Non-negative matrix factorization temporal topic models and clinical text data identify COVID-19 pandemic effects on primary healthcare and community health in Toronto, Canada.

OBJECTIVE: To demonstrate how non-negative matrix factorization can be used to learn a temporal topic model over a large collection of primary care clinical notes, characterizing diverse COVID-19 pandemic effects on the physical/mental/social health of residents of Toronto, Canada. MATERIALS AND METHODS: The study employs a retrospective open cohort design, consisting of 382,666 primary care progress notes from 44,828 patients, 54 physicians, and 12 clinics collected 01/01/2017 through 31/12/2020. Non-negative matrix factorization uncovers a meaningful latent topical structure permeating the corpus of primary care notes. The learned latent topical basis is transformed into a multivariate time series data structure. Time series methods and plots showcase the evolution/dynamics of learned topics over the study period and allow the identification of COVID-19 pandemic effects. We perform several post-hoc checks of model robustness to increase trust that descriptive/unsupervised inferences are stable over hyper-parameter configurations and/or data perturbations. RESULTS: Temporal topic modelling uncovers a myriad of pandemic-related effects from the expressive clinical text data. In terms of direct effects on patient-health, topics encoding respiratory disease symptoms display altered dynamics during the pandemic year. Further, the pandemic was associated with a multitude of indirect patient-level effects on topical domains representing mental health, sleep, social and familial dynamics, measurement of vitals/labs, uptake of prevention/screening maneuvers, and referrals to medical specialists. Finally, topic models capture changes in primary care practice patterns resulting from the pandemic, including changes in EMR documentation strategies and the uptake of telemedicine. CONCLUSION: Temporal topic modelling applied to a large corpus of rich primary care clinical text data, can identify a meaningful topical/thematic summarization which can provide policymakers and public health stakeholders a passive, cost-effective, technology for understanding holistic impacts of the COVID-19 pandemic on the primary healthcare system and community/public-health.

Over-use of thyroid testing in Canadian and UK primary care in frequent attenders: A cross-sectional study.

BACKGROUND: Thyroid-stimulating hormone (TSH) is a common test used to detect and monitor clinically significant hypo- and hyperthyroidism. Population-based screening of asymptomatic adults for thyroid disorders is not recommended. OBJECTIVE: The research objectives were to determine patterns of TSH testing in Canadian and English primary care practices, as well as patient and physician practice characteristics associated with testing TSH for primary care patients with no identifiable indication. METHODS: In this 2-year cross-sectional observational study, Canadian and English electronic medical record databases were used to identify patients and physician practices. Cohorts of patients aged 18 years or older, without identifiable indications for TSH testing, were generated from these databases. Analyses were performed using a random-effects logistic regression to determine patient and physician practice characteristics associated with increased testing. We determined the proportion of TSH tests performed concurrently with at least one common screening blood test (lipid profile or hemoglobin A1c). Standardised proportions of TSH test per family practice were used to examine the heterogeneity in the populations. RESULTS: At least one TSH test was performed in 35.97% (N = 489 663) of Canadian patients and 29.36% (N = 1 030 489) of English patients. Almost all TSH tests in Canada and England (95.69% and 99.23% respectively) were within the normal range (0.40-5.00 mU/L). A greater number of patient-physician encounters was the strongest predictor of TSH testing. It was determined that 51.40% of TSH tests in Canada and 76.55% in England were performed on the same day as at least one other screening blood test. There was no association between the practice size and proportion of asymptomatic patients tested. CONCLUSIONS: This comparative binational study found TSH patterns suggestive of over-testing and potentially thyroid disorder screening in both countries. There may be significant opportunities to improve the appropriateness of TSH ordering in Canada and England and therefore improve the allocation of limited system resources.

Implementation of targeted screening for poverty in a large primary care team in Toronto, Canada: a feasibility study.

