Urban Indian Health Institute的动态

Harnessing Data for Public Health Impact: Applied Medical Statistics & Public Health Specialist Do you thrive at the intersection of healthcare and data analysis? Are you passionate about using data to improve population health? What you'll do: ?? Design and analyze complex medical studies to uncover population health trends. ?? Translate statistical findings into actionable insights for public health professionals. ?? Collaborate with epidemiologists, healthcare providers, and policymakers to develop evidence-based public health interventions. ?? Monitor and evaluate the effectiveness of public health programs. Why it matters: ?? Your work will directly contribute to improving public health outcomes and reducing health disparities. ?? You'll be at the forefront of using data to solve some of the world's most pressing health challenges. ?? This is a growing field with excellent job prospects and competitive salaries. Skills you'll need: ?? Strong foundation in statistics, epidemiology, and research methods. ?? Proficiency in statistical software (e.g., SAS, R, Stata, SPSS). ?? Excellent communication and collaboration skills. ?? Ability to translate complex data into clear and concise reports. #publichