Machine and Deep Learning for Tuberculosis Detection on Chest X-Rays: Systematic Literature Review.

BACKGROUND: Tuberculosis (TB) was the leading infectious cause of mortality globally prior to COVID-19 and chest radiography has an important role in the detection, and subsequent diagnosis, of patients with this disease. The conventional experts reading has substantial within- and between-observer variability, indicating poor reliability of human readers. Substantial efforts have been made in utilizing various artificial intelligence-based algorithms to address the limitations of human reading of chest radiographs for diagnosing TB. OBJECTIVE: This systematic literature review (SLR) aims to assess the performance of machine learning (ML) and deep learning (DL) in the detection of TB using chest radiography (chest x-ray [CXR]). METHODS: In conducting and reporting the SLR, we followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A total of 309 records were identified from Scopus, PubMed, and IEEE (Institute of Electrical and Electronics Engineers) databases. We independently screened, reviewed, and assessed all available records and included 47 studies that met the inclusion criteria in this SLR. We also performed the risk of bias assessment using Quality Assessment of Diagnostic Accuracy Studies version 2 (QUADAS-2) and meta-analysis of 10 included studies that provided confusion matrix results. RESULTS: Various CXR data sets have been used in the included studies, with 2 of the most popular ones being Montgomery County (n=29) and Shenzhen (n=36) data sets. DL (n=34) was more commonly used than ML (n=7) in the included studies. Most studies used human radiologist's report as the reference standard. Support vector machine (n=5), k-nearest neighbors (n=3), and random forest (n=2) were the most popular ML approaches. Meanwhile, convolutional neural networks were the most commonly used DL techniques, with the 4 most popular applications being ResNet-50 (n=11), VGG-16 (n=8), VGG-19 (n=7), and AlexNet (n=6). Four performance metrics were popularly used, namely, accuracy (n=35), area under the curve (AUC; n=34), sensitivity (n=27), and specificity (n=23). In terms of the performance results, ML showed higher accuracy (mean ~93.71%) and sensitivity (mean ~92.55%), while on average DL models achieved better AUC (mean ~92.12%) and specificity (mean ~91.54%). Based on data from 10 studies that provided confusion matrix results, we estimated the pooled sensitivity and specificity of ML and DL methods to be 0.9857 (95% CI 0.9477-1.00) and 0.9805 (95% CI 0.9255-1.00), respectively. From the risk of bias assessment, 17 studies were regarded as having unclear risks for the reference standard aspect and 6 studies were regarded as having unclear risks for the flow and timing aspect. Only 2 included studies had built applications based on the proposed solutions. CONCLUSIONS: Findings from this SLR confirm the high potential of both ML and DL for TB detection using CXR. Future studies need to pay a close attention on 2 aspects of risk of bias, namely, the reference standard and the flow and timing aspects. TRIAL REGISTRATION: PROSPERO CRD42021277155; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=277155.

Seek COVER: using a disease proxy to rapidly develop and validate a personalized risk calculator for COVID-19 outcomes in an international network.

