Analysis Of Health Data Research Articles

Wait a Minute? Uniting Physical Exercise and Data Science in Higher Learning

Purpose: This study explores the design and implementation of a novel Data Science course that integrates physical activity and physiological data collection. The course, scheduled twice weekly, combines a 30-minute physical exercise session with data collection activities focusing on metrics such as temperature, three separate heart rates, blood pressure, Functional Movement Screening (FMS), nutritional assessment using MyPlate.gov categories, urinalysis, and other metrics. The primary objective was to provide students with practical experience in data science by engaging them in real-time physiological data collection and analysis. Methods: This study involved a structured course format where students engaged in physical activities and recorded physiological data using standardized measurement tools and techniques. The course structure included instructional sessions on data science principles, data collection procedures, and statistical analysis. Students then applied their knowledge by undertaking a project that involved analyzing the collected data to address specific research questions or hypotheses related to physical health and performance. Each student presented their findings through a PowerPoint presentation, fostering peer review and collaborative learning. Results: Findings indicated that students successfully developed data science skills while gaining insights into the relationship between physical activity and physiological metrics. The projects revealed varied patterns and correlations, demonstrating the practical applicability of data science in health and fitness contexts. Discussion/Conclusions: The study highlights the effectiveness of integrating physical activity with data science education, enhancing both engagement and learning outcomes. The course equips students with technical skills while emphasizing the importance of holistic health data analysis.

A collaborative approach to advancing research and training in Public Health Data Science—challenges, opportunities, and lessons learnt

The unprecedented availability of increasingly complex, voluminous, and multi-dimensional data as well as the emergence of data science as an evolving field provide ideal opportunities to address the multi-faceted public health challenges faced by low and middle income countries (LMIC), especially those in sub-Saharan Africa. However, there is a severe lack of well-trained data scientists and home-grown educational programs to enable context-specific training. The lack of human capacity and resources for public health data analysis as well as the dire need to use modern technology for better understanding and possible intervention cannot be dealt with currently available educational programs and computing infrastructure, demanding a great deal of collaboration and investments within Africa and with the Global North This paper describes processes undertaken to establish sustainable research training programs and to train a new generation of data scientists with knowledge, mentoring, professional skills, and research immersion. The goal is to position them for rigorous, biomedically grounded and ethically conscious Public Health Data Science practice with a focus on Ethiopia and Kenya. The programs are realized through partnership among Columbia University (CU, USA), Addis Ababa University (AAU, Ethiopia), and the University of Nairobi (UoN, Kenya). In this paper, we describe the collaborative project named “Advancing Public Health Research in Eastern Africa through Data Science Training (APHREA-DST)” delving on its conceptualization, implementation framework and activities undertaken. We adopted both qualitative and quantitative approaches to understand the needs of the stakeholders for such educational and training programs. Through harmonized online surveys and stakeholder engagements via focus group discussions in Ethiopia and Kenya, a curriculum was developed for a masters degree program in Public Health Data Science (PHDS). Moreover, the engagement with local projects in both countries as well as active collaboration with other data science related projects in Africa under DSI-Africa consortium benefited the project to start the M. Sc. program successfully. So far, the launching of the graduate program in both countries and the two-cycle experience sharing program done at Columbia University as well as the numerous MoUs signed between partners for data sharing and internships are the major successes of the project. In this paper, we discuss in detail the challenges faced as well as the existing opportunities and lessons learnt this far in implementing this tripartite collaborative teaching and research project.

Understanding the demographic and socioeconomic determinants of morbidity in Eastern Uganda: a retrospective analysis of the Iganga-Mayuge health and demographic surveillance data

IntroductionUnderstanding the determinants of disease burden is imperative in enhancing population health outcomes. This study uses data from the Iganga-Mayuge Health and Demographic Surveillance Site, to understand demographic and socioeconomic...