BACKGROUND: Poverty has a significant influence on health. Efforts to optimize income and reduce poverty could make a difference to the lives of patients and their families. Routine screening for poverty in primary care is an important first step but rarely occurs in Canada. We aimed to implement a targeted screening and referral process in a large, distributed primary care team in Toronto, Ontario, Canada. The main outcome was the proportion of targeted patients screened. METHODS: This implementation evaluation was conducted with a large community-based primary care team in north Toronto. The primary care team serves relatively wealthy neighborhoods with pockets of poverty. Physicians were invited to participate. We implemented targeted screening by combining census information on neighborhood-level deprivation with postal codes in patient records. For physicians agreeing to participate, we added prompts to screen for poverty to the charts of adult patients living in the most deprived areas. Standardized electronic medical record templates recommended a referral to a team case worker for income optimization, for those patients screening positive. We recorded the number and percentages of participants at each stage, from screening to receiving advice on income optimization. RESULTS: 128 targeted patients with at least one visit (25%) were screened. The primary care team included 86 physicians distributed across 19 clinical locations. Thirty-four physicians (39%) participated. Their practices provided care for 27,290 patients aged 18 or older; 852 patients (3%) were found to be living in the most deprived neighborhoods. 509 (60%) had at least one office visit over the 6 months of follow up. 25 patients (20%) screened positive for poverty, and 13 (52%) were referred. Eight patients (62% of those referred) were ultimately seen by a caseworker for income optimization. CONCLUSIONS: We implemented a targeted poverty screening program combined with resources to optimize income for patients in a large, distributed community-based primary care team. Screening was feasible; however, only a small number of patients were linked to the intervention Further efforts to scale and spread screening and mitigation of poverty are warranted; these should include broadening the targeted population beyond those living in the most deprived areas.

Choosing Wisely Canada campaign associated with less overuse of thyroid testing: Retrospective parallel cohort study.

OBJECTIVE: To assess the effectiveness of a Choosing Wisely Canada (CWC) initiative to improve thyroid-stimulating hormone (TSH) test ordering for patients with no identified indication for this test. DESIGN: Retrospective parallel cohort study using routinely collected electronic medical record (EMR) data. The CWC initiative included supporting primary care leads in each participating family health team, providing education on better test ordering, and allowing adaptation appropriate to each local context. SETTING: Toronto, Ont, and surrounding areas. PARTICIPANTS: Family physicians contributing EMR data to the University of Toronto Practice-Based Research Network and their patients aged 18 or older. MAIN OUTCOME MEASURES: Proportion of adult patients with a TSH test done in a 2-year period (2016 to 2017) in the absence of EMR data with an indication for TSH testing; proportion of TSH test results in the normal range for those patients; and change in the rate of TSH screening in sites participating in the CWC initiative compared with sites not participating. RESULTS: A total of 150 944 patients (51.7% of studied adults) had no identified indication for TSH testing; 33.4% of those patients were seen by physicians in the family health teams participating in the CWC initiative. Overall, 35.1% of all patients with no identified indication had at least 1 TSH test between January 1, 2016, and December 31, 2017. The 119 physicians participating in the CWC initiative decreased their monthly rate of testing by 0.23% from 2016 to 2017, a relative reduction of 13.2%. The 233 physicians not participating decreased testing by 0.04%, a relative reduction of 1.8%. The monthly difference between the 2 groups was 0.19% (95% CI -0.02 to -0.35 P = .03), a relative difference of 11.4%. The TSH testing decreased for almost all CWC patient subgroups. More than 95% of patients tested in both groups had TSH results in the normal range. CONCLUSION: This study found high rates of TSH testing without identified indications in the practices studied. A CWC initiative implemented in primary care was effective in reducing TSH testing.

Association between socio-economic status and hemoglobin A1c levels in a Canadian primary care adult population without diabetes.

