2DM.</p><p><strong>Methods: </strong>A qualitative case study approach was employed, utilizing semistructured interviews with 30 Asian American individuals in Missouri. Thematic analysis was used to identify key barriers.</p><p><strong>Results: </strong>Four major barriers emerged: (1) language and cultural barriers-limited availability of translated materials and interpreters; (2) limited digital literacy and access-older adults and individuals with low technological exposure struggled with telehealth platforms; (3) limited provider recommendations-health care providers did not actively endorse telehealth, reducing patient awareness of telehealth as an option; and (4) technology access and infrastructure disparities-low-income participants faced challenges with the costs of and access to broadband and telehealth-compatible devices.</p><p><strong>Conclusions: </strong>Addressing cultural and technological barriers is crucial to increasing telehealth adoption among Asian American people with T2DM. Culturally tailored interventions, provider engagement, and digital literacy programs should be prioritized. Policy efforts must focus on expanding broadband access and providing multilingual telehealth resources.</p>","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e75689"},"PeriodicalIF":2.6,"publicationDate":"2026-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12867477/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146114632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mercia Napame, Sylvie Picard, Tony Foglia, Anne Rubenstrunk, Florence Baudoux, Francoise Giroud, Sandrine Lablanche, Sophie Borot
<p><strong>Background: </strong>Closed-loop insulin delivery is the new standard of care for patients with type 1 diabetes (T1D). However, in France, its implementation remains predominantly hospital based. Expanding access to this treatment through alternative care models looks essential.</p><p><strong>Objective: </strong>This study (cost-effectiveness analysis) compares 2 care models for people with T1D implementing a closed-loop system in France: outpatient care in the Inter-Regional Center for Automated Insulin in Diabetes (CIRDIA) and inpatient care.</p><p><strong>Methods: </strong>We conducted a cost-effectiveness analysis using retrospective observational data from individuals with T1D aged 16 years and older from the implementation of the closed loop to a 12-month follow-up either in the CIRDIA (CIRDIA group) or in a hospital center setting (hospital center [HC] group). The cost analyses were based on patient records and public databases: the French Medical Information Systems Program and the French General Nomenclature of Professional Acts. Closed-loop efficacy was assessed using a time in range (TIR) of 70 to 180 mg/dL, and closed-loop safety was assessed using the glycemia risk index (GRI), a single indicator that represents the risk of hypoglycemia or hyperglycemia and ranges from 0 (minimal risk) to 100 (maximal risk).</p><p><strong>Results: </strong>A total of 201 patients were included: 128 in the CIRDIA group and 73 in the HC group. The mean (SD) age was 43 (14) years and 46 (15) years, respectively. Mean (SD) baseline TIR was 52.9% (16%) in the CIRDIA group versus 65.9% (15.1%) in the HC group (P<.001), whereas mean (SD) baseline GRI was 56.4 (21) in the CIRDIA group versus 37.8 (19.8) in the HC group (P<.001). After 12 months, both groups achieved similar efficacy and safety outcomes with a mean (SD) TIR at 72.7% (11.6%) in the CIRDIA group versus 71.9% (10.5%) in the HC group, and a mean GRI at 30.1 (14.1) versus 30.3 (13), respectively. There were no significant between-group differences (P=.60 for TIR; P=.91 for GRI). However, the CIRDIA was associated with significantly lower management costs with a mean cost of €8373.12 (SD €427.30; €1=US $1.10 at the time of the study) per patient in the CIRDIA group versus €8814.32 (SD €192) per patient in the HC group (P<.001). The estimated saving was €626 per percentage point of increase in TIR and €2011 per point of reduction in GRI, indicating that the HC closed-loop initiation was dominated by the CIRDIA. The CIRDIA was less costly than HC in 8600 (86%) out of 10,000 simulations in a probabilistic sensitivity analysis.</p><p><strong>Conclusions: </strong>These findings suggest the potential of the CIRDIA to represent a viable alternative organizational model for closed-loop initiation in France, achieving comparable effectiveness at lower cost in our population. Further research with longer follow-up is warranted. From a policy perspective, the resources saved could be at least part