Performance Assessment of Some Count Data Models to Immunization Coverage Data

This research evaluates the performance of various count data models, including Poisson Regression (PR), Zero-Inflated Poisson Regression (ZIP), Zero-Truncated Poisson Regression (ZTP), Truncated Negative Binomial Poisson Regression (TNBP), and Negative Binomial Poisson Regression (NBP), using immunization coverage data from the National Primary Health Care Development Agency (NPHCDA). The study focuses on children under 12 months, assessing model fit using Likelihood Ratio (LR), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC) criteria. Analysis conducted with STATA indicates that the Truncated Negative Binomial Poisson Regression (TNBP) outperformed other models in fit and efficiency. Both the ZTeeP and TNBP models demonstrated the best fit, with lower AIC (1959.107) and BIC (2037.649) values and higher Pseudo R-squared values (0.0677 for ZTP and 0.0590 for TNBP), compared to standard models. Age was identified as a significant predictor, negatively associated with immunization status, implying that older infants in the under-12-month category are less likely to receive all vaccinations. The ZTP model showed significant positive effects for antigens such as HepB0, OPV0, BCG, and Measles, with age having a significant negative association. The findings highlight the importance of selecting appropriate statistical models for accurate public health data analysis, enhancing decision-making in immunization programs.

LEGAL REVIEW OF CHILD NUTRITION HEALTH REGARDING THE CHALLENGES OF STUNTING IMPLICATIONS IN CHILDREN WITH SPECIAL NEEDS IN RURAL AREAS THROUGH THE UTILIZATION OF ARTIFICIAL INTELLIGENCE

Children with special needs are vulnerable in rural areas to nutrition problems, including stunting, due to limited access to health services and minimal nutrition knowledge. Article 11 of Law No. 17 of 2023 on Health emphasizes the responsibility of the Central and Local Governments to provide access to healthcare facilities and health education. In this case, the utilization of artificial intelligence (AI) can help monitor, diagnose, and intervene more effectively against stunting through accurate data analysis and timely recommendations. The objective of this study is to analyse the legal challenges related to child nutritional health, especially stunting in children with special needs in rural areas, and explore the potential of utilizing AI in addressing this issue. The research method used is normative juridical, using a statutory approach and an analytical approach. The results show that children with special needs in rural areas face various challenges in accessing health services and nutritional information. Limited access, lack of community knowledge about the importance of nutrition, and lack of effective intervention programs are the main obstacles. On the other hand, the use of AI can help in early detection and treatment of stunting through health data analysis, and provide appropriate intervention recommendations. The use of artificial intelligence has great potential in improving the effectiveness of child nutrition health programs and reducing stunting cases among children with special needs in rural areas. Collaboration between the government, health institutions, and technology is needed to implement this solution effectively and sustainably.

Theoretical perspectives on biostatistics and its multifaceted applications in global health studies

Biostatistics plays a crucial role in global health studies by providing the tools necessary for analyzing complex health data and informing evidence-based public health decisions. This theoretical review explores the foundations of biostatistics and its diverse applications in the context of global health. At its core, biostatistics utilizes mathematical and statistical principles to analyze health data, enabling researchers to identify patterns, determine risk factors, assess interventions, and predict health outcomes across populations. Key theoretical concepts, including probability theory, statistical inference, and regression models, form the backbone of biostatistical analysis and are integral to understanding the dynamic relationships between health determinants and outcomes. The review highlights the various applications of biostatistics in global health, focusing on areas such as disease surveillance, epidemiology, and public health intervention evaluation. Biostatistical techniques are essential for analyzing epidemiological data, monitoring disease trends, and evaluating the effectiveness of health policies and interventions on a global scale. Furthermore, advanced statistical methods, including machine learning and Bayesian approaches, are becoming increasingly important in global health research. These methods allow for the analysis of large-scale data, predictive modeling, and more precise risk assessments, particularly in the context of emerging diseases, genetic research, and environmental health studies. Despite its strengths, biostatistics faces challenges in global health, including issues related to data quality, ethical considerations, and cross-cultural variations in health outcomes. Addressing these challenges is critical for ensuring accurate and equitable health assessments. Looking forward, biostatistics continues to evolve with innovations in data science, big data analytics, and real-time health data collection, making it an indispensable tool for improving global health outcomes and supporting health policies. This review emphasizes the theoretical underpinnings of biostatistics while highlighting its transformative potential in global health research and practice. Keywords: Theoretical Perspectives, Biostatistics, Multifaceted Applications, Global Health Studies.