BACKGROUND: Hgb A1c levels may be higher in persons without diabetes of lower socio-economic status (SES) but evidence about this association is limited; there is therefore uncertainty about the inclusion of SES in clinical decision support tools informing the provision and frequency of Hgb A1c tests to screen for diabetes. We studied the association between neighborhood-level SES and Hgb A1c in a primary care population without diabetes. METHODS: This is a retrospective study using data routinely collected in the electronic medical records (EMRs) of forty six community-based family physicians in Toronto, Ontario. We analysed records from 4,870 patients without diabetes, age 45 and over, with at least one clinical encounter between January 1st 2009 and December 31st 2011 and one or more Hgb A1c report present in their chart during that time interval. Residential postal codes were used to assign neighborhood deprivation indices and income levels by quintiles. Covariates included elements known to be associated with an increase in the risk of incident diabetes: age, gender, family history of diabetes, body mass index, blood pressure, LDL cholesterol, HDL cholesterol, triglycerides, and fasting blood glucose. RESULTS: The difference in mean Hgb A1c between highest and lowest income quintiles was -0.04% (p = 0.005, 95% CI -0.07% to -0.01%), and between least deprived and most deprived was -0.05% (p = 0.003, 95% CI -0.09% to -0.02%) for material deprivation and 0.02% (p = 0.2, 95% CI -0.06% to 0.01%) for social deprivation. After adjustment for covariates, a marginally statistically significant difference in Hgb A1c between highest and lowest SES quintile (p = 0.04) remained in the material deprivation model, but not in the other models. CONCLUSIONS: We found a small inverse relationship between Hgb A1c and the material aspects of SES; this was largely attenuated once we adjusted for diabetes risk factors, indicating that an independent contribution of SES to increasing Hgb A1c may be limited. This study does not support the inclusion of SES in clinical decision support tools that inform the use of Hgb A1c for diabetes screening.

Developing a method to estimate practice denominators for a national Canadian electronic medical record database.

BACKGROUND: Calculating disease prevalence requires both a numerator (number of persons with a disease) and a matching denominator (the 'population at risk' being studied). Determining primary care practice denominators is challenging. OBJECTIVE: To develop and test a method to calculate primary care practice denominators. METHODS: We compared a 'corrected yearly contact group', or practice population, with the number of patients enrolled with practices. The yearly contact group was the set of patients with a visit noted in the electronic medical records during the past year. The correction factor was the proportion of patients that reported contacting their physician in the past year. Eighty-one physicians from Toronto and Kingston, Ontario, provided data. The main outcome measure was the ratio of practice population to the number of enrolled patients. Other measures included the change in ratio over 2 years, differences between locations, and differences by provider, practice and patient characteristics. RESULTS: The ratio of practice population to enrolled patients was 1.03 in 2010 (95% confidence interval 1.00 to 1.05) and 1.03 in 2011 (95% confidence interval 1.00 to 1.05). There was no change in the ratio over time. Ratios by location, provider or practice characteristics differed by less than 10%. There was a slight under-estimation of practice population for younger male patients and over-estimation for female patients. CONCLUSION: This method provided a denominator that was reasonably similar to the enrolled population and was stable over time and by location, provider and practice characteristics. In regions without patient enrollment, this may provide an estimate of practice denominators.

Using ICD-9 diagnostic codes for external validation of topic models derived from primary care electronic medical record clinical text data.

Background/Objectives: Unsupervised topic models are often used to facilitate improved understanding of large unstructured clinical text datasets. In this study we investigated how ICD-9 diagnostic codes, collected alongside clinical text data, could be used to establish concurrent-, convergent- and discriminant-validity of learned topic models. Design/Setting: Retrospective open cohort design. Data were collected from primary care clinics located in Toronto, Canada between 01/01/2017 through 12/31/2020. Methods: We fit a non-negative matrix factorization topic model, with K = 50 latent topics/themes, to our input document term matrix (DTM). We estimated the magnitude of association between each Boolean-valued ICD-9 diagnostic code and each continuous latent topical vector. We identified ICD-9 diagnostic codes most strongly associated with each latent topical vector; and qualitatively interpreted how these codes could be used for external validation of the learned topic model. Results: The DTM consisted of 382,666 documents and 2210 words/tokens. We correlated concurrently assigned ICD-9 diagnostic codes with learned topical vectors, and observed semantic agreement for a subset of latent constructs (e.g. conditions of the breast, disorders of the female genital tract, respiratory disease, viral infection, eye/ear/nose/throat conditions, conditions of the urinary system, and dermatological conditions, etc.). Conclusions: When fitting topic models to clinical text corpora, researchers can leverage contemporaneously collected electronic medical record data to investigate the external validity of fitted latent variable models.