{"title":"Inter-Regional Center for Automated Insulin in Diabetes (CIRDIA) and Hospital-Based Approaches to Closed-Loop Therapy in Type 1 Diabetes: Cost-Effectiveness Analysis.","authors":"Mercia Napame, Sylvie Picard, Tony Foglia, Anne Rubenstrunk, Florence Baudoux, Francoise Giroud, Sandrine Lablanche, Sophie Borot","doi":"10.2196/86690","DOIUrl":"10.2196/86690","url":null,"abstract":"<p><strong>Background: </strong>Closed-loop insulin delivery is the new standard of care for patients with type 1 diabetes (T1D). However, in France, its implementation remains predominantly hospital based. Expanding access to this treatment through alternative care models looks essential.</p><p><strong>Objective: </strong>This study (cost-effectiveness analysis) compares 2 care models for people with T1D implementing a closed-loop system in France: outpatient care in the Inter-Regional Center for Automated Insulin in Diabetes (CIRDIA) and inpatient care.</p><p><strong>Methods: </strong>We conducted a cost-effectiveness analysis using retrospective observational data from individuals with T1D aged 16 years and older from the implementation of the closed loop to a 12-month follow-up either in the CIRDIA (CIRDIA group) or in a hospital center setting (hospital center [HC] group). The cost analyses were based on patient records and public databases: the French Medical Information Systems Program and the French General Nomenclature of Professional Acts. Closed-loop efficacy was assessed using a time in range (TIR) of 70 to 180 mg/dL, and closed-loop safety was assessed using the glycemia risk index (GRI), a single indicator that represents the risk of hypoglycemia or hyperglycemia and ranges from 0 (minimal risk) to 100 (maximal risk).</p><p><strong>Results: </strong>A total of 201 patients were included: 128 in the CIRDIA group and 73 in the HC group. The mean (SD) age was 43 (14) years and 46 (15) years, respectively. Mean (SD) baseline TIR was 52.9% (16%) in the CIRDIA group versus 65.9% (15.1%) in the HC group (P<.001), whereas mean (SD) baseline GRI was 56.4 (21) in the CIRDIA group versus 37.8 (19.8) in the HC group (P<.001). After 12 months, both groups achieved similar efficacy and safety outcomes with a mean (SD) TIR at 72.7% (11.6%) in the CIRDIA group versus 71.9% (10.5%) in the HC group, and a mean GRI at 30.1 (14.1) versus 30.3 (13), respectively. There were no significant between-group differences (P=.60 for TIR; P=.91 for GRI). However, the CIRDIA was associated with significantly lower management costs with a mean cost of €8373.12 (SD €427.30; €1=US $1.10 at the time of the study) per patient in the CIRDIA group versus €8814.32 (SD €192) per patient in the HC group (P<.001). The estimated saving was €626 per percentage point of increase in TIR and €2011 per point of reduction in GRI, indicating that the HC closed-loop initiation was dominated by the CIRDIA. The CIRDIA was less costly than HC in 8600 (86%) out of 10,000 simulations in a probabilistic sensitivity analysis.</p><p><strong>Conclusions: </strong>These findings suggest the potential of the CIRDIA to represent a viable alternative organizational model for closed-loop initiation in France, achieving comparable effectiveness at lower cost in our population. Further research with longer follow-up is warranted. From a policy perspective, the resources saved could be at least part","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e86690"},"PeriodicalIF":2.6,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12854399/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146087953","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
<p><strong>Background: </strong>Diabetes prediction requires accurate, privacy-preserving, and scalable solutions. Traditional machine learning models rely on centralized data, posing risks to data privacy and regulatory compliance. Moreover, health care settings are highly heterogeneous, with diverse participants, hospitals, clinics, and wearables, producing nonindependent and identically distributed data and operating under varied computational constraints. Learning in isolation at individual institutions limits model generalizability and effectiveness. Collaborative federated learning (FL) enables institutions to jointly train models without sharing raw data, but current approaches often struggle with heterogeneity, security threats, and system coordination.</p><p><strong>Objective: </strong>This study aims to develop a secure, scalable, and privacy-preserving framework for diabetes prediction by integrating FL with ensemble modeling, blockchain-based access control, and knowledge distillation. The framework is designed to handle data heterogeneity, nonindependent and identically distributed distributions, and varying computational capacities across diverse health care participants while simultaneously enhancing data privacy, security, and trust.</p><p><strong>Methods: </strong>We propose a federated ensemble learning framework, FedEnTrust, that enables decentralized health care participants to collaboratively train models without sharing raw data. Each participant shares soft label outputs, which are distilled and aggregated through adaptive weighted voting to form a global consensus. The framework supports heterogeneous participants by assigning model architectures based on local computational capacity. To ensure secure and transparent coordination, a blockchain-enabled smart contract governs participant registration, role assignment, and model submission with strict role-based access control. We evaluated the system on the PIMA Indians Diabetes Dataset, measuring prediction accuracy, communication efficiency, and blockchain performance.