Exploring the association of social connections and food security among adults with uncontrolled type 2 diabetes: a population-based study.

BACKGROUND: Primary health care professionals are held accountable for various quality measures in the treatment of patients with chronic diseases such as diabetes. Uncontrolled type 2 diabetes (T2D) remains a considerable health problem; thus, further studying patients with this condition is important for delivering effective interventions. Social determinants of health (SDoH) have been shown to affect various aspects of diabetes care in different subpopulations. We studied the association of SDoH with uncontrolled T2D in a population of adult primary care patients. METHODS: We retrospectively searched our electronic health record for adult patients (≥18 years) with a diagnosis of T2D and a hemoglobin A1c (HbA1c) level of 8% or higher. Patients were empaneled to 2 primary care clinic sites between January 1, 2021, and January 31, 2022. Patients were grouped by HbA1c level to stratify patients according to the extent of uncontrolled T2D. Patient characteristics were compared among groups. Unadjusted and adjusted multinomial logistic regression analysis was used to estimate the odds of various SDoH factors among patient groups with different levels of uncontrolled T2D. RESULTS: The study cohort included 1,596 patients. Most patients were White (79%), and the median age was 58.8 years. The median HbA1c level was 8.9%, and approximately 68% of patients were obese (body mass index [BMI] ≥30). When the study population was grouped by HbA1c level (8% to < 9% [n = 806], ≥9% to < 12% [n = 684], and ≥12% [n = 106]), significant differences among groups were observed in age group (P < .001), marital status (P < .001), race (P < .001), ethnicity (P = .001), and BMI category (P = .01). In groups with higher HbA1c levels, we noticed a higher percentage of patients who were aged 51 to 65 years or single. Among patients with uncontrolled HbA1c levels, more patients were obese than overweight. Patients in the intermediate HbA1c group had increased odds of food insecurity and some decreased social connections, even after adjusting for age, sex, race, ethnicity, and marital status. CONCLUSIONS: Among patients with uncontrolled T2D, higher HbA1c levels were associated with decreased social connections and increased food insecurity. Our findings provide insight into the role of these SDoH in managing T2D and have important implications for primary care practice.

Validation of Different Dementia Code-Based Definitions in a Population-Based Study of Rheumatoid Arthritis.

OBJECTIVE: Rheumatoid arthritis (RA) has been associated with an elevated dementia risk. This study aimed to examine how different diagnostic dementia definitions perform in patients with RA compared to individuals without RA. METHODS: The study population included 2050 individuals (1025 with RA) from a retrospective, population-based cohort in southern Minnesota and compared the performance of 3 code-based dementia diagnostic algorithms with medical record review diagnosis of dementia. For the overall comparison, each patient's complete medical history was used, with no time frames. Sensitivity analyses were performed using 1-, 2-, and 5-year windows around the date that dementia was identified in the medical record (reference standard). RESULTS: Algorithms performed very similarly in persons with and without RA. The algorithms generally had high specificity, negative predictive values, and accuracy, regardless of the time window studied (> 88%). Sensitivity and positive predictive values varied depending on the algorithm and the time window. Sensitivity values ranged 56.5-95.9%, and positive predictive values ranged 55.2-83.1%. Performance measures declined with more restrictive time windows. CONCLUSION: Routinely collected electronic health record (EHR) data were used to define code-based dementia diagnostic algorithms with good performance (vs diagnosis by medical record review). These results can inform future studies that use retrospective databases, especially in the same or a similar EHR infrastructure, to identify dementia in individuals with RA.

Parent-reported Barriers and Parental Beliefs Associated with Intentions to Obtain HPV Vaccination for Children in a Primary care Patient Population in Minnesota, USA.

Human papillomavirus (HPV) vaccine uptake among adolescents remains suboptimal in the US. The COVID-19 pandemic posed new challenges to increase HPV vaccination rates. To characterize parent-reported barriers to obtain HPV vaccination for their children and to identify psychosocial factors associated with parents' intention to vaccinate their children for HPV, we administered parent surveys between April 2020 and January 2022 during a randomized pragmatic trial assessing the impact of evidence-based implementation strategies on HPV vaccination rates for adolescent patients at six Mayo Clinic primary care practices in Southeast Minnesota. A total of 342 surveys were completed (response rate 34.1%). Analyses were focused on parents of unvaccinated children (n = 133). The survey assessed the main reason the child did not receive the HPV vaccine, parental beliefs about the vaccine, and the parent's intention to vaccinate the child for HPV in the next 12 months. Frequently reported awareness and access barriers to HPV vaccination included not knowing the child was due (17.8%) and COVID-19 related delay (11.6%). Frequently reported attitudinal barriers include the belief that the child was too young for the vaccine (17.8%) and that the vaccine is not proven to be safe (16.3%). Injunctive social norm (Adjusted-OR = 3.15, 95%CI: 1.94, 5.41) and perceived harm beliefs (Adjusted-OR = 0.58, 95%CI: 0.35, 0.94) about the HPV vaccine were positively and negatively associated with HPV vaccination intention, respectively. Our findings suggest that continued efforts to overcome parental awareness, access, and attitudinal barriers to HPV vaccination are needed and underscore the importance of utilizing evidence-based health system-level interventions.