Machine Learning in Health Anylitics and Patient Monitoring Solomon Kavuta and Mr Joel Mulepa Dmi St John The Baptist University, Lilongwe, Malawi

The integration of machine learning techniques in healthcare has revolutionized the way predictive analytics and patient monitoring are conducted. This transformation is reshaping traditional healthcare practices by harnessing the power of advanced computational methods to analyze vast amounts of patient data. Machine learning algorithms, which excel at detecting patterns and making predictions, leverage this data to forecast disease onset, progression, and treatment outcomes. This capability enables healthcare providers to adopt proactive measures, facilitating early intervention and the development of personalized treatment plans tailored to individual patients' needs. The potential of machine learning in healthcare is vast and multifaceted. By continuously analyzing real-time patient data streams from various sources such as electronic health records (EHRs), medical imaging, genomic information, and data from wearable devices, these algorithms can identify subtle changes in a patient's condition that may indicate the early stages of a disease or the risk of an adverse event. This continuous monitoring allows for timely interventions that can prevent complications, reduce the severity of diseases, and improve overall patient outcomes. For example, machine learning can help predict the likelihood of a patient developing conditions like diabetes or heart disease, allowing for lifestyle changes and medical interventions that can mitigate these risks before they become severe health issues. Moreover, machine learning facilitates remote patient monitoring, which has become increasingly important in the context of modern healthcare. With the advent of wearable devices and Internet of Things (IoT) sensors, it is now possible to collect and analyze health data outside of traditional clinical settings. These technologies enable healthcare providers to monitor patients in real-time, regardless of their location, ensuring continuous care and oversight. This is particularly beneficial for managing chronic diseases, where regular monitoring can significantly improve patient management and outcomes. For instance, remote monitoring of patients with heart conditions can detect early signs of arrhythmias or other issues, prompting immediate medical responses that can prevent more serious complications. The transformative impact of machine learning on healthcare delivery extends beyond patient monitoring and predictive analytics. It also includes enhancing diagnostic accuracy, optimizing treatment protocols, and streamlining administrative processes. By reducing the need for frequent hospital visits and enabling more efficient use of healthcare resources, machine learning can help lower healthcare costs while improving the quality of care provided to patients. Additionally, the ability to analyze large datasets quickly and accurately can aid in medical research, uncovering new insights into disease mechanisms and potential treatments.

Investigating Cancer Inequalities in Urbanizing Texas with Plausible Reasoning

This article contributes to geographical methodology by presenting an epistemology of plausible reasoning (PR) as a complement to realist and reflexive frameworks for science. I introduce PR’s theory of evidence, after Harold Jeffreys and George Pólya, and then extrapolate principles for nonexperimental study designs. This article leverages these principles to structure an investigation into the causes of a racial disparity in the colorectal cancer (CRC) burden in the Houston and Dallas-Fort Worth metropolitan areas. Centered on a series of empirical questions pertaining to racial ghettoization, contemporary urbanization, and cancer inequalities, the study uncovers important geographic and social changes in the CRC burden in the study areas between 1999 and 2019. These include the emergence of a substantial class-related inequality in CRC incidence, the rise of aggregate racial disparities in CRC incidence outside of historically ghettoized areas, and an overall intensification of spatial CRC inequality. Drawing on fundamental cause theory and critical urban studies, I propose that these changes are a product of the uneven spread of new preventive technologies (screening colonoscopy) amid extensive and polarizing urbanization processes that are reshaping these regions. The study highlights the value of incorporating concern for dynamic macrostructures and political economy into analyses of spatial health data, and illustrates how PR can contribute to such work.