Multimorbidity and Blood Pressure Control in Patients Attending Primary Care in Canada.

BACKGROUND: There has been conflicting evidence on the association between multimorbidity and blood pressure (BP) control. This study aimed to investigate this associations in people with hypertension attending primary care in Canada, and to assess whether individual long-term conditions are associated with BP control. METHODS: This was a cross-sectional study in people with hypertension attending primary care in Toronto between January 1, 2017 and December 31, 2019. Uncontrolled BP was defined as systolic BP ≥ 140 mmHg or diastolic BP ≥ 90 mmHg. A list of 11 a priori selected chronic conditions was used to define multimorbidity. Multimorbidity was defined as having ≥1 long-term condition in addition to hypertension. Logistic regression models were used to estimate the association between multimorbidity (or individual long-term conditions) with uncontrolled BP. RESULTS: A total of 67 385 patients with hypertension were included. They had a mean age of 70, 53.1% were female, 80.6% had multimorbidity, and 35.7% had uncontrolled BP. Patients with multimorbidity had lower odds of uncontrolled BP than those without multimorbidity (adjusted OR = 0.72, 95% CI 0.68-0.76). Among the long-term conditions, diabetes (aOR = 0.73, 95%CI 0.70-0.77), heart failure (aOR = 0.81, 95%CI 0.73-0.91), ischemic heart disease (aOR = 0.74, 95%CI 0.69-0.79), schizophrenia (aOR = 0.79, 95%CI 0.65-0.97), depression/anxiety (aOR = 0.91, 95%CI 0.86-0.95), dementia (aOR = 0.87, 95%CI 0.80-0.95), and osteoarthritis (aOR = 0.89, 95%CI 0.85-0.93) were associated with a lower likelihood of uncontrolled BP. CONCLUSION: We found that multimorbidity was associated with better BP control. Several conditions were associated with better control, including diabetes, heart failure, ischemic heart disease, schizophrenia, depression/anxiety, dementia, and osteoarthritis.

Using Primary Care Clinical Text Data and Natural Language Processing to Identify Indicators of COVID-19 in Toronto, Canada.

The objective of this study was to investigate whether a rule-based natural language processing (NLP) system, applied to primary care clinical text data, could be used to monitor COVID-19 viral activity in Toronto, Canada. We employed a retrospective cohort design. We included primary care patients with a clinical encounter between January 1, 2020 and December 31, 2020 at one of 44 participating clinical sites. During the study timeframe, Toronto first experienced a COVID-19 outbreak between March-2020 and June-2020; followed by a second viral resurgence from October-2020 through December-2020. We used an expert derived dictionary, pattern matching tools and contextual analyzer to classify primary care documents as 1) COVID-19 positive, 2) COVID-19 negative, or 3) unknown COVID-19 status. We applied the COVID-19 biosurveillance system across three primary care electronic medical record text streams: 1) lab text, 2) health condition diagnosis text and 3) clinical notes. We enumerated COVID-19 entities in the clinical text and estimated the proportion of patients with a positive COVID-19 record. We constructed a primary care COVID-19 NLP-derived time series and investigated its correlation with independent/external public health series: 1) lab confirmed COVID-19 cases, 2) COVID-19 hospitalizations, 3) COVID-19 ICU admissions, and 4) COVID-19 intubations. A total of 196,440 unique patients were observed over the study timeframe, of which 4,580 (2.3%) had at least one positive COVID-19 document in their primary care electronic medical record. Our NLP-derived COVID-19 time series describing the temporal dynamics of COVID-19 positivity status over the study timeframe demonstrated a pattern/trend which strongly mirrored that of other external public health series under investigation. We conclude that primary care text data passively collected from electronic medical record systems represent a high quality, low-cost source of information for monitoring/surveilling COVID-19 impacts on community health.

Marginal Structural Models Using Calibrated Weights With SuperLearner: Application to Type II Diabetes Cohort.