</p><p><strong>Results: </strong>The FedEnTrust framework achieved 84.2% accuracy, with precision, recall, and F1-score of 84.6%, 88.6%, and 86.4%, respectively, outperforming existing decentralized models and nearing centralized deep learning benchmarks. The blockchain-based smart contract ensured 100% success for authorized transactions and rejected all unauthorized attempts, including malicious submissions. The average blockchain latency was 210 milliseconds, with a gas cost of ~107,940 units, enabling secure, real-time interaction. Throughout, patient privacy was preserved by exchanging only model metadata, not raw data.</p><p><strong>Conclusions: </strong>FedEnTrust offers a deployable, privacy-preserving solution for decentralized health care prediction by integrating FL, ensemble modeling, blockchain-based access control, and knowledge distillation. It balances accuracy, scalability, and ethical data u
{"title":"Privacy-Preserving Collaborative Diabetes Prediction in Heterogeneous Health Care Systems: Algorithm Development and Validation of a Secure Federated Ensemble Framework.","authors":"Md Rakibul Hasan, Juan Li","doi":"10.2196/79166","DOIUrl":"10.2196/79166","url":null,"abstract":"<p><strong>Background: </strong>Diabetes prediction requires accurate, privacy-preserving, and scalable solutions. Traditional machine learning models rely on centralized data, posing risks to data privacy and regulatory compliance. Moreover, health care settings are highly heterogeneous, with diverse participants, hospitals, clinics, and wearables, producing nonindependent and identically distributed data and operating under varied computational constraints. Learning in isolation at individual institutions limits model generalizability and effectiveness. Collaborative federated learning (FL) enables institutions to jointly train models without sharing raw data, but current approaches often struggle with heterogeneity, security threats, and system coordination.</p><p><strong>Objective: </strong>This study aims to develop a secure, scalable, and privacy-preserving framework for diabetes prediction by integrating FL with ensemble modeling, blockchain-based access control, and knowledge distillation. The framework is designed to handle data heterogeneity, nonindependent and identically distributed distributions, and varying computational capacities across diverse health care participants while simultaneously enhancing data privacy, security, and trust.</p><p><strong>Methods: </strong>We propose a federated ensemble learning framework, FedEnTrust, that enables decentralized health care participants to collaboratively train models without sharing raw data. Each participant shares soft label outputs, which are distilled and aggregated through adaptive weighted voting to form a global consensus. The framework supports heterogeneous participants by assigning model architectures based on local computational capacity. To ensure secure and transparent coordination, a blockchain-enabled smart contract governs participant registration, role assignment, and model submission with strict role-based access control. We evaluated the system on the PIMA Indians Diabetes Dataset, measuring prediction accuracy, communication efficiency, and blockchain performance.</p><p><strong>Results: </strong>The FedEnTrust framework achieved 84.2% accuracy, with precision, recall, and F1-score of 84.6%, 88.6%, and 86.4%, respectively, outperforming existing decentralized models and nearing centralized deep learning benchmarks. The blockchain-based smart contract ensured 100% success for authorized transactions and rejected all unauthorized attempts, including malicious submissions. The average blockchain latency was 210 milliseconds, with a gas cost of ~107,940 units, enabling secure, real-time interaction. Throughout, patient privacy was preserved by exchanging only model metadata, not raw data.</p><p><strong>Conclusions: </strong>FedEnTrust offers a deployable, privacy-preserving solution for decentralized health care prediction by integrating FL, ensemble modeling, blockchain-based access control, and knowledge distillation. It balances accuracy, scalability, and ethical data u","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e79166"},"PeriodicalIF":2.6,"publicationDate":"2026-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12834198/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146054479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Debbie Lam, Poonamdeep Jhajj, Diana Sherifali, Frances S Chen, Tricia S Tang
Background: Existing qualitative research in peer support interventions has largely focused on the recipients of support rather than those delivering support. Exploring the perspectives of both roles may provide a holistic understanding of the peer support experience.