Multiple feature selection based on an optimization strategy for causal analysis of health data

PurposeRecent advancements in information technology and wearable devices have revolutionized healthcare through health data analysis. Identifying significant relationships in complex health data enhances healthcare and public health strategies. In health analytics, causal graphs are important for investigating the relationships among health features. However, they face challenges owing to the large number of features, complexity, and computational demands. Feature selection methods are useful for addressing these challenges. In this paper, we present a framework for multiple feature selection based on an optimization strategy for causal analysis of health data.MethodsWe select multiple health features based on an optimization strategy. First, we define a Weighted Total Score (WTS) index to assess the feature importance after the combination of different feature selection methods. To explore an optimal set of weights for each method, we design a multiple feature selection algorithm integrated with the greedy algorithm. The features are then ranked according to their WTS, enabling selection of the most important ones. After that, causal graphs are constructed based on the selected features, and the statistical significance of the paths is assessed. Furthermore, evaluation experiments are conducted on an experiment dataset collected for this study and an open dataset for diabetes.ResultsThe results demonstrate that our approach outperforms baseline models by reducing the number of features while improving model performance. Moreover, the statistical significance of the relationships between features uncovered through causal graphs is validated for both datasets.ConclusionBy using the proposed framework for multiple feature selection based on an optimization strategy for causal analysis, the number of features is reduced and the causal relationships are uncovered and validated.

Abstract 4123308: Trends in Gestational Diabetes Mellitus By US State, 2019 -2023

Introduction/ Background: Gestational diabetes mellitus (GDM) is one of the most frequent adverse pregnancy outcomes and increases the risk of lifetime cardiometabolic disease. Given known geographic disparities in maternal morbidity and mortality in the US, we examined state-level prevalence and trends in GDM from 2019-2023. Methods: We conducted a serial, cross-sectional analysis of maternal health data recorded on birth certificates from all livebirths in the US using the National Center for Health Statistics Natality Files. We included pregnant individuals aged 15-44 years who gave birth from 2019-2023, had a singleton, live birth, and did not have pre-pregnancy diabetes. We calculated the age-standardized prevalence and average-annual percent change (AAPC) in GDM from 2019-2023, overall, by U.S. census region (Northeast, Midwest, South, and West), and in each US state and the District of Columbia (DC). Results: Of the 17,432,486 individuals with live births between 2019 and 2023, overall prevalence of GDM per 100 live births (95% confidence interval) was 6.9 (6.9, 6.9) in 2019 and 8.0 (8.0, 8.1) in 2023 (AAPC 3.2 [-1.7, 8.5]). There was significant geographic variation in overall prevalence of GDM by US region and state ( Figure 1 ). In 2023, prevalence was higher in the Midwest (8.8 [8.7, 8.9]) and Western states (8.6, [8.5,8.6]) compared with the Southern (7.3, [7.3, 7.4]) and Northeastern states (8.1, [8.0,8.2]), ranging from 5.5 (4.9,6.0) in DC to 13.0 (12.2,13.8) in Alaska. Prevalence was higher in 2023 compared with 2019 in all US regions and states and DC, except Alaska, Connecticut, Idaho, Maine, New Jersey, and Wyoming. Prevalence increased most rapidly in the Western US states (AAPC 4.2%/year [0.1, 8.4]) with the greatest change observed in Montana from 5.1 (4.7, 5.6) in 2019 to 8.0 (7.5, 8.5) in 2023 (AAPC 10.3%/year [5.6, 15.3]). Conclusions: GDM prevalence increased significantly from 2019-2023 with substantial heterogeneity by US region and state. More localized state-level analysis with context-specific polices that address factors associated with geographic differences in GDM are needed to promote maternal cardiometabolic health equity in the US.

Are EU member states ready for the European Health Data Space? Lessons learnt on the secondary use of health data from the TEHDAS Joint Action.