As different scientific disciplines begin to converge on machine learning for causal inference, we demonstrate the application of machine learning algorithms in the context of longitudinal causal estimation using electronic health records. Our aim is to formulate a marginal structural model for estimating diabetes care provisions in which we envisioned hypothetical (i.e. counterfactual) dynamic treatment regimes using a combination of drug therapies to manage diabetes: metformin, sulfonylurea and SGLT-2i. The binary outcome of diabetes care provisions was defined using a composite measure of chronic disease prevention and screening elements [27] including (i) primary care visit, (ii) blood pressure, (iii) weight, (iv) hemoglobin A1c, (v) lipid, (vi) ACR, (vii) eGFR and (viii) statin medication. We used several statistical learning algorithms to describe causal relationships between the prescription of three common classes of diabetes medications and quality of diabetes care using the electronic health records contained in National Diabetes Repository. In particular, we generated an ensemble of statistical learning algorithms using the SuperLearner framework based on the following base learners: (i) least absolute shrinkage and selection operator, (ii) ridge regression, (iii) elastic net, (iv) random forest, (v) gradient boosting machines, and (vi) neural network. Each statistical learning algorithm was fitted using the pseudo-population generated from the marginalization of the time-dependent confounding process. Covariate balance was assessed using the longitudinal (i.e. cumulative-time product) stabilized weights with calibrated restrictions. Our results indicated that the treatment drop-in cohorts (with respect to metformin, sulfonylurea and SGLT-2i) may have improved diabetes care provisions in relation to treatment naïve (i.e. no treatment) cohort. As a clinical utility, we hope that this article will facilitate discussions around the prevention of adverse chronic outcomes associated with type II diabetes through the improvement of diabetes care provisions in primary care.

Specialist wait time reporting using family physicians' electronic medical record data: a mixed method study of feasibility and clinical utility.

BACKGROUND: More than 50% of Canadian adult patients wait longer than four weeks to see a specialist after referral from primary care. Access to accurate wait time information may help primary care physicians choose the timeliest specialist to address a patient's specific needs. We conducted a mixed-methods study to assess if primary to specialist care wait times can be extracted from electronic medical records (EMR), analyzed the wait time information, and used focus groups and interviews to assess the potential clinical utility of the wait time information. METHODS: Two family practices were recruited to examine primary care physician to specialist wait times between January 2016 and December 2017, using EMR data. The primary outcome was the median wait time from physician referral to specialist appointment for each specialty service. Secondary outcomes included the physician and patient characteristics associated with wait times as well as qualitative analyses of physician interviews about the resulting wait time reports. RESULTS: Wait time data can be extracted from the primary care EMR and converted to a report format for family physicians and specialists to review. After data cleaning, there were 7141 referrals included from 4967 unique patients. The 5 most common specialties referred to were Dermatology, Gastroenterology, Ear Nose and Throat, Obstetrics and Gynecology and Urology. Half of the patients were seen by a specialist within 42 days, 75% seen within 80 days and all patients within 760 days. There were significant differences in wait times by specialty, for younger patients, and those with urgently labelled medical situations. Overall, wait time reports were perceived by clinicians to be important since they could help family physicians decide how to triage referrals and might lead to system improvements. CONCLUSIONS: Wait time information from primary to specialist care can aid in decision-making around specialist referrals, identify bottlenecks, and help with system planning. This mixed method study is a starting point to review the importance of providing wait time data for both family physicians, specialists and local health systems. Future work can be directed towards developing wait time reporting functionality and evaluating if wait time information will help increase system efficiency and/or improve provider and patient satisfaction.

Health system use among patients with mental health conditions in a community based sample in Toronto, Canada: A retrospective cohort study.