Objective: This study elicits the experiences of recipients and providers of support who participated in REACHOUT, a 6-month peer-led mental health support intervention delivered via mobile app for adults with type 1 diabetes. REACHOUT offered multiple support delivery modalities (one-on-one, group-based texting, and virtual face-to-face small group sessions) that could be customized by recipients.
Methods: A total of 32 study participants (recipients and peer supporters) attended focus group discussions following the completion of REACHOUT. Thematic analysis was performed in an inductive approach.
Results: Four major themes were identified by thematic analysis: (1) need for a sense of community and belonging, (2) factors to enhance the recipient-peer supporter experience, (3) key aspects of the peer supporter experience, and (4) importance of personalizing the user experience while using the REACHOUT mobile app. REACHOUT successfully fostered connectedness by bringing together adults with type 1 diabetes who previously felt isolated. Recipients felt greater agency when given the opportunity to self-select a peer supporter. The main factors considered during the matching process included insulin delivery and glucose monitoring systems, duration of diabetes, shared hobbies, life stage, and age. While support was designed to be unidirectional from peer supporter to recipient, the former also derived benefits. Peer supporters expressed the need for greater guidance around navigating boundaries and responding to emotionally charged conversations. Finally, the REACHOUT app was able to accommodate a heterogeneity of support needs by offering one-on-one and group support across multiple communication platforms including text, audio, and video.
Conclusions: The success of peer-led mental health support interventions such as REACHOUT is likely associated with the recipient-peer supporter dynamic. By offering a range of support delivery and communication modalities, participants can better personalize solutions to meet their unique support needs. Understanding the perspectives of both recipients and peer supporters is essential to refining interventions and optimizing digitally delivered mental health support models.
{"title":"Exploring the REACHOUT Mental Health Support App for Type 1 Diabetes From the Perspectives of Recipients and Providers of Peer Support: Qualitative Study.","authors":"Debbie Lam, Poonamdeep Jhajj, Diana Sherifali, Frances S Chen, Tricia S Tang","doi":"10.2196/72779","DOIUrl":"10.2196/72779","url":null,"abstract":"<p><strong>Background: </strong>Existing qualitative research in peer support interventions has largely focused on the recipients of support rather than those delivering support. Exploring the perspectives of both roles may provide a holistic understanding of the peer support experience.</p><p><strong>Objective: </strong>This study elicits the experiences of recipients and providers of support who participated in REACHOUT, a 6-month peer-led mental health support intervention delivered via mobile app for adults with type 1 diabetes. REACHOUT offered multiple support delivery modalities (one-on-one, group-based texting, and virtual face-to-face small group sessions) that could be customized by recipients.</p><p><strong>Methods: </strong>A total of 32 study participants (recipients and peer supporters) attended focus group discussions following the completion of REACHOUT. Thematic analysis was performed in an inductive approach.</p><p><strong>Results: </strong>Four major themes were identified by thematic analysis: (1) need for a sense of community and belonging, (2) factors to enhance the recipient-peer supporter experience, (3) key aspects of the peer supporter experience, and (4) importance of personalizing the user experience while using the REACHOUT mobile app. REACHOUT successfully fostered connectedness by bringing together adults with type 1 diabetes who previously felt isolated. Recipients felt greater agency when given the opportunity to self-select a peer supporter. The main factors considered during the matching process included insulin delivery and glucose monitoring systems, duration of diabetes, shared hobbies, life stage, and age. While support was designed to be unidirectional from peer supporter to recipient, the former also derived benefits. Peer supporters expressed the need for greater guidance around navigating boundaries and responding to emotionally charged conversations. Finally, the REACHOUT app was able to accommodate a heterogeneity of support needs by offering one-on-one and group support across multiple communication platforms including text, audio, and video.</p><p><strong>Conclusions: </strong>The success of peer-led mental health support interventions such as REACHOUT is likely associated with the recipient-peer supporter dynamic. By offering a range of support delivery and communication modalities, participants can better personalize solutions to meet their unique support needs. Understanding the perspectives of both recipients and peer supporters is essential to refining interventions and optimizing digitally delivered mental health support models.</p>","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e72779"},"PeriodicalIF":2.6,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12822868/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146020653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Simon K Kjær, Lukas Ochsner Reynaud Ridder, Mads Svart, Nikolaj Rittig, Lise Nørkjær Bjerg, Birgitte Sandfeld-Paulsen, Henrik Holm Thomsen
Unlabelled: In our study, a commercially available continuous ketone monitoring device captured β-Hydroxybutyrate (BHB) dynamics during exogenous ketosis but revealed a gradual decline day-to-day BHB concentrations over 14 days in both ketone ester and placebo groups, likely reflecting sensor drift.