The proposal for a regulation on the European Health Data Space (EHDS) contains provisions that would significantly change health data management systems in European member states (MS). This article presents results of a country mapping exercise conducted during the Joint Action 'Towards the European Health Data Space' (TEHDAS) in 2022. It presents the state-of-play of health data management systems in 12 MS and their preparedness to comply with the EHDS provisions. The country mapping exercise consisted of virtual or face-to-face semi-structured interviews to a selection of key stakeholders of the health information systems. A semi-quantitative analysis of the reports was conducted and is presented here, focusing on key aspects related to the user journey through the EHDS. This article reveals a heterogenous picture in countries' readiness to comply with the EHDS provisions. There is a need to improve digitalization and quality of health data at source across most countries. Less than half of the countries visited have or are developing a national datasets catalogue. Although the process to access health data varies, researchers can analyse health data in secure processing environments in all countries visited. Most of the countries use a unique personal identifier for health to facilitate data linkage. The study concluded that the current landscape is heterogeneous, and no member state is fully ready yet to comply with the future regulation. However, there is general political will and ongoing efforts to align health data management systems with the provisions in the EHDS legislative proposal.

Association of METS-IR index with Type 2 Diabetes: A cross-sectional analysis of national health and nutrition examination survey data from 2009 to 2018.

With rising global diabetes prevalence, precise early identification and management of diabetes risk are critical research areas. The metabolic score for insulin resistance (METS-IR), a novel non-insulin-based tool, is gaining attention for quantifying insulin resistance using multiple metabolic parameters. Despite its potential in predicting diabetes and its precursors, evidence on its specific relationship with diabetes is limited, especially in large-scale population validation and mechanistic exploration. This study aims to analyze the association between METS-IR and type 2 diabetes (T2DM) in American adults. We conducted a cross-sectional analysis of the National Health and Nutrition Examination Survey (NHANES) data from 2009 to 2018. Participants aged 20 years and above were included, excluding individuals with missing data on BMI, fasting blood glucose, high-density lipoprotein cholesterol glycated hemoglobin and diabetes status. Logistic regression analysis, subgroup analysis, and restricted cubic spline analysis were used to assess the association between METS-IR and T2DM, controlling for potential confounding factors. After adjusting for age, gender, race, education level, smoking status, drinking habits, depression, physical activity, hypertension, and hyperlipidemia, we found a positive association between METS-IR and the risk of T2DM. Specifically, each unit increase in METS-IR was associated with a 7% increase in the risk of T2DM (OR = 1.07, 95% CI: 1.06, 1.08). Subgroup analysis showed that the association between METS-IR and T2DM incidence was significantly positive in the highest quartile group, particularly among Mexican Americans over 40 years old and those diagnosed with depression, hypertension, or hyperlipidemia. Our study revealed a significant positive association between METS-IR and the prevalence of T2DM, indicating that this relationship persists even after controlling for various confounding factors. Therefore, monitoring METS-IR may provide a valuable tool for the early identification of individuals at risk of glucose metabolism disorders. Further research should focus on the applicability of METS-IR in different populations and its potential impact on clinical practice.

Revealed and stated preferences for sharing of patient-reported data in Austria

Abstract Objectives Routine collection of health data is essential for creating sustainable, efficient, and equitable health systems. While administrative health data are collected automatically, successful routine collection of patient-reported outcome measures (PROMs) and patient-reported experience measures (PREMs) depends on voluntary participation. The aim of this study is to investigate how to increase participation in routine patient-reported data collection. Methods This mixed methods study used representative survey data from the Austrian population aged 18-70 (N = 1260) and 15 semi-structured narrative interviews with citizens and health data experts collected in 2024. Survey respondents were shown examples of PREMs and PROMs and indicated their willingness to share them, depending on the recipient and purpose. Interviewees answered the same questions while providing insight into their decision-making process, which was analyzed using thematic analysis. Wilcoxon rank sum tests were used to assess differences in willingness to share across recipients and data collection purposes. Results Survey respondents were more willing to share PREMs than PROMs (“rather yes” vs. “rather no”, p < 0.001). One explanation given in the interviews is that PROMs are more personal and convey sensitive information. In terms of sharing PROMs, people were most likely to share with their GP (81%) and least likely to share with private companies (7%). Facilitators for sharing identified in the interviews were improving personal care and the health system, while barriers were use for profit generation and fear of negative consequences for the person sharing. Conclusions The results suggest handling the collection of PROMs demands greater caution compared to PREMs. Furthermore, clearly communicating personal and societal benefits arising from the collection and analysis of self-reported health data may increase participation in routine self-reported health data collection. Key messages • The collection of self-reported health data depends on voluntary participation. Clearly communicating personal and societal benefits of sharing self-reported data can motivate participation. • Self-reported data collection is essential for sustainable and efficient health systems. Compared to PREMs, people are more reluctant to share PROMs because they contain sensitive information.