OBJECTIVE: To identify hospital and primary care health service use among people with mental health conditions or addictions in an integrated primary-secondary care database in Toronto, Ontario. METHOD: This was a retrospective cohort study of adults with mental health diagnoses using data from the Health Databank Collaborative (HDC), a primary care-hospital linked database in Toronto. Data were included up to March 31st 2019. Negative binomial and logistic regression were used to evaluate associations between health care utilization and various patient characteristics and mental health diagnoses. RESULTS: 28,482 patients age 18 or older were included. The adjusted odds of at least one mental health diagnosis were higher among younger patients (18-30 years vs. 81+years aOR = 1.87; 95% CI:1.68-2.08) and among female patients (aOR = 1.35; 95% CI: 1.27-1.42). Patients with one or more mental health diagnoses had higher adjusted rates of hospital visits compared to those without any mental health diagnosis including addiction (aRR = 1.74, 95% CI: 1.58-1.91) and anxiety (aRR = 1.28, 95% CI: 1.23-1.32). 14.5% of patients with a psychiatric diagnosis were referred to the hospital for specialized psychiatric services, and 38% of patients referred were eventually seen in consultation. The median wait time from the date of referral to the date of consultation was 133 days. CONCLUSIONS: In this community, individuals with mental health diagnoses accessed primary and hospital-based health care at greater rates than those without mental health diagnoses. Wait times for specialized psychiatric care were long and most patients who were referred did not have a consultation. Information about services for patients with mental health conditions can be used to plan and monitor more integrated care across sectors, and ultimately improve outcomes.

Primary Care Severe Asthma Registry and Education Project (PCSAR-EDU): Phase 1 - an e-Delphi for registry definitions and indices of clinician behaviour.

INTRODUCTION: Although most asthma is mild to moderate, severe asthma accounts for disproportionate personal and societal costs. Poor co-ordination of care between primary care and specialist settings is recognised as a barrier to achieving optimal outcomes. The Primary Care Severe Asthma Registry and Education (PCSAR-EDU) project aims to address these gaps through the interdisciplinary development and evaluation of both a 'real-world' severe asthma registry and an educational programme for primary care providers. This manuscript describes phase 1 of PCSAR-EDU which involves establishing interdisciplinary consensus on criteria for the: (1) definition of severe asthma; (2) generation of a severe asthma registry and (3) definition of an electronic-medical record data-based Clinician Behaviour Index (CBI). METHODS AND ANALYSIS: In phase 1, a modified e-Delphi activity will be conducted. Delphi panellists (n≥13) will be invited to complete a 30 min online survey on three separate occasions (i.e., three separate e-Delphi 'rounds') over a 3-month period. Expert opinion will be collected via an open-ended survey ('Open' round 1) and 5-point Likert scale and ranking surveys ('Closed' round 2 and 3). A fourth and final Delphi round will occur via synchronous meeting, whereby panellists approve a finalised ideal 'core criteria list', CBI and corresponding item weighting. ETHICS AND DISSEMINATION: Ethical approval has been obtained for the activities involved in phase 1 from the University of Toronto's Human Research Ethics Programme (approval number 39695). Future ethics approvals will depend on information gathered in the proceeding phase; thus, ethical approval for phase 2 and 3 of this study will be sought sequentially. Findings will be disseminated through conference presentations, peer-reviewed publications and knowledge translation tools.

Diabetes screening with hemoglobin A1c prior to a change in guideline recommendations: prevalence and patient characteristics.

BACKGROUND: In January 2010, the American Diabetes Association recommended the use of hemoglobin A1c (Hgb A1c) to screen and diagnose diabetes. This study explored the prevalence and clinical context of Hgb A1c tests done for non-diabetic primary care patients for the three years prior to the release of the new guidelines. We sought to determine the provision of tests in non-diabetic patients age 19 or over, patients age 45 and over (eligible for routine diabetes screening), the annual change in the rate of this screening test, and the patient characteristics associated with the provision of Hgb A1c screening. METHODS: We conducted a retrospective study using data routinely collected in Electronic Medical Records. The participants were thirteen community-based family physicians in Toronto, Ontario. We calculated the proportion of non diabetic patients who had at least one Hbg A1c done in three years. We used logistic generalized estimating equation with year treated as a continuous variable to test for a non-zero slope in yearly Hbg A1c provision. We modelled screening using multivariable logistic regression. RESULTS: There were 11,792 non-diabetic adults. Of these, 1,678 (14.2%; 95%CI 13.6%-14.9%) had at least one Hgb A1c test done; this was higher for patients 45 years of age or older (20.2%; 95% CI 19.3% - 21.2%). The proportion of non-diabetic patients with an A1c test increased from 5.2% in 2007 to 8.8% in 2009 (p < 0.0001 for presence of slope). Factors associated with significantly greater adjusted odds ratios of having the test done included increasing diastolic blood pressure, increasing fasting glucose, increasing body mass index, increasing age, as well as male gender and presence of hypertension, but not smoking status or LDL cholesterol. Patients living in the highest income quintile neighbourhoods had significantly lower odds ratios of having this test done than those in the lowest quintile (p < 0.001). CONCLUSIONS: A large and increasing proportion of the non-diabetic patients we studied have had an Hgb A1c for screening prior to guidelines recommending the test for this purpose. Several risk factors for cardiovascular disease or diabetes were associated with the provision of the Hgb A1c. Early uptake of the test may represent appropriate utilization.

