Background: Diabetes Mellitus is an increasingly prevalent chronic disease characterized by the body’s inability to metabolize glucose. The objective of this study was to build an effective predictive model with high sensitivity and selectivity to better identify Canadian patients at risk of having Diabetes Mellitus based on patient demographic data and the laboratory results during their visits to medical facilities. Methods: Using the most recent records of 13,309 Canadian patients aged between 18 and 90 years, along with their laboratory information (age, sex, fasting blood glucose, body mass index, high-density lipoprotein, triglycerides, blood pressure, and low-density lipoprotein), we built predictive models using Logistic Regression and Gradient Boosting Machine (GBM) techniques. The area under the receiver operating characteristic curve (AROC) was used to evaluate the discriminatory capability of these models. We used the adjusted threshold method and the class weight method to improve sensitivity – the proportion of Diabetes Mellitus patients correctly predicted by the model. We also compared these models to other learning machine techniques such as Decision Tree and Random Forest. Results: The AROC for the proposed GBM model is 84.7% with a sensitivity of 71.6% and the AROC for the proposed Logistic Regression model is 84.0% with a sensitivity of 73.4%. The GBM and Logistic Regression models perform better than the Random Forest and Decision Tree models. Conclusions: The ability of our model to predict patients with Diabetes using some commonly used lab results is high with satisfactory sensitivity. These models can be built into an online computer program to help physicians in predicting patients with future occurrence of diabetes and providing necessary preventive interventions. The model is developed and validated on the Canadian population which is more specific and powerful to apply on Canadian patients than existing models developed from US or other populations. Fasting blood glucose, body mass index, high-density lipoprotein, and triglycerides were the most important predictors in these models.

Diabetes Mellitus is one of the major health challenges all over the world. The prevalence of diabetes is increasing at a fast pace, deteriorating human, economic and social fabric. Prevention and prediction of diabetes mellitus is increasingly gaining interest in healthcare community. Although several clinical decision support systems have been proposed that incorporate several data mining techniques for diabetes prediction and course of progression. These conventional systems are typically based either just on a single classifier or a plain combination thereof. Recently extensive endeavors are being made for improving the accuracy of such systems using ensemble classifiers. This study follows the adaboost and bagging ensemble techniques using J48 (c4.5) decision tree as a base learner along with standalone data mining technique J48 to classify patients with diabetes mellitus using diabetes risk factors. This classification is done across three different ordinal adults groups in Canadian Primary Care Sentinel Surveillance network. Experimental result shows that, overall performance of adaboost ensemble method is better than bagging as well as standalone J48 decision tree

Background: Diabetes mellitus is a complex disease with serious complications. Electronic decision support, providing information that is shared and discussed by both patient and physician, encourages timely interventions and may improve the management of this chronic disease. However, it has rarely been tested in community-based primary care. Methods: In this pragmatic randomized trial, we randomly assigned adult primary care patients with type 2 diabetes to receive the intervention or usual care. The intervention involved shared access by the primary care provider and the patient to a Web-based, colour-coded diabetes tracker, which provided sequential monitoring values for 13 diabetes risk factors, their respective targets and brief, prioritized messages of advice. The primary outcome measure was a process composite score. Secondary outcomes included clinical composite scores, quality of life, continuity of care and usability. The outcome assessors were blinded to each patient’s intervention status. Results: We recruited sequentially 46 primary care providers and then 511 of their patients (mean age 60.7 [standard deviation 12.5] years). Mean follow-up was 5.9 months. The process composite score was significantly better for patients in the intervention group than for control patients (difference 1.27, 95% confidence interval [CI] 0.79–1.75, p < 0.001); 61.7% (156/253) of patients in the intervention group, compared with 42.6% (110/258) of control patients, showed improvement (difference 19.1%, p < 0.001). The clinical composite score also had significantly more variables with improvement for the intervention group (0.59, 95% CI 0.09–1.10, p = 0.02), including significantly greater declines in blood pressure (−3.95 mm Hg systolic and −2.38 mm Hg diastolic) and glycated hemoglobin (−0.2%). Patients in the intervention group reported greater satisfaction with their diabetes care. Interpretation: A shared electronic decision-support system to support the primary care of diabetes improved the process of care and some clinical markers of the quality of diabetes care. (ClinicalTrials.gov trial register no. NCT00813085.)

Objectives: To assess patients' preferred method of consent for the use of information from electronic medical records for research. Design: Interviews and a structured survey of patients in practices with electronic medical records. Setting: Family practices in southern Ontario, Canada. Participants: 123 patients: 17 were interviewed and 106 completed a survey. Main outcome measures: Patients' opinions and concerns on use of information from their medical records for research and their preferences for method of consent. Results: Most interviewees were willing to allow the use of their information for research purposes, although the majority preferred that consent was sought first. The seeking of consent was considered an important element of respect for the individual. Most interviewees made little distinction between identifiable and anonymised data. Research sponsored by private insurance firms generated the greatest concern, and research sponsored by foundation the least. Sponsorship by drug companies evoked negative responses during interview and positive responses in the survey. Conclusions: Patients are willing to allow information from their medical records to be used for research, but most prefer to be asked for consent either verbally or in writing.

The development of a pan-Canadian network of primary care research networks for studying issues in primary care has been the vision of Canadian primary care researchers for many years. With the opportunity for funding from the Public Health Agency of Canada and the support of the College of Family Physicians of Canada, we have planned and developed a project to assess the feasibility of a network of networks of family medicine practices that exclusively use electronic medical records. The Canadian Primary Care Sentinel Surveillance Network will collect longitudinal data from practices across Canada to assess the primary care epidemiology and management of 5 chronic diseases: hypertension, diabetes, depression, chronic obstructive lung disease, and osteoarthritis. This article reports on the 7-month first phase of the feasibility project of 7 regional networks in Canada to develop a business plan, including governance, mission, and vision; develop memorandum of agreements with the regional networks and their respective universities; develop and obtain approval of research ethics board applications; develop methods for data extraction, a Canadian Primary Care Sentinel Surveillance Network database, and initial assessment of the types of data that can be extracted; and recruitment of 10 practices at each network that use electronic medical records. The project will continue in phase 2 of the feasibility testing until April 2010.

Prevention and diagnosis of NAFLD is an ongoing area of interest in the healthcare community. Screening is complicated by the fact that the accuracy of noninvasive testing lacks specificity and sensitivity to make and stage the diagnosis. Currently no non-invasive ATP III criteria based prediction method is available to diagnose NAFLD risk. Firstly, the objective of this research is to develop machine learning based method in order to identify individuals at an increased risk of developing NAFLD using risk factors of ATP III clinical criteria updated in 2005 for Metabolic Syndrome (MetS). Secondly, to validate the relative ability of quantitative score defined by Italian Association for the Study of the Liver (IASF) and guideline explicitly defined for the Canadian population based on triglyceride thresholds to predict NAFLD risk. We proposed a Decision Tree based method to evaluate the risk of developing NAFLD and its progression in the Canadian population, using Electronic Medical Records (EMRs) by exploring novel risk factors for NAFLD. Our results show proposed method could potentially help physicians make more informed choices about their management of patients with NAFLD. Employing the proposed application in ordinary medical checkup is expected to lessen healthcare expenditures compared with administering additional complicated test.