Leveraging digital platforms to improve Public Health in hospital smart buildings

Abstract The integration of digital platforms into hospital infrastructure represents a significant leap forward in public health management. This session explores how smart hospital buildings, equipped with advanced digital technologies, can enhance public health outcomes through more effective disease surveillance, improved environmental controls and enhanced patient care management. Smart hospital buildings utilize a range of new technologies such as IoT sensors, AI-driven analytics and big data solutions to monitor and optimize various aspects of hospital operations and care. These technologies enable real-time data collection on environmental conditions, patient health metrics and resource utilization, creating an ecosystem that not only supports the immediate health needs of patients but also contributes to broader public health goals. The session will address four key areas where digital platforms can make an impact: -Disease Surveillance and Management: Digital tools can track health data trends and patterns within hospital settings, enabling early detection of infectious outbreaks and more precise tracking of patient outcomes. -Environmental Monitoring: IoT sensors can continuously assess conditions such as air quality, temperature and humidity, which are critical for preventing hospital-acquired infections and ensuring a safe environment for patients and staff. -Resource Optimization: AI algorithms can analyse data to optimize resource allocation, ensuring that medical supplies and human resources are used efficiently, which is crucial in health emergencies. -Patient-Centred Care: Digital platforms can facilitate personalized patient care through real-time health monitoring and data analysis, ensuring timely medical interventions and better health management. This session will highlight case studies demonstrating successful implementation of digital technologies in smart hospital buildings and discuss the challenges and lessons learned from these initiatives.

How Avoidable Mortality is influenced by health expenditure: an Italian 14-year panel data analysis

Abstract Background Avoidable Mortality is a widely used public health indicator to assess the quality of healthcare systems. The aim of this study was to investigate how the allocation of private and public health expenditure (PHE) could influence avoidable mortality (AM), treatable mortality (TM) and preventable mortality (PM) of Italian citizens. Methods A pooled cross-sectional time series study was conducted. Italian healthcare data from 2009 to 2019 was extracted from ‘Health for All’, a WHO-related database released periodically by the Italian National Institute of Statistics. PHE was divided by providers area (directly provided services (DPS), general and specialist medicine, pharmaceutical, rehabilitation). Data about private health expenditure, gross domestic product per capita (GDPPC), level of education, medical drug consumption per capita (MDCPC), general practitioners per capita (GPPC), physical exercise, tobacco use, and obesity were collected. AM was divided into TM and PM using OECD/Eurostat lists of causes of death (2022). A fixed-effects regression was used to assess the effects of economic and social variables on AM, TM and PM. Results Higher spending on DPS (-0.027), higher MDCPC (-0.859) and higher GDPPC (-0.001) were associated with lower rates of AM. Higher sport activity (c -0.673) was associated with lower rates of TM. Higher levels of private health expenditure (0.022) and higher number of GPPC (0.023) were associated with higher rates of PM; these counterintuitive results could be caused by the inappropriateness of medical treatment directly requested by patients. Other variables were not significantly associated with AM, TM or PM. Conclusions The results highlight the importance of medical DPS by NHS (emergency departments, hospital care, specialized care, primary care, etc.). Analyzing health data divided by areas of NHS we could help policy makers to allocate human and economic resources more efficiently. Key messages • The allocation of public health resources could influence the health status of the polulazion. • Italian National Health System could reduce avoidable mortality through the use of directly provided services.