Using a data entry clerk to improve data quality in primary care electronic medical records: a pilot study.

BACKGROUND: The quality of electronic medical record (EMR) data is known to be problematic; research on improving these data is needed. OBJECTIVE: The primary objective was to explore the impact of using a data entry clerk to improve data quality in primary care EMRs. The secondary objective was to evaluate the feasibility of implementing this intervention. METHODS: We used a before and after design for this pilot study. The participants were 13 community based family physicians and four allied health professionals in Toronto, Canada. Using queries programmed by a data manager, a data clerk was tasked with re-entering EMR information as coded or structured data for chronic obstructive pulmonary disease (COPD), smoking, specialist designations and interprofessional encounter headers. We measured data quality before and three to six months after the intervention. We evaluated feasibility by measuring acceptability to clinicians and workload for the clerk. RESULTS: After the intervention, coded COPD entries increased by 38% (P = 0.0001, 95% CI 23 to 51%); identifiable data on smoking categories increased by 27% (P = 0.0001, 95% CI 26 to 29%); referrals with specialist designations increased by 20% (P = 0.0001, 95% CI 16 to 22%); and identifiable interprofessional headers increased by 10% (P = 0.45, 95 CI -3 to 23%). Overall, the intervention was rated as being at least moderately useful and moderately usable. The data entry clerk spent 127 hours restructuring data for 11 729 patients. CONCLUSIONS: Utilising a data manager for queries and a data clerk to re-enter data led to improvements in EMR data quality. Clinicians found this approach to be acceptable.

Must we appear to be all-knowing?: patients' and family physicians' perspectives on information seeking during consultations.

OBJECTIVE: To compare patients' opinions about family physicians looking up medical information during consultations with family physicians' expectations of how patients would respond to their using sources to find answers to medical questions. DESIGN: Survey. SETTING: North York, Ont. PARTICIPANTS: One hundred fifty-three family practice patients, 54 family physicians, and 21 family practice residents. MAIN OUTCOME MEASURES: Patients' self-reported confidence in their family physicians and their perceptions of the quality of care after seeing physicians look up medical information, both without specifying the physician's source of information and with reference to several specific information media. Family physicians' predictions for how patients would respond to their using resources to answer medical questions. RESULTS: When the information source used by physicians was not specified, 9% and 7% of patients reported decreased confidence and perceived lower quality of care, respectively. When the information source used by physicians was specified, the proportions of negative responses for patients' confidence and their perceptions of quality of care were 39% and 31%, respectively, for Internet search engines (ISEs); 8% and 7% for online resources designed for physicians (ORDP); 27% and 27% for personal digital assistants (PDAs); and 10% and 9% for hard-copy medical textbooks (HMTs). When the information source was not specified, 32% and 12% of physicians expected patients to report negative responses for confidence and perceptions of quality of care, respectively. When the information source was specified, 51% and 33% of physicians expected patients to report negative responses for confidence and perceptions of quality of care, respectively, for their use of ISEs; 16% and 8% for ORDP; 20% and 12% for PDAs; and 36% and 21% for HMTs. Younger patients were more likely to respond negatively to physicians' use of resources, especially if the source was an ISE (P<.001). Physicians earlier in their careers were more likely to expect negative patient responses (P<.05). CONCLUSION: Family physicians overestimated the decrease in patients' confidence caused by seeing them look up medical questions. While most patients responded positively, a substantial proportion of younger patients reported decreased confidence. Patients believed the best sources of information were ORDP and HMTs.

