An interpretable predictive deep learning platform for pediatric metabolic diseases.

OBJECTIVES: Metabolic disease in children is increasing worldwide and predisposes a wide array of chronic comorbid conditions with severe impacts on quality of life. Tools for early detection are needed to promptly intervene to prevent or slow the development of these long-term complications. MATERIALS AND METHODS: No clinically available tools are currently in widespread use that can predict the onset of metabolic diseases in pediatric patients. Here, we use interpretable deep learning, leveraging longitudinal clinical measurements, demographical data, and diagnosis codes from electronic health record data from a large integrated health system to predict the onset of prediabetes, type 2 diabetes (T2D), and metabolic syndrome in pediatric cohorts. RESULTS: The cohort included 49 517 children with overweight or obesity aged 2-18 (54.9% male, 73% Caucasian), with a median follow-up time of 7.5 years and mean body mass index (BMI) percentile of 88.6%. Our model demonstrated area under receiver operating characteristic curve (AUC) accuracies up to 0.87, 0.79, and 0.79 for predicting T2D, metabolic syndrome, and prediabetes, respectively. Whereas most risk calculators use only recently available data, incorporating longitudinal data improved AUCs by 13.04%, 11.48%, and 11.67% for T2D, syndrome, and prediabetes, respectively, versus models using the most recent BMI (P < 2.2 × 10-16). DISCUSSION: Despite most risk calculators using only the most recent data, incorporating longitudinal data improved the model accuracies because utilizing trajectories provides a more comprehensive characterization of the patient's health history. Our interpretable model indicated that BMI trajectories were consistently identified as one of the most influential features for prediction, highlighting the advantages of incorporating longitudinal data when available.

Diagnosis and management of post-bariatric surgery hypoglycemia.

INTRODUCTION: While bariatric surgery remains the most effective treatment for obesity that allows substantial weight loss with improvement and possibly remission of obesity-associated comorbidities, some postoperative complications may occur. Managing physicians need to be familiar with the common problems to ensure timely and effective management. Of these complications, postoperative hypoglycemia is an increasingly recognized complication of bariatric surgery that remains underreported and underdiagnosed. AREA COVERED: This article highlights the importance of identifying hypoglycemia in patients with a history of bariatric surgery, reviews pathophysiology and addresses available nutritional, pharmacological and surgical management options. Systemic evaluation including careful history taking, confirmation of hypoglycemia and biochemical assessment is essential to establish accurate diagnosis. Understanding the weight-dependent and weight-independent mechanisms of improved postoperative glycemic control can provide better insight into the causes of the exaggerated responses that lead to postoperative hypoglycemia. EXPERT OPINION: Management of post-operative hypoglycemia can be challenging and requires a multidisciplinary approach. While dietary modification is the mainstay of treatment for most patients, some patients may benefit from pharmacotherapy (e.g. GLP-1 receptor antagonist); Surgery (e.g. reversal of gastric bypass) is reserved for unresponsive severe cases. Additional research is needed to understand the underlying pathophysiology with a primary aim in optimizing diagnostics and treatment options.

Psychological Characteristics and Goals of Caregivers Choosing Continuous Glucose Monitoring for Children With Type 1 Diabetes.

Background: Continuous glucose monitoring (CGM) can improve glycemic outcomes in pediatric type 1 diabetes management. However, its impact on the psychosocial functioning of caregivers is less well described. The objectives of this pilot study were to explore caregiver reasons for adding CGM to their child's type 1 diabetes management, parental psychosocial function before initiating CGM, and the relationship between the two. Methods: Thirty-two families with a child with type 1 diabetes from Rainbow Babies and Children's Hospital diabetes clinics who were initiating CGM were recruited over 3 months. Before CGM initiation, the caregivers completed the Fear of Hypoglycemia Scale, State-Trait Anxiety Inventory, Problem Areas in Diabetes Scale, and a questionnaire assessing their primary reason for starting CGM. Participant characteristics and questionnaire results were summarized with descriptive statistics. Participants were grouped by reason for starting CGM, and results were compared among groups using ANOVA and reporting the global F test. Results: Fifty percent of respondents indicated that they were starting CGM to optimize glycemic control. The majority of parents (71.8%) expressed worry about helplessness during a hypoglycemic episode. There were no statistically significant differences in patient characteristics or questionnaire results between groups. Forty-three percent of participating families started using CGM during the study's 8-month follow-up period. The main reason (64%) for not starting CGM was not having the chance to start the process of obtaining a CGM system. There were no statistically significant differences between children who did and did not start CGM. Conclusion: Caregivers have different reasons for starting CGM for their child with type 1 diabetes. Further studies are needed to understand whether these reasons are related to differences in psychosocial functioning. Despite interest in starting CGM, there remain barriers to implementation.

Extent of physician-pharmaceutical industry interactions in low- and middle-income countries: a systematic review.

Background: Pharmaceutical companies spend large amounts of money promoting their products to physicians. There is evidence that physicians' interactions with pharmaceutical companies negatively affect their prescribing patterns. The objective of this study was to systematically review the extent of the relationship between physicians and pharmaceutical companies in low- and middle-income countries (LMICs). Methods: Studies assessing the extent of any type of interaction between practicing physicians and pharmaceutical companies were eligible. We searched MEDLINE and EMBASE databases in July 2016. Reviewers worked in duplicate and independently to complete study selection, data abstraction and assessment of methodological features. We summarized the findings narratively. Results: We included 11 eligible studies (7 quantitative and 4 qualitative). Quantitative studies found that pharmaceutical company representatives visited at least 90% of physicians. Printed material, stationery items and drug samples were the most frequently received gifts. Two of the studies assessing direct payment found percentages of 16 and 5%, respectively. Findings of qualitative studies were consistent with those of quantitative studies. In addition, they revealed an increasing tendency for pharmaceutical companies to provide expensive personal gifts, sponsor social events and offer cash as inducements to physicians based on their demands. They also identified building personal relationships, creating a sense of indebtedness and emotional blackmailing as commonly used techniques to influence physicians. Conclusion: A relatively high percentage of physicians in LMICs interact with pharmaceutical companies. Findings have implications for policy and practice, given the current extent of interaction is likely affecting the prescribing habits and professional behaviour of physicians.

Knowledge, beliefs and attitudes of physicians in low and middle-income countries regarding interacting with pharmaceutical companies: a systematic review.

BACKGROUND: Understanding the perceptions and attitudes of physicians is important. This knowledge assists in the efforts to reduce the impact of their interactions with the pharmaceutical industry on clinical practice. It appears that most studies on such perceptions and attitudes have been conducted in high-income countries. The objective was to systematically review the knowledge, beliefs and attitudes of physicians in low and middle-income countries regarding interactions with pharmaceutical companies. METHODS: Eligible studies addressed any type of interaction between physicians and pharmaceutical companies. The outcomes of interest included knowledge, beliefs and attitudes of practicing physicians. The search strategy covered MEDLINE and EMBASE databases. Two reviewers completed in duplicate and independently study selection, data abstraction, and assessment of methodological features. The data synthesis consisted of a narrative summary of the findings stratified by knowledge, beliefs and attitudes. RESULTS: We included ten reports from nine eligible studies, each of which had a number of methodological limitations. Four studies found that the top perceived benefits of this interaction were receiving information and rewards. In five out of eight studies assessing the perception regarding the impact of the interaction on the behavior of physician prescription, the majority of participants believed it to be minor. In one of these studies, participants perceived that impact to be lesser when asked about their own behavior. The attitudes of physicians towards information and rewards provided by pharmaceutical company representatives (PCRs) (assessed in 5 and 2 studies respectively) varied across studies. In the only study assessing their attitudes towards pharmaceutical-sponsored Continuing Medical Education, physicians considered local conferences to have higher impact. Their attitudes towards developing policies restricting physicians' interactions with PCRs were positive in two studies. In one study, the majority of participants did not mind the public knowing that physicians were receiving gifts and awards from drug companies. CONCLUSIONS: This review identified few studies conducted in low and middle-income countries. While physicians generally perceived the impact of interactions on their behavior to be minor, their attitudes toward receiving information and rewards varied across studies.

Validity of tools used for surveying physicians about their interactions with pharmaceutical company: a systematic review.

BACKGROUND: There is evidence that physicians' prescription behavior is negatively affected by the extent of their interactions with pharmaceutical companies. In order to develop and implement policies and interventions for better management of interactions, we need to understand physicians' perspectives on this issue. Surveys addressing physicians' interactions with pharmaceutical companies need to use validated tools to ensure the validity of their findings. OBJECTIVE: To assess the validity of tools used in surveys of physicians about the extent and nature of their interactions with pharmaceutical companies, and about their knowledge, beliefs and attitudes towards such interactions; and to identify those tools that have been formally validated. METHODS: We developed a search strategy with the assistance of a medical librarian. We electronically searched MEDLINE and EMBASE databases in September 2015. Teams of two reviewers conducted data selection and data abstraction. They identified eligible studies in one table and then abstracted the relevant data from the studies with validated tools in another table. Tables were piloted and standardized. RESULTS: We identified one validated questionnaire out of the 11 assessing the nature and extent of the interaction, and three validated questionnaires out of the 47 assessing knowledge, beliefs and attitudes of physicians toward the interaction. None of these validated questionnaires were used in more than one survey. CONCLUSION: The available supporting evidence of the issue of physicians' interaction with pharmaceutical company is of low quality. There is a need for research to develop and validate tools to survey physicians about their interactions with pharmaceutical companies.

Vitamin D supplementation in children with asthma: a systematic review and meta-analysis.

BACKGROUND: Epidemiologic studies suggest an association between vitamin D deficiency and atopic diseases, including asthma. The objective of this study was to systematically review the benefits and harms of vitamin D supplementation in children with asthma. METHODS: We used standard Cochrane systematic review methodology. The search strategy included an electronic search in February 2013 of MEDLINE and EMBASE. Two reviewers completed in duplicate and independently study selection, data abstraction, and assessment of risk of bias. We pooled the results of trials using a random-effects model. We assessed the quality of evidence by outcome using the GRADE methodology. RESULTS: Four trials with a total of 149 children met eligibility criteria. The trials had major methodological limitations. Given the four studies reporting on asthma symptoms used different instruments to measure that outcome, we opted not to conduct a meta-analysis. Three of those studies reported improvement in asthma symptoms in the vitamin D supplemented group study, while the fourth reported no effect (very low quality evidence). For the lung function outcome, a meta-analysis of two trials assessing post treatment FEV-1 found a mean difference of 0.54 liters per second (95% CI -5.28; 4.19; low quality evidence). For the vitamin D level outcome, a meta-analysis of three trials found a mean difference of 6.56 ng/ml (95% CI -0.64; 13.77; very low quality evidence). CONCLUSIONS: The available very low to low quality evidence does not confirm or rule out beneficial effects of vitamin D supplementation in children with asthma. Large-scale, well-designed and executed randomized controlled trials are needed to better understand the effectiveness and safety of vitamin D in children with asthma.

Legislative, educational, policy and other interventions targeting physicians' interaction with pharmaceutical companies: a systematic review.

BACKGROUND: Pharmaceutical company representatives likely influence the prescribing habits and professional behaviour of physicians. OBJECTIVE: The objective of this study was to systematically review the effects of interventions targeting practising physicians' interactions with pharmaceutical companies. ELIGIBILITY CRITERIA: We included observational studies, non-randomised controlled trials (non-RCTs) and RCTs evaluating legislative, educational, policy or other interventions targeting the interactions between physicians and pharmaceutical companies. DATA SOURCES: The search strategy included an electronic search of MEDLINE and EMBASE. Two reviewers performed duplicate and independent study selection, data abstraction and assessment of risk of bias. APPRAISAL AND SYNTHESIS METHODS: We assessed the risk of bias in each included study. We summarised the findings narratively because the nature of the data did not allow a meta-analysis to be conducted. We assessed the quality of evidence by outcome using the GRADE methodology. RESULTS: Of 11 189 identified citations, one RCT and three observational studies met the eligibility criteria. All four studies specifically targeted one type of interaction with pharmaceutical companies, that is, interactions with drug representatives. The RCT provided moderate quality evidence of no effect of a 'collaborative approach' between the pharmaceutical industry and a health authority. The three observational studies provided low quality evidence suggesting a positive effect of policies aiming to reduce interaction between physicians and pharmaceutical companies (by restricting free samples, promotional material, and meetings with pharmaceutical company representatives) on prescription behaviour. LIMITATIONS: We identified too few studies to allow strong conclusions. CONCLUSIONS: Available evidence suggests a potential impact of policies aiming to reduce interaction between physicians and drug representatives on physicians' prescription behaviour. We found no evidence concerning interventions affecting other types of interaction with pharmaceutical companies.

Association between low vitamin D levels and the diagnosis of asthma in children: a systematic review of cohort studies.

BACKGROUND: There is conflicting evidence about the association between low vitamin D levels in children and development of asthma in later life. The objective of this study was to systematically review the evidence for an epidemiological association between low serum levels of vitamin D and the diagnosis of asthma in children. METHODS: We used the Cochrane methodology for conducting systematic reviews. The search strategy included an electronic search of MEDLINE and EMBASE in February 2013. Two reviewers completed, in duplicate and independently, study selection, data abstraction, and assessment of risk of bias. RESULTS: Of 1081 identified citations, three cohort studies met eligibility criteria. Two studies found that low serum vitamin D level is associated with an increased risk of developing asthma late in childhood, while the third study found no association with either vitamin D2 or vitamin D3 levels. All three studies suffer from major methodological shortcomings that limit our confidence in their results. CONCLUSIONS: Available epidemiological evidence suggests a potential association between low serum levels of vitamin D and the diagnosis of asthma in children. High quality studies are needed to reliably answer the question of interest.