Low-to-Moderate Dosage and Short-Term Use of Corticosteroids Benefit Patients With Severe COVID-19 Infections.

Although the FDA has given emergency use authorization (EUA) for some antiviral drugs for the treatment of COVID-19, no direct antiviral drugs have been identified for the treatment of critically ill patients, the most important treatment is suppression of the hyperinflammation. The purpose of this study was to evaluate the role of corticosteroids in hospitalized severe or critical patients positive for COVID-19. This is a retrospective single-center descriptive study. Patients classified as having severe or critical COVID-19 infections with acute respiratory dysfunction syndrome in Shenzhen Third People's Hospital were enrolled from January 11th to March 30th, 2020. Ninety patients were classified as having severe or critical COVID-19 infections. The patients were treated with methylprednisolone with a low-to-moderate dosage and short duration. The days from the symptom onset to methylprednisolone were about 8 days. Eighteen patients were treated with invasive ventilation and intensive care unit (ICU) care. All the patients in the severe group and ten in the critical group recovered and were discharged. Three critical cases with invasive ventilation died. Although cases were much more severe in the corticosteroid-treated group, the mortality was not significantly increased. Early use of low-to-moderate dosage and short duration of corticosteroid may be the more accurate immune-modulatory treatment and brings more benefits to severe patients with COVID-19.

A hybrid model to identify fall occurrence from electronic health records.

INTRODUCTION: Falls are a leading cause of unintentional injury in the elderly. Electronic health records (EHRs) offer the unique opportunity to develop models that can identify fall events. However, identifying fall events in clinical notes requires advanced natural language processing (NLP) to simultaneously address multiple issues because the word "fall" is a typical homonym. METHODS: We implemented a context-aware language model, Bidirectional Encoder Representations from Transformers (BERT) to identify falls from the EHR text and further fused the BERT model into a hybrid architecture coupled with post-hoc heuristic rules to enhance the performance. The models were evaluated on real world EHR data and were compared to conventional rule-based and deep learning models (CNN and Bi-LSTM). To better understand the ability of each approach to identify falls, we further categorize fall-related concepts (i.e., risk of fall, prevention of fall, homonym) and performed a detailed error analysis. RESULTS: The hybrid model achieved the highest f1-score on sentence (0.971), document (0.985), and patient (0.954) level. At the sentence level (basic data unit in the model), the hybrid model had 0.954, 1.000, 0.988, and 0.999 in sensitivity, specificity, positive predictive value, and negative predictive value, respectively. The error analysis showed that that machine learning-based approaches demonstrated higher performance than a rule-based approach in challenging cases that required contextual understanding. The context-aware language model (BERT) slightly outperformed the word embedding approach trained on Bi-LSTM. No single model yielded the best performance for all fall-related semantic categories. CONCLUSION: A context-aware language model (BERT) was able to identify challenging fall events that requires context understanding in EHR free text. The hybrid model combined with post-hoc rules allowed a custom fix on the BERT outcomes and further improved the performance of fall detection.

Natural Language Processing for the Evaluation of Methodological Standards and Best Practices of EHR-based Clinical Research.

The effective use of EHR data for clinical research is challenged by the lack of methodologic standards, transparency, and reproducibility. For example, our empirical analysis on clinical research ontologies and reporting standards found little-to-no informatics-related standards. To address these issues, our study aims to leverage natural language processing techniques to discover the reporting patterns and data abstraction methodologies for EHR-based clinical research. We conducted a case study using a collection of full articles of EHR-based population studies published using the Rochester Epidemiology Project infrastructure. Our investigation discovered an upward trend of reporting EHR-related research methodologies, good practice, and the use of informatics related methods. For example, among 1279 articles, 24.0% reported training for data abstraction, 6% reported the abstractors were blinded, 4.5% tested the inter-observer agreement, 5% reported the use of a screening/data collection protocol, 1.5% reported that team meetings were organized for consensus building, and 0.8% mentioned supervision activities by senior researchers. Despite that, the overall ratio of reporting/adoption of methodologic standards was still low. There was also a high variation regarding clinical research reporting. Thus, continuously developing process frameworks, ontologies, and reporting guidelines for promoting good data practice in EHR-based clinical research are recommended.