Design and testing of an architecture for a national primary care chronic disease surveillance network in Canada.

Chronic diseases are a growing concern around the globe. In Canada, chronic disease care is taking an increasing share of health care budgets, and with an aging population, is threatening to overwhelm Provincial budgets where most health care is paid for. A chronic disease surveillance network fills an important gap in current public health surveillance systems. This paper describes the design and feasibility testing of an information technology and privacy architecture to extract, transform and transfer data from 7 electronic medical record systems used by 100 primary care providers in 6 province to a central data repository at the High Performance Computing Virtual Laboratory at Queen's University in Canada.

Trends in diabetes medication use in Australia, Canada, England, and Scotland: a repeated cross-sectional analysis in primary care.

BACKGROUND: Several new classes of glucose-lowering medications have been introduced in the past two decades. Some, such as sodium-glucose cotransporter 2 inhibitors (SGLT2s), have evidence of improved cardiovascular outcomes, while others, such as dipeptidyl peptidase-4 inhibitors (DPP4s), do not. It is therefore important to identify their uptake in order to find ways to support the use of more effective treatments. AIM: To analyse the uptake of these new classes among patients with type 2 diabetes. DESIGN AND SETTING: This was a retrospective repeated cross-sectional analysis in primary care. Rates of medication uptake in Australia, Canada, England, and Scotland were compared. METHOD: Primary care Electronic Medical Data on prescriptions (Canada, UK) and dispensing data (Australia) from 2012 to 2017 were used. Individuals aged ≥40 years on at least one glucose-lowering drug class in each year of interest were included, excluding those on insulin only. Proportions of patients in each nation, for each year, on each class of medication, and on combinations of classes were determined. RESULTS: Data from 238 619 patients were included in 2017. The proportion of patients on sulfonylureas (SUs) decreased in three out of four nations, while metformin decreased in Canada. Use of combinations of metformin and new drug classes increased in all nations, replacing combinations involving SUs. In 2017, more patients were on DPP4s (between 19.1% and 27.6%) than on SGLT2s (between 10.1% and 15.3%). CONCLUSION: New drugs are displacing SUs. However, despite evidence of better outcomes, the adoption of SGLT2s lagged behind DPP4s.

Cardiovascular risk factor documentation and management in primary care electronic medical records among people with schizophrenia in Ontario, Canada: retrospective cohort study.

OBJECTIVES: In order to address the substantial increased risk of cardiovascular disease among people with schizophrenia, it is necessary to identify the factors responsible for some of that increased risk. We analysed the extent to which these risk factors were documented in primary care electronic medical records (EMR), and compared their documentation by patient and provider characteristics. DESIGN: Retrospective cohort study. SETTING: EMR database of the University of Toronto Practice-Based Research Network Data Safe Haven. PARTICIPANTS: 197 129 adults between 40 and 75 years of age; 4882 with schizophrenia and 192 427 without. PRIMARY AND SECONDARY OUTCOME MEASURES: Documentation of cardiovascular disease risk factors (age, sex, smoking history, presence of diabetes, blood pressure, whether a patient is currently on medication to reduce blood pressure, total cholesterol and high-density lipoprotein cholesterol). RESULTS: Documentation of cardiovascular risk factors was more complete among people with schizophrenia (74.5% of whom had blood pressure documented at least once in the last 2 years vs 67.3% of those without, p>0.0001). Smoking status was not documented in 19.8% of those with schizophrenia and 20.8% of those without (p=0.0843). Factors associated with improved documentation included older patients (OR for ages 70-75 vs 45-49=3.51, 95% CI 3.26 to 3.78), male patients (OR=1.39, 95% CI 1.33 to 1.45), patients cared for by a female provider (OR=1.52, 95% CI 1.12 to 2.07) and increased number of encounters (OR for ≥10 visits vs 3-5 visits=1.53, 95% CI 1.46 to 1.60). CONCLUSIONS: Documentation of cardiovascular risk factors was better among people with schizophrenia than without, although overall documentation was inadequate. Efforts to improve documentation of risk factors are warranted in order to facilitate improved management.