{"title":"Continuous Ketone Monitoring: Data From a Randomized Controlled Trial.","authors":"Simon K Kjær, Lukas Ochsner Reynaud Ridder, Mads Svart, Nikolaj Rittig, Lise Nørkjær Bjerg, Birgitte Sandfeld-Paulsen, Henrik Holm Thomsen","doi":"10.2196/85548","DOIUrl":"10.2196/85548","url":null,"abstract":"<p><strong>Unlabelled: </strong>In our study, a commercially available continuous ketone monitoring device captured β-Hydroxybutyrate (BHB) dynamics during exogenous ketosis but revealed a gradual decline day-to-day BHB concentrations over 14 days in both ketone ester and placebo groups, likely reflecting sensor drift.</p>","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e85548"},"PeriodicalIF":2.6,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12799078/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145967800","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Background: Diabetic kidney disease (DKD) is a major complication of diabetes and the leading cause of end-stage renal disease globally. Artificial intelligence (AI) technologies have shown increasing potential in DKD research for early detection, risk prediction, and disease management. However, the landscape of AI applications in this field remains incompletely mapped, especially in terms of collaboration networks, thematic evolution, and clinical translation.
Objective: This study aims to perform a comprehensive bibliometric and translational analysis of AI-related DKD research published between 2006 and 2024, identifying publication trends, research hotspots, key contributors, collaboration patterns, and the extent of clinical validation and explainability.
Methods: A systematic search of the Web of Science Core Collection was conducted to identify English-language original articles applying AI technologies to DKD. Articles were screened following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines. Bibliometric visualization was performed using CiteSpace and VOSviewer to assess coauthorship, institutional and country collaboration, keyword evolution, and citation bursts. A qualitative review was conducted to evaluate clinical validation, model explainability, and real-world implementation.
Results: Out of 1158 retrieved records, 384 studies met the inclusion criteria. Global publications on AI in DKD increased rapidly after 2019. China led in publication volume, followed by the United States, India, and Iran. Keyword analysis showed a thematic transition from early biomarker and proteomic research to deep learning, clinical prediction models, and management tools. Despite methodological advances, few studies included external validation or explainability frameworks. Notable translational efforts included DeepMind's acute kidney injury predictor and a chronic kidney disease prediction model developed by Sumit, yet widespread real-world integration remains limited.
Conclusions: AI research in DKD has grown substantially over the past 2 decades, with expanding international collaboration and diversification of research themes. However, challenges persist in clinical applicability, model transparency, and global inclusivity. Future research should prioritize explainable AI, multicenter validation, and integration into clinical workflows to support effective translation of AI innovations into DKD care.
背景：糖尿病肾病（DKD）是糖尿病的主要并发症，也是全球终末期肾脏疾病的主要原因。人工智能（AI）技术在DKD研究的早期发现、风险预测和疾病管理方面显示出越来越大的潜力。然而，人工智能在这一领域的应用前景仍然不完整，特别是在协作网络、主题演变和临床翻译方面。目的：对2006年至2024年间发表的人工智能相关DKD研究进行综合文献计量学和翻译分析，确定出版趋势、研究热点、主要贡献者、合作模式以及临床验证和可解释性程度。方法：对Web of Science核心馆藏进行系统搜索，以识别将AI技术应用于DKD的英语原创文章。文章按照PRISMA（系统评价和荟萃分析首选报告项目）2020指南进行筛选。使用CiteSpace和VOSviewer进行文献计量可视化，以评估合著者、机构和国家合作、关键词演变和引文爆发。进行了定性回顾，以评估临床有效性，模型的可解释性和现实世界的实施。结果：在1158份检索记录中，384项研究符合纳入标准。2019年之后，全球关于DKD领域人工智能的出版物迅速增加。中国的出版物数量最多，其次是美国、印度和伊朗。关键词分析显示了从早期生物标志物和蛋白质组学研究到深度学习、临床预测模型和管理工具的主题转变。尽管方法学有所进步，但很少有研究包括外部验证或可解释性框架。值得注意的转化工作包括DeepMind的急性肾损伤预测器和Sumit开发的慢性肾病预测模型，但广泛的现实应用仍然有限。结论：在过去的20年里，随着国际合作的扩大和研究主题的多样化，DKD领域的人工智能研究取得了长足的发展。然而，在临床适用性、模型透明度和全球包容性方面仍然存在挑战。未来的研究应优先考虑可解释的人工智能，多中心验证，并整合到临床工作流程中，以支持将人工智能创新有效地转化为DKD护理。