Deep Learning Prediction of Mild Cognitive Impairment using Electronic Health Records.

About 44.4 million people have been diagnosed with dementia worldwide, and it is estimated that this number will be almost tripled by 2050. Predicting mild cognitive impairment (MCI), an intermediate state between normal cognition and dementia and an important risk factor for the development of dementia is crucial in aging populations. MCI is formally determined by health professionals through a comprehensive cognitive evaluation, together with a clinical examination, medical history and often the input of an informant (an individual that know the patient very well). However, this is not routinely performed in primary care visits, and could result in a significant delay in diagnosis. In this study, we used deep learning and machine learning techniques to predict the progression from cognitively unimpaired to MCI and also to analyze the potential for patient clustering using routinely-collected electronic health records (EHRs). Our analysis of EHRs indicates that temporal characteristics of patient data incorporated in a deep learning model provides increased power in predicting MCI.

Wellness Coaching to Improve Lifestyle Behaviors Among Adults With Prediabetes: Patients' Experience and Perceptions to Participation.

BACKGROUND: Health and Wellness Coaching has been shown to enhance treatment outcomes in the primary care setting. However, little is known about the experience and perceptions of patients who worked with a wellness coach as an integrated member of their primary health-care team. OBJECTIVE: This project assessed patients' experience and obtained their perceptions on barriers and facilitators to participation in a primary care-based wellness coaching program. METHOD: A survey was mailed to 99 primary care patients with prediabetes who participated in a 12-week wellness coaching program. RESULTS: Sixty-two (63%) completed the survey; responders felt that participation in the wellness coaching program helped move them toward healthier lifestyle behavior and created a personal vision of wellness. Major themes associated with participation were supportive coaching relationship, increased self-accountability, increased goal-setting, and healthy behavior strategies. No significant barrier to participation was reported. CONCLUSION: Participants reported highly positive experience with the program; how to best integrate health and wellness coaching into the primary care setting needs to be explored.

Development of Distinct Body Mass Index Trajectories Among Children Before Age 5 Years: A Population-Based Study.

The prevalence of childhood obesity has increased over the past 3 decades. This study was designed to understand how childhood body mass index (BMI) influences later risk of obesity. We calculated BMIs for children residing in Olmsted County, Minnesota, between January 1, 2005 and December 31, 2012 using medical records data. We defined homogenous BMI trajectory clusters using a nonparametric hill-climbing algorithm. Overall, 16,538 (47%) children had >3 weight assessments at least 1 year apart and were included in the analyses. Within the 8-year follow-up period, children who were younger than 2 years and overweight had a 3- fold increase of obesity (adjusted hazard ratio [HR] = 3.24; 95% confidence interval [CI] = 2.69-3.89) and those aged 5 years and overweight had a 10-fold increased risk of obesity (adjusted HR = 9.97, 95% CI = 8.55-11.62). Three distinct BMI trajectories could be distinguished prior to 5 years of age. The risk of developing obesity in those who are overweight increased dramatically with increasing age. Interventions to prevent obesity need to occur prior to school age to prevent children from entering unhealthy BMI trajectories.

Incidence Rate of Prediabetes Progression to Diabetes: Modeling an Optimum Target Group for Intervention.

Thirty-seven percent of US adults have prediabetes. Various interventions can delay diabetes progression; however, the optimum target group for risk reduction is uncertain. This study estimated rate of progression to diabetes at 1 and 5 years among a cohort of patients from 3 primary care clinics and modeled the potential magnitude in diabetes incidence risk reduction of an intervention program among specific subgroups. Records of 106,821 empaneled patients in 2005 were reviewed. Generalized population attributable risk (PAR) statistics were calculated to estimate the impact of reducing fasting blood glucose on diabetes progression. Multiple intervention effects (varying levels of glucose reduction along with multiple adherence rates) were examined for those with baseline glucose from 110 to 119 mg/dL and ≥120 mg/dL. Ten percent of patients (n = 10,796) met criteria for prediabetes. The 1- and 5-year diabetes incidence rate was 38.6 and 40.24 per 1000 person-years, respectively. Age and obesity were independent predictors of increased progression rate. The generalized PAR for a 10-point reduction in the 110-119 mg/dL subgroup with 25% adherence was 7.6%. The generalized PAR for similar percent reduction and adherence level in patients with baseline glucose of ≥120 mg/dL was only 3.0%. Rate of progression to diabetes increased over time and with associated independent risk factors. Greater risk reduction in diabetes progression within the target population can be achieved when the intervention is successful in those with baseline glucose of 110-119 mg/dL. Modeling an optimum target group for a diabetes prevention intervention offers a novel and useful guide to planning and allocating resources in population health management.