Community Paramedicine Program in Social Housing and Health Service Utilization

Community Paramedicine at Clinic (CP@clinic) is a chronic disease prevention program that decreases 911 calls for emergency medical services, but its wider system effects are unknown. To evaluate the effects of CP@clinic vs usual care on individual-level health service utilization outcomes. This open-label, pragmatic cluster randomized clinical trial evaluated all residents 55 years or older in 30 social housing buildings in Ontario, Canada, that had (1) a unique postal code, (2) at least 50 apartments, (3) 60% or more residents 55 years or older, and (4) a similar building for pairing (15 intervention and 15 control buildings, pair-matched randomization). The 12-month intervention had a staggered start date from January 1, 2015, to December 1, 2015, and ended between December 31, 2015, and November 30, 2016. Administrative health data analysis was conducted in May 2022. CP@clinic was a health promotion and disease prevention program led by specially trained community paramedics who held weekly drop-in sessions in social housing buildings. These paramedics conducted 1-on-1 risk assessments, provided health education and referrals to relevant community resources, and, with consent, sent assessments to family physicians. Control buildings received usual care (universal health care, including free primary and specialty medical care). Individual-level health service utilization was measured from administrative health data, with ED visits via ambulance as the primary outcome; secondary outcomes included ED visits for any reason, primary care visits, hospitalizations, length of hospital stay, laboratory tests, receipt of home care, transfer to long-term care, and medication initiation. Generalized estimating equations were used to estimate intervention effects on individual-level health service utilization, accounting for trial design and individual-level baselines. The 30 social housing buildings had 3695 residents (1846 control and 1849 intervention participants; mean [SD] age, 72.8 [9.1] years; 2400 [65.0%] female). Intention-to-treat analysis found no significant difference in ED visits by ambulance (445 of 1849 [24.1%] vs 463 of 1846 [25.1%]; adjusted odds ratio [AOR], 0.97; 95% CI, 0.89-1.05) but found higher antihypertensive medication initiation (74 of 500 [14.8%] vs 47 of 552 [8.5%]; AOR, 1.74; 95% CI, 1.19-2.53) and lower anticoagulant initiation (48 of 1481 [3.2%] vs 69 of 1442 [4.8%]; AOR, 0.68; 95% CI, 0.53-0.86) in the intervention arm vs the control arm. CP@clinic attendance was associated with higher incidence of primary care visits (adjusted incidence rate ratio, 1.10; 95% CI, 1.03-1.17), higher odds of receiving home care (AOR, 1.07; 95% CI, 1.01-1.13), and lower odds of long-term care transfers (AOR, 0.32; 95% CI, 0.13-0.81). In this cluster randomized clinical trial of CP@clinic, the intervention did not affect the rate of ED visits by ambulance; however, there were increased primary care visits and connections to home care services, which may have increased antihypertensive medication initiation and reduced long-term care transfers from social housing. Health policymakers should consider CP@clinic's impact as an upstream approach to improve care for older adults with low income. ClinicalTrials.gov Identifier: NCT02152891.

A Welfare Test for Sharing Health Data

AbstractMachine learning and artificial intelligence methods are being increasingly used in the analysis of patient level and other health data, resulting in new insights and significant societal benefits. However, researchers are often denied access to sensitive health data due to concerns on privacy and security breaches. A key reason for reluctance in data sharing by data custodians is the lack of a specific legislative or other methodological framework in balancing societal benefits against costs from potential privacy breaches. Consistent with traditional cost–benefit analysis and the use of Quality Adjusted Life Years (QALYs), this study proposes an economic test that would enable Research Ethics Boards (REBs) to assess the benefits versus risks of data sharing. Our test can also be used to assess whether REBs employ reasonable effort to protect individual privacy and avoid tort liability. Constructing an acceptable welfare test for health data sharing will also lead to a better understanding of the value of data.

Receipt of social services intervention in childhood, educational attainment and emergency hospital admissions: longitudinal analyses of national administrative health, social care, and education data in Wales, UK

BackgroundResearch consistently finds poorer health and educational outcomes for children who have experienced out-of-home care relative to the general population. Few studies have explored differences between those in care and those in receipt of intervention from social services but not in care. Children receiving social services interventions often experience Adverse Childhood Experiences (ACEs), and deprivation, which are known to negatively impact outcomes. We aimed to estimate the association of different social services interventions with educational outcomes and hospital admissions, while adjusting for ACEs and deprivation.MethodsWe linked retrospective, routinely collected administrative records from health, education, and social care to create a cohort via the Secure Anonymised Information Linkage (SAIL) databank in Wales, UK. We analysed data for children and household members (N = 30,439) across four different groups: [1] no social care intervention; [2] children in need but not in care (CIN); [3] children on the Child Protection Register but not in care (CPR); [4] children in care - i.e. removed from the family home and looked after by the local authority (CLA). Our primary outcome was education outcomes at age 16 years. Secondary outcomes were all cause emergency hospital admissions, and emergency admissions for external causes/injuries.ResultsChildren in receipt of social services intervention were more likely to not attain the expected level upon leaving statutory education at age 16 after adjusting for ACEs and other characteristics (for children who had been in out-of-home care (conditional OR: 1·76, (95%CI) 1·25 − 2·48), in need (2·51, 2·00–3·15) and those at risk (i.e., on the child protection register) (4·04, 2·44 − 6·68). For all-cause emergency admissions, all social care groups were at greater risk compared to children in the general population (children in care (conditional HR: 1·31, 1·01–1·68), children in need (1·62, 1·38 − 1·90), and children at risk (1·51, 1·11 − 2·04).ConclusionsAll groups receiving social service intervention experience poorer educational and health outcomes than peers in the general population. Children who remain with their home parents or caregivers but are identified as ‘in need’ or ‘at risk’ by social care practitioners require further research. Integrated support is needed from multiple sectors, including health, educational and social care.

Discrimination, Unfair Treatment by Police, and Intimate Partner Violence Perpetration Among Black and Latino Men

ABSTRACT Discrimination and unfair treatment by police are salient racial injustices for Black and Latino men. General Strain Theory and the Theory of African American Offending point to the potential of these leading to outcomes like intimate partner violence perpetration (IPV). Drawing on theoretical predictions and analyses of the Add Health data, I found evidence that discrimination and unfair treatment by police had independent effects on IPV perpetration for Black and Latino men. The effect of racial injustices varied for Latino men such that both types were strong contributors to IPV perpetration for immigrant Latino men but not US born Latino men.

Comparative Analysis of K-Means and Divisive Clustering Techniques on Balanced Mental Health Data

Effective Clustering of mental health data can provide significant insights into patterns and relationships that are critical for understanding mental health conditions. This study investigated various clustering techniques applied to balanced mental health data to avoid biases associated with an imbalanced data. Clustering of the balanced mental health data was done with respect to the area of Residence feature. Firstly, Random undersampling and SMOTE techniques were incorporated to the imbalanced data set as balancing techniques so as to improve model performance. Random Undersampling Technique turned out to be the most ideal balancing technique with its accuracy, recall, precision and F-score values as 1. After balancing the data, two clustering techniques were applied to the Random Undersampled balanced data. The two techniques were namely: K-means and Divisive techniques. In order to select which of the two clustering techniques is ideal, two test statistics namely Internal Validation and Stability Validation were applied. Results showed that K-means clustering technique indicated slightly lower Average Propotion of None-overlap, Average Distance between Means and Figure Of Merit values given as 0.12, 0.41 and 0.9972 as compared to Divisive clustering technique which were 0.14, 0.42 and 0.9999. The conclusion was that K-means clustering has a better performance. This study's findings will help guide future researchers dealing with mental health data analysis on ways to improve model performance for better and more reliable predictions.