{"title":"Artificial Intelligence in Diabetic Kidney Disease Research: Bibliometric Analysis From 2006 to 2024.","authors":"Xingyuan Li, Liming Xiao, Fenghao Yang, Fang Liu","doi":"10.2196/72616","DOIUrl":"10.2196/72616","url":null,"abstract":"<p><strong>Background: </strong>Diabetic kidney disease (DKD) is a major complication of diabetes and the leading cause of end-stage renal disease globally. Artificial intelligence (AI) technologies have shown increasing potential in DKD research for early detection, risk prediction, and disease management. However, the landscape of AI applications in this field remains incompletely mapped, especially in terms of collaboration networks, thematic evolution, and clinical translation.</p><p><strong>Objective: </strong>This study aims to perform a comprehensive bibliometric and translational analysis of AI-related DKD research published between 2006 and 2024, identifying publication trends, research hotspots, key contributors, collaboration patterns, and the extent of clinical validation and explainability.</p><p><strong>Methods: </strong>A systematic search of the Web of Science Core Collection was conducted to identify English-language original articles applying AI technologies to DKD. Articles were screened following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines. Bibliometric visualization was performed using CiteSpace and VOSviewer to assess coauthorship, institutional and country collaboration, keyword evolution, and citation bursts. A qualitative review was conducted to evaluate clinical validation, model explainability, and real-world implementation.</p><p><strong>Results: </strong>Out of 1158 retrieved records, 384 studies met the inclusion criteria. Global publications on AI in DKD increased rapidly after 2019. China led in publication volume, followed by the United States, India, and Iran. Keyword analysis showed a thematic transition from early biomarker and proteomic research to deep learning, clinical prediction models, and management tools. Despite methodological advances, few studies included external validation or explainability frameworks. Notable translational efforts included DeepMind's acute kidney injury predictor and a chronic kidney disease prediction model developed by Sumit, yet widespread real-world integration remains limited.</p><p><strong>Conclusions: </strong>AI research in DKD has grown substantially over the past 2 decades, with expanding international collaboration and diversification of research themes. However, challenges persist in clinical applicability, model transparency, and global inclusivity. Future research should prioritize explainable AI, multicenter validation, and integration into clinical workflows to support effective translation of AI innovations into DKD care.</p>","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e72616"},"PeriodicalIF":2.6,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12786635/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145946729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Edmund Evangelista, Fathima Ruba, Salman Bukhari, Amril Nazir, Ravishankar Sharma
Background: Gestational diabetes mellitus (GDM) is a prevalent chronic condition that affects maternal and fetal health outcomes worldwide, increasingly in underserved populations. While generative artificial intelligence (AI) and large language models (LLMs) have shown promise in health care, their application in GDM management remains underexplored.
Objective: This study aimed to investigate whether retrieval-augmented generation techniques, when combined with knowledge graphs (KGs), could improve the contextual relevance and accuracy of AI-driven clinical decision support. For this, we developed and validated a graph-based retrieval-augmented generation (GraphRAG)-enabled local LLM as a clinical support tool for GDM management, assessing its performance against open-source LLM tools.
Methods: A prototype clinical AI assistant was developed using a GraphRAG constructed from 1212 peer-reviewed research articles on GDM interventions, retrieved from the Semantic Scholar API (2000-2024). The GraphRAG prototype integrated entity extraction, KG construction using Neo4j, and retrieval-augmented response generation. The performance was evaluated in a simulated environment using clinical and layperson prompts, comparing the outputs of the systems against ChatGPT (OpenAI), Claude (Anthropic), and BioMistral models across 5 common natural language generation metrics.