BACKGROUND: We investigated whether we could use influenza data to develop prediction models for COVID-19 to increase the speed at which prediction models can reliably be developed and validated early in a pandemic. We developed COVID-19 Estimated Risk (COVER) scores that quantify a patient's risk of hospital admission with pneumonia (COVER-H), hospitalization with pneumonia requiring intensive services or death (COVER-I), or fatality (COVER-F) in the 30-days following COVID-19 diagnosis using historical data from patients with influenza or flu-like symptoms and tested this in COVID-19 patients. METHODS: We analyzed a federated network of electronic medical records and administrative claims data from 14 data sources and 6 countries containing data collected on or before 4/27/2020. We used a 2-step process to develop 3 scores using historical data from patients with influenza or flu-like symptoms any time prior to 2020. The first step was to create a data-driven model using LASSO regularized logistic regression, the covariates of which were used to develop aggregate covariates for the second step where the COVER scores were developed using a smaller set of features. These 3 COVER scores were then externally validated on patients with 1) influenza or flu-like symptoms and 2) confirmed or suspected COVID-19 diagnosis across 5 databases from South Korea, Spain, and the United States. Outcomes included i) hospitalization with pneumonia, ii) hospitalization with pneumonia requiring intensive services or death, and iii) death in the 30 days after index date. RESULTS: Overall, 44,507 COVID-19 patients were included for model validation. We identified 7 predictors (history of cancer, chronic obstructive pulmonary disease, diabetes, heart disease, hypertension, hyperlipidemia, kidney disease) which combined with age and sex discriminated which patients would experience any of our three outcomes. The models achieved good performance in influenza and COVID-19 cohorts. For COVID-19 the AUC ranges were, COVER-H: 0.69-0.81, COVER-I: 0.73-0.91, and COVER-F: 0.72-0.90. Calibration varied across the validations with some of the COVID-19 validations being less well calibrated than the influenza validations. CONCLUSIONS: This research demonstrated the utility of using a proxy disease to develop a prediction model. The 3 COVER models with 9-predictors that were developed using influenza data perform well for COVID-19 patients for predicting hospitalization, intensive services, and fatality. The scores showed good discriminatory performance which transferred well to the COVID-19 population. There was some miscalibration in the COVID-19 validations, which is potentially due to the difference in symptom severity between the two diseases. A possible solution for this is to recalibrate the models in each location before use.

Statin Prescription Patterns and Associations with Subclinical Inflammation.

Background and Objectives: Statins have been extensively utilised in atherosclerotic cardiovascular disease (ASCVD) prevention and can inhibit inflammation. However, the association between statin therapy, subclinical inflammation and associated health outcomes is poorly understood in the primary care setting. Materials and Methods: Primary care electronic health record (EHR) data from the electronic Practice-Based Research Network (ePBRN) from 2012−2019 was used to assess statin usage and adherence in South-Western Sydney (SWS), Australia. Independent determinants of elevated C-reactive protein (CRP) were determined. The relationship between baseline CRP levels and hospitalisation rates at 12 months was investigated. Results: The prevalence of lipid-lowering medications was 14.0% in all adults and 44.6% in the elderly (≥65 years). The prevalence increased from 2012 to 2019 despite a drop in statin use between 2013−2015. A total of 55% of individuals had good adherence (>80%). Hydrophilic statin use and higher intensity statin therapy were associated with elevated CRP levels. However, elevated CRP levels were not associated with all-cause or ASCVD hospitalisations after adjusting for confounders. Conclusions: The prevalence and adherence patterns associated with lipid-lowering medications highlighted the elevated ASCVD-related burden in the SWS population, especially when compared with the Australian general population. Patients in SWS may benefit from enhanced screening protocols, targeted health literacy and promotion campaigns, and timely incorporation of evidence into ASCVD clinical guidelines. This study, which used EHR data, did not support the use of CRP as an independent marker of future short-term hospitalisations.

A New Method for Estimating the Incidence of Infectious Diseases.

Ambitious World Health Organization targets for disease elimination require monitoring of epidemics using routine health data in settings of decreasing and low incidence. We evaluated 2 methods commonly applied to routine testing results to estimate incidence rates that assume a uniform probability of infection between consecutive negative and positive tests based on 1) the midpoint of this interval and 2) a randomly selected point in this interval. We compared these with an approximation of the Poisson binomial distribution, which assigns partial incidence to time periods based on the uniform probability of occurrence in these intervals. We assessed bias, variance, and convergence of estimates using simulations of Weibull-distributed failure times with systematically varied baseline incidence and varying trend. We considered results for quarterly, half-yearly, and yearly incidence estimation frequencies. We applied the methods to assess human immunodeficiency virus (HIV) incidence in HIV-negative patients from the Treatment With Antiretrovirals and Their Impact on Positive and Negative Men (TAIPAN) Study, an Australian study of HIV incidence in men who have sex with men, between 2012 and 2018. The Poisson binomial method had reduced bias and variance at low levels of incidence and for increased estimation frequency, with increased consistency of estimation. Application of methods to real-world assessment of HIV incidence found decreased variance in Poisson binomial model estimates, with observed incidence declining to levels where simulation results had indicated bias in midpoint and random-point methods.

National Heart Foundation of Australia: position statement on coronary artery calcium scoring for the primary prevention of cardiovascular disease in Australia.

INTRODUCTION: This position statement considers the evolving evidence on the use of coronary artery calcium scoring (CAC) for defining cardiovascular risk in the context of Australian practice and provides advice to health professionals regarding the use of CAC scoring in primary prevention of cardiovascular disease in Australia. Main recommendations: CAC scoring could be considered for selected people with moderate absolute cardiovascular risk, as assessed by the National Vascular Disease Prevention Alliance (NVDPA) absolute cardiovascular risk algorithm, and for whom the findings are likely to influence the intensity of risk management. (GRADE evidence certainty: Low. GRADE recommendation strength: Conditional.) CAC scoring could be considered for selected people with low absolute cardiovascular risk, as assessed by the NVDPA absolute cardiovascular risk algorithm, and who have additional risk-enhancing factors that may result in the underestimation of risk. (GRADE evidence certainty: Low. GRADE recommendation strength: Conditional.) If CAC scoring is undertaken, a CAC score of 0 AU could reclassify a person to a low absolute cardiovascular risk status, with subsequent management to be informed by patient-clinician discussion and follow contemporary recommendations for low absolute cardiovascular risk. (GRADE evidence certainty: Very low. GRADE recommendation strength: Conditional.) If CAC scoring is undertaken, a CAC score > 99 AU or ≥ 75th percentile for age and sex could reclassify a person to a high absolute cardiovascular risk status, with subsequent management to be informed by patient-clinician discussion and follow contemporary recommendations for high absolute cardiovascular risk. (GRADE evidence certainty: Very low. GRADE recommendation strength: Conditional.) CHANGES IN MANAGEMENT AS A RESULT OF THIS STATEMENT: CAC scoring can have a role in reclassification of absolute cardiovascular risk for selected patients in Australia, in conjunction with traditional absolute risk assessment and as part of a shared decision-making approach that considers the preferences and values of individual patients.

Online training in evidence-based medicine and research methods for GP registrars: a mixed-methods evaluation of engagement and impact.

BACKGROUND: Evidence-based medicine (EBM) is a core skillset for enhancing the quality and safety of patients' care. Online EBM education could improve clinicians' skills in EBM, particularly when it is conducted during vocational training. There are limited studies on the impact of online EBM training on clinical practice among general practitioner (GP) registrars (trainees in specialist general practice). We aimed to describe and evaluate the acceptability, utility, satisfaction and applicability of the GP registrars experience with the online course. The course was developed by content-matter experts with educational designers to encompass effective teaching methods (e.g. it was interactive and used multiple teaching methods). METHODS: Mixed-method data collection was conducted after individual registrars' completion of the course. The course comprised six modules that aimed to increase knowledge of research methods and application of EBM skills to everyday practice. GP registrars who completed the online course during 2016-2020 were invited to complete an online survey about their experience and satisfaction with the course. Those who completed the course within the six months prior to data collection were invited to participate in semi-structured phone interviews about their experience with the course and the impact of the course on clinical practice. A thematic analysis approach was used to analyse the data from qualitative interviews. RESULTS: The data showed the registrars were generally positive towards the course and the concept of EBM. They stated that the course improved their confidence, knowledge, and skills and consequently impacted their practice. The students perceived the course increased their understanding of EBM with a Cohen's d of 1.6. Registrars identified factors that influenced the impact of the course. Of those, some were GP-related including their perception of EBM, and being comfortable with what they already learnt; some were work-place related such as time, the influence of supervisors, access to resources; and one was related to patient preferences. CONCLUSIONS: This study showed that GP registrars who attended the online course reported that it improved their knowledge, confidence, skill and practice of EBM over the period of three months. The study highlights the supervisor's role on GP registrars' ability in translating the EBM skills learnt in to practice and suggests exploring the effect of EBM training for supervisors.

Cultural respect in general practice: a cluster randomised controlled trial.

OBJECTIVE: To examine whether the Ways of Thinking and Ways of Doing (WoTWoD) cultural respect framework improves clinically appropriate anticipatory care in general practice and the cultural respect levels of medical practice staff. DESIGN: Mixed methods, cluster randomised controlled trial with a participatory action research approach. SETTING, PARTICIPANTS: Fifty-six general practices in Sydney and Melbourne, 2014-2017. INTERVENTION: WoTWoD encompasses a toolkit (ten scenarios illustrating cross-cultural behaviour in clinical practice), one half-day workshop, cultural mentor support for practices, and a local care partnership between participating Medicare locals/primary health networks and local Aboriginal Community Controlled Health Services for guiding the program and facilitating community engagement. The intervention lasted 12 months at each practice. MAJOR OUTCOMES: Rates of claims for MBS item 715 (health assessment for Aboriginal and Torres Strait Islander People) and recording of chronic disease risk factors; changes in cultural quotient (CQ) scores of practice staff. RESULTS: Complete results were available for 28 intervention (135 GPs, 807 Indigenous patients) and 25 control practices (210 GPs, 1554 Indigenous patients). 12-Month rates of MBS item 715 claims and recording of risk factors for the two groups were not statistically significantly different, nor were mean changes in CQ scores, regardless of staff category and practice attributes. CONCLUSION: The WoTWoD program did not increase the rate of Indigenous health checks or improve cultural respect scores in general practice. Conceptual, methodologic, and contextual factors that influence cultural mentorship, culturally respectful clinical practice, and Indigenous health care require further investigation. TRIAL REGISTRATION: Australia New Zealand Clinical Trials Registry ACTRN12614000797673.

Statistical supervised meta-ensemble algorithm for medical record linkage.

Identifying unique patients across multiple care facilities or services is a major challenge in providing continuous care and undertaking health research. Identifying and linking patients without compromising privacy and security is an emerging issue in the big data era. The large quantity and complexity of the patient data emphasize the need for effective linkage methods that are both scalable and accurate. In this study, we aim to develop and evaluate an ensemble classification method using the three most typically used supervised learning methods, namely support vector machines, logistic regression and standard feed-forward neural networks, to link records that belong to the same patient across multiple service locations. Our ensemble method is the combination of bagging and stacking. Each base learner's critical hyperparameters were selected through grid search technique. Two synthetic datasets were used in this study namely FEBRL and ePBRN. ePBRN linkage dataset was based on linkage errors noticed in the Australian primary care setting. The overall linkage performance was determined by assessing the blocking performance and classification performance. Our ensemble method outperformed the base learners in all evaluation metrics on one dataset. More specifically, the precision, which is average of individual precision scores in case of base learners increased from 90.70% to 94.85% in FEBRL, and from 62.17% to 99.28% in ePBRN. Similarly, the F-score increased from 94.92% to 98.18% in FEBRL, and from 72.99% to 91.72% in ePBRN. Our experiments suggest that we can significantly improve the linkage performance of individual algorithms by employing ensemble strategies.

Linking observational data from general practice, hospital admissions and diabetes clinic databases: can it be used to predict hospital admission?

BACKGROUND: Linking process of care data from general practice (GP) and hospital data may provide more information about the risk of hospital admission and re-admission for people with type-2 diabetes mellitus (T2DM). This study aimed to extract and link data from a hospital, a diabetes clinic (DC). A second aim was to determine whether the data could be used to predict hospital admission for people with T2DM. METHODS: Data were extracted using the GRHANITE™ extraction and linkage tool. The data from nine GPs and the DC included data from the two years prior to the hospital admission. The date of the first hospital admission for patients with one or more admissions was the index admission. For those patients without an admission, the census date 31/03/2014 was used in all outputs requiring results prior to an admission. Readmission was any admission following the index admission. The data were summarised to provide a comparison between two groups of patients: 1) Patients with a diagnosis of T2DM who had been treated at a GP and had a hospital admission and 2) Patients with a diagnosis of T2DM who had been treated at a GP and did not have a hospital admission. RESULTS: Data were extracted for 161,575 patients from the three data sources, 644 patients with T2DM had data linked between the GPs and the hospital. Of these, 170 also had data linked with the DC. Combining the data from the different data sources improved the overall data quality for some attributes particularly those attributes that were recorded consistently in the hospital admission data. The results from the modelling to predict hospital admission were plausible given the issues with data completeness. CONCLUSION: This project has established the methodology (tools and processes) to extract, link, aggregate and analyse data from general practices, hospital admission data and DC data. This study methodology involved the establishment of a comparator/control group from the same sites to compare and contrast the predictors of admission, addressing a limitation of most published risk stratification and admission prediction studies. Data completeness needs to be improved for this to be useful to predict hospital admissions.

Gender differentials in readiness and use of mHealth services in a rural area of Bangladesh.

BACKGROUND: Traditional gender roles result in women lagging behind men in the use of modern technologies, especially in developing countries. Although there is rapid uptake of mobile phone use in Bangladesh, investigation of gender differences in the ownership, access and use of mobile phones in general and mHealth in particular has been limited. This paper presents gender differentials in the ownership of mobile phones and knowledge of available mHealth services in a rural area of Bangladesh. METHODS: We interviewed 4915 randomly selected respondents aged 18 years and above. Associations between gender and knowledge of available mHealth services, use of existing mHealth services and intentions to use mHealth services in the future were examined by multivariate logistic regression analysis, controlling for the effect of categorised covariates. RESULTS: Of the 4915 respondents to the survey, 61.8% of men (1213/1964) and 34.4% of women (1015/2951) owned a mobile phone. For men, mobile phone ownership was highest among those aged 18-29 years (n = 663, 76.3%), and for women among those aged 30-39 years (n = 825, 44.7%). A higher proportion of men owned phones compared to women, irrespective of socioeconomic status (SES) as indicated by asset index (p < 0.001). Although mobile phone ownership on average was lower among women, they were more likely to share their mobile phone with their family members (19.7%) compared to men (11.6%, p < 0.001). Greater number of men were more likely to be aware of the use of mobile phones for healthcare compared to women (38.5% vs 26.5%, p < 0.001). Knowledge about available mHealth services was lower among women than men; however, intention to use mHealth services in the future was high for both genders, irrespective of age, education and socioeconomic status. CONCLUSIONS: Compared to men, women are less likely to own a mobile phone and less aware of available mHealth services, despite high intention to use mHealth among both genders. To optimise the use of mHealth services and to achieve equity of use, uptake strategies should target women, with a focus on the poorer and less educated groups.

Can research that is not intended or unlikely to be published be considered ethical?

BACKGROUND: For research to be ethically acceptable, the potential benefits must justify any risks involved for participants. Dissemination of research findings through publication is one way of creating benefit, but not all researchers intend to publish their research. Other factors, such as lack of size or representativeness, generalisability or innovativeness, or negative findings mean the research is unlikely to be published in a peer-reviewed medical journal. OBJECTIVE: This paper discusses ethical considerations in research where peer-reviewed publication is not intended or unlikely. DISCUSSION: Proposing research that is not intended or unlikely to be published in a peer-reviewed journal does not preclude it from being considered ethical. Additional benefits of such projects may include professional development of investigators, pilot data collection leading to more definitive studies, or developing collaborations with research users that increase relevance and improve utility of findings.

A survey of Sydney general practitioners' management of patients with chronic hepatitis B.

OBJECTIVE: To examine the chronic hepatitis B (CHB) assessment and management practices of general practitioners in the Sydney and South Western Sydney Local Health Districts, areas with a high prevalence of CHB, and to obtain their views on alternative models of care. DESIGN, SETTING AND PARTICIPANTS: We used a descriptive, cross-sectional study design to survey GPs who had seen at least one patient aged 18 years or over who had been notified as having CHB to the Public Health Unit between 1 June 2012 and 31 May 2013. There were 213 eligible GPs; the response rate was 57.7%. MAIN OUTCOME MEASURES: The CHB assessment, management and referral practices of the GPs, and their opinions about different models of care. RESULTS: Most GPs (78.9%) were at least reasonably confident about managing CHB. GPs were generally most comfortable with a model of care that involved initial referral to a specialist; managing CHB without specialist input or with only review by a specialised nurse practitioner were less popular. CONCLUSION: These results suggest that barriers, including dependence on specialist input, still hinder the appropriate assessment and management of CHB patients by GPs. Well designed and targeted support programs that include specialist support are needed if there is to be a successful shift to an increased role for GPs in the model of care for managing CHB.

Alcohol screening and brief interventions in primary care - Evidence and a pragmatic practice-based approach.

BACKGROUND: Risky alcohol drinking is a common problem in adults presenting in Australian general practice. Preventive health guidelines recommend routine delivery of alcohol screening and brief intervention (ASBI) by general practitioners (GPs). However, ASBIs have rarely been implemented successfully in a sustainable manner. OBJECTIVE: In this article, we explain the current state of empirical evidence for the effectiveness of ASBI in primary care and describe a pragmatic interpretation of how this evidence applies to routine care. DISCUSSION: The empirical evidence surrounding ASBIs is complex. ASBIs are efficacious in research settings, but their effectiveness when compared with control interventions in real-world practice is less certain. Alcohol assessment within therapeutic doctor-patient relationships, rather than the specific formal tools, may be the 'active ingredient'. A pragmatic, practice-based approach to early detection of risky drinking is to focus on strategies that allow asking patients about their drinking more regularly, documenting it, and using quality improvement methodology to improve alcohol recording data completeness for the practice population.

Ethical considerations in recruiting primary care patients to research studies.

BACKGROUND: How patients are selected and subsequently invited to take part in research has important implications for gaining informed, voluntary consent. OBJECTIVE: This article identifies and discusses common ethical issues that are faced by researchers when recruiting patients from primary care settings. DISCUSSION: Recruiting primary care patients for research studies should be guided by the core ethical values of merit and integrity, respect, justice and beneficence. Issues of patient privacy and risk of coercion are major concerns when selecting and recruiting primary care patients, but the ethical issues will depend on the type of research and the potential risks to participants. The National Statement on Ethical Conduct in Human Research, and Australian privacy laws and principles, should be reviewed to ensure recruitment meets contemporary ethical standards prior to submitting a study protocol for ethical review.

Safe and effective cultural mentorship in general practice.

BACKGROUND: Closing the gap in health and welfare for Aboriginal and Torres Strait Islander peoples is an ongoing challenge. OBJECTIVE: The objectives of this article are to conceptualise and operationalise models of cultural mentorship within a multifaceted practice-based program to facilitate culturally and clinically appropriate care. METHODS: Participatory action research and workshops were conducted with Aboriginal Elders, Aboriginal health workers (AHWs), Indigenous health project officers (IHPOs) and staff from participating Medicare Locals (MLs). RESULTS: Roles and responsibilities in a cultural mentorship relationship were defined, along with potential benefits and harm. Mentors and mentees should be comfortable with their own identity and/or ethnicity before engaging in a mentorship relationship. Mutual trust is implicit and participants must be prepared, flexible and mutually respectful to achieve mutual goals. The cultural mentorship model includes Aboriginal Elders and local care partnerships of Aboriginal community-controlled and primary care organisations, and practice mentorship teams of a local AHW/IHPO, research project officer and, where available, ML practice support officer. DISCUSSION: A successful cultural mentorship model is multi-level and safe, and requires mutual trust and respect, time and resources. Potential benefits include equitable access to, and use of, safe quality care for Aboriginal and Torres Strait Islander patients in general practice.

Optimising the use of observational electronic health record data: Current issues, evolving opportunities, strategies and scope for collaboration.

BACKGROUND: With increasing computerisation in general practice, national primary care networks are mooted as sources of data for health services and population health research and planning. Existing data collection programs - MedicinesInsight, Improvement Foundation, Bettering the Evaluation and Care of Health (BEACH) - vary in purpose, governance, methodologies and tools. General practitioners (GPs) have significant roles as collectors, managers and users of electronic health record (EHR) data. They need to understand the challenges to their clinical and managerial roles and responsibilities. OBJECTIVE: The aim of this article is to examine the primary and secondary use of EHR data, identify challenges, discuss solutions and explore directions. DISCUSSION: Representatives from existing programs, Medicare Locals, Local Health Districts and research networks held workshops on the scope, challenges and approaches to the quality and use of EHR data. Challenges included data quality, interoperability, fragmented governance, proprietary software, transparency, sustainability, competing ethical and privacy perspectives, and cognitive load on patients and clinicians. Proposed solutions included effective change management; transparent governance and management of intellectual property, data quality, security, ethical access, and privacy; common data models, metadata and tools; and patient/community engagement. Collaboration and common approaches to tools, platforms and governance are needed. Processes and structures must be transparent and acceptable to GPs.

Is there a role for a primary health nurse in a learning support team in a disadvantaged high school? Evaluation of a pilot study.

Disadvantaged children experience more health problems and have poorer educational outcomes compared with students from advantaged backgrounds. This paper presents the quantitative and qualitative findings from a pilot study to determine the impact of the Healthy Learner model, where an experienced primary care nurse was embedded in a learning support team in a disadvantaged high school. Students entering high school with National Assessment Program, Literacy and Numeracy (NAPLAN) scores in the lowest quartile for the school were assessed by the nurse and identified health issues addressed. Thirty-nine students were assessed in 2012-13 and there were up to seven health problems identified per student, ranging from serious neglect to problems such as uncorrected vision or hearing. Many of these problems were having an impact on the student and their ability to engage in learning. Families struggled to navigate the health system, they had difficulty explaining the student's problems to health professionals and costs were a barrier. Adding a nurse to the learning support team in this disadvantaged high school was feasible and identified considerable unmet health needs that affect a student's ability to learn. The families needed extensive support to access any subsequent health care they required.

Psychological distress among Vietnamese adults attending Vietnamese-speaking general practices in South Western Sydney: prevalence and associations.

Objective: The aim of the present study was to determine the prevalence of psychological distress among Vietnamese adults attending Vietnamese-speaking general practices and explore possible risk factors in this population.Methods: A cross-sectional survey of Vietnamese adult patients was conducted at 25 general practices with Vietnamese-speaking general practitioners (GPs) in south-western Sydney between October 2012 and February 2013. Patients completed the Kessler (K10) scale and a demographic questionnaire, available in Vietnamese or English. Data were analysed using SPSS version 21.Results: Of the 350 patients invited to participate, 247 completed surveys (response rate 71%). One-quarter (25%) of participants had a very high K10 score for psychological distress, nearly twice that reported in the NSW Health Survey. Participants with high exposure to trauma were at increased risk of psychological distress (odds ratio 5.9, 95% confidence interval 2.4-14.4; P < 0.0001) compared with those with mild or no trauma exposure. Similarly, risk was increased if there was a past history of mental health problems and a lack of personal and social support.Conclusion: The high prevalence of mental health problems in adult Vietnamese people attending Vietnamese-speaking general practices is associated with exposure to trauma. This highlights the importance of personal, social and professional support in effective management. Vietnamese-speaking GPs who see Vietnamese or similar refugee groups should actively seek out a history of exposure to trauma, a past history of mental illness and the existence of support systems.

Coronary artery disease risk assessment from unstructured electronic health records using text mining.

Coronary artery disease (CAD) often leads to myocardial infarction, which may be fatal. Risk factors can be used to predict CAD, which may subsequently lead to prevention or early intervention. Patient data such as co-morbidities, medication history, social history and family history are required to determine the risk factors for a disease. However, risk factor data are usually embedded in unstructured clinical narratives if the data is not collected specifically for risk assessment purposes. Clinical text mining can be used to extract data related to risk factors from unstructured clinical notes. This study presents methods to extract Framingham risk factors from unstructured electronic health records using clinical text mining and to calculate 10-year coronary artery disease risk scores in a cohort of diabetic patients. We developed a rule-based system to extract risk factors: age, gender, total cholesterol, HDL-C, blood pressure, diabetes history and smoking history. The results showed that the output from the text mining system was reliable, but there was a significant amount of missing data to calculate the Framingham risk score. A systematic approach for understanding missing data was followed by implementation of imputation strategies. An analysis of the 10-year Framingham risk scores for coronary artery disease in this cohort has shown that the majority of the diabetic patients are at moderate risk of CAD.