Results: The GraphRAG-enabled local LLM showed higher accuracy in generating clinically relevant responses. It achieved a bilingual evaluation understudy score of 0.99, Jaccard similarity of 0.98, and BERTScore of 0.98, outperforming the benchmark LLMs. The prototype also produced accurate, evidence-based recommendations for clinicians and patients, demonstrating its feasibility as a clinical support tool.
Conclusions: GraphRAG-enabled local LLMs show much potential for improving personalized GDM care by integrating domain-specific evidence and contextual retrieval. Our prototype proof-of-concept serves two purposes: (1) the local LLM architecture gives practitioners from underserved locations access to state-of-the-art medical research in the treatment of chronic conditions and (2) the KG schema may be feasibly built on peer-reviewed, indexed publications, devoid of hallucinations and contextualized with patient data. We conclude that advanced AI techniques such as KGs, retrieval-augmented generation, and local LLMs improve GDM management decisions and other similar conditions and advance equitable health care delivery in resource-constrained health care environments.
{"title":"GraphRAG-Enabled Local Large Language Model for Gestational Diabetes Mellitus: Development of a Proof-of-Concept.","authors":"Edmund Evangelista, Fathima Ruba, Salman Bukhari, Amril Nazir, Ravishankar Sharma","doi":"10.2196/76454","DOIUrl":"10.2196/76454","url":null,"abstract":"<p><strong>Background: </strong>Gestational diabetes mellitus (GDM) is a prevalent chronic condition that affects maternal and fetal health outcomes worldwide, increasingly in underserved populations. While generative artificial intelligence (AI) and large language models (LLMs) have shown promise in health care, their application in GDM management remains underexplored.</p><p><strong>Objective: </strong>This study aimed to investigate whether retrieval-augmented generation techniques, when combined with knowledge graphs (KGs), could improve the contextual relevance and accuracy of AI-driven clinical decision support. For this, we developed and validated a graph-based retrieval-augmented generation (GraphRAG)-enabled local LLM as a clinical support tool for GDM management, assessing its performance against open-source LLM tools.</p><p><strong>Methods: </strong>A prototype clinical AI assistant was developed using a GraphRAG constructed from 1212 peer-reviewed research articles on GDM interventions, retrieved from the Semantic Scholar API (2000-2024). The GraphRAG prototype integrated entity extraction, KG construction using Neo4j, and retrieval-augmented response generation. The performance was evaluated in a simulated environment using clinical and layperson prompts, comparing the outputs of the systems against ChatGPT (OpenAI), Claude (Anthropic), and BioMistral models across 5 common natural language generation metrics.</p><p><strong>Results: </strong>The GraphRAG-enabled local LLM showed higher accuracy in generating clinically relevant responses. It achieved a bilingual evaluation understudy score of 0.99, Jaccard similarity of 0.98, and BERTScore of 0.98, outperforming the benchmark LLMs. The prototype also produced accurate, evidence-based recommendations for clinicians and patients, demonstrating its feasibility as a clinical support tool.</p><p><strong>Conclusions: </strong>GraphRAG-enabled local LLMs show much potential for improving personalized GDM care by integrating domain-specific evidence and contextual retrieval. Our prototype proof-of-concept serves two purposes: (1) the local LLM architecture gives practitioners from underserved locations access to state-of-the-art medical research in the treatment of chronic conditions and (2) the KG schema may be feasibly built on peer-reviewed, indexed publications, devoid of hallucinations and contextualized with patient data. We conclude that advanced AI techniques such as KGs, retrieval-augmented generation, and local LLMs improve GDM management decisions and other similar conditions and advance equitable health care delivery in resource-constrained health care environments.</p>","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"11 ","pages":"e76454"},"PeriodicalIF":2.6,"publicationDate":"2026-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12767777/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145907331","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Michelle I Knopp, Ann Marie Castleman, Anna Schwarz, Jamarin Belger-Wallace, Mercedes Falciglia, Aleona Zuzek, Eneida Mendonca
Background: Continuous glucose monitors (CGM) reduce the burden of glycemic monitoring and improve glycemic control, quality of life, and decreased health care use. Despite expanded insurance coverage and adoption, barriers remain, especially in primary care. Existing research largely evaluates specific populations or interventions, leaving limited insight into the broader primary care experience.
Objective: This study aims to examine the experiences of adults with type 2 diabetes mellitus (T2DM) using CGM in primary care, guided by the Health Belief Model and Technology Acceptance Model.
Methods: This qualitative study included in-person semistructured sessions (interviews or a focus group), surveys, and electronic health record data. Participants were recruited from 3 urban primary care (internal medicine and internal medicine-pediatrics) clinics affiliated with a large academic health system in Southwest Ohio, United States, with high rates of public insurance (Medicare or Medicaid). Eligible participants were adults (≥18 y) with T2DM and a CGM prescription. Data were analyzed using theme generation guided by directed content analysis in MAXQDA (VERBI Software GmbH) with codes derived from Health Belief Model and Technology Acceptance Model constructs. Survey data were used to triangulate to enhance validity.
Results: Overall, 16 participants (interviews: n=12; 1 focus group: n=4) were recruited for the study with a mean age of 56.9 (SD 10.5) years. In total, 69% (11/16) identified as Black, 100% (16/16) as Non-Hispanic, and 69% (11/16) as female, and 94% (15/16) used public insurance. Six themes emerged: disease susceptibility, disease severity, influential drivers, perceived ease of use, perceived usefulness, and attitude toward using CGM. All participants found CGM helpful and would recommend it to others. While affirming numerous barriers well-described in other populations, this study uniquely describes the burden of comorbidities, the trust in CGM data compared to glucometer-based monitoring, and the reliance on receivers to use CGM technology in this patient population.
Conclusions: CGM is valued by adults with T2DM in primary care, yet barriers remain. Tailored support for initiation, troubleshooting, and education (especially alarm management and data interpretation) is needed. These insights can inform scalable strategies to enhance CGM use and experience in primary care.
{"title":"Continuous Glucose Monitors Among Adults With Type 2 Diabetes Mellitus in the Primary Care Setting: Qualitative Study Informed by Technology Acceptance Model and Health Belief Model.","authors":"Michelle I Knopp, Ann Marie Castleman, Anna Schwarz, Jamarin Belger-Wallace, Mercedes Falciglia, Aleona Zuzek, Eneida Mendonca","doi":"10.2196/73446","DOIUrl":"10.2196/73446","url":null,"abstract":"<p><strong>Background: </strong>Continuous glucose monitors (CGM) reduce the burden of glycemic monitoring and improve glycemic control, quality of life, and decreased health care use. Despite expanded insurance coverage and adoption, barriers remain, especially in primary care. Existing research largely evaluates specific populations or interventions, leaving limited insight into the broader primary care experience.</p><p><strong>Objective: </strong>This study aims to examine the experiences of adults with type 2 diabetes mellitus (T2DM) using CGM in primary care, guided by the Health Belief Model and Technology Acceptance Model.</p><p><strong>Methods: </strong>This qualitative study included in-person semistructured sessions (interviews or a focus group), surveys, and electronic health record data. Participants were recruited from 3 urban primary care (internal medicine and internal medicine-pediatrics) clinics affiliated with a large academic health system in Southwest Ohio, United States, with high rates of public insurance (Medicare or Medicaid). Eligible participants were adults (≥18 y) with T2DM and a CGM prescription. Data were analyzed using theme generation guided by directed content analysis in MAXQDA (VERBI Software GmbH) with codes derived from Health Belief Model and Technology Acceptance Model constructs. Survey data were used to triangulate to enhance validity.</p><p><strong>Results: </strong>Overall, 16 participants (interviews: n=12; 1 focus group: n=4) were recruited for the study with a mean age of 56.9 (SD 10.5) years. In total, 69% (11/16) identified as Black, 100% (16/16) as Non-Hispanic, and 69% (11/16) as female, and 94% (15/16) used public insurance. Six themes emerged: disease susceptibility, disease severity, influential drivers, perceived ease of use, perceived usefulness, and attitude toward using CGM. All participants found CGM helpful and would recommend it to others. While affirming numerous barriers well-described in other populations, this study uniquely describes the burden of comorbidities, the trust in CGM data compared to glucometer-based monitoring, and the reliance on receivers to use CGM technology in this patient population.</p><p><strong>Conclusions: </strong>CGM is valued by adults with T2DM in primary care, yet barriers remain. Tailored support for initiation, troubleshooting, and education (especially alarm management and data interpretation) is needed. These insights can inform scalable strategies to enhance CGM use and experience in primary care.</p>","PeriodicalId":52371,"journal":{"name":"JMIR Diabetes","volume":"10 ","pages":"e73446"},"PeriodicalIF":2.6,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12753101/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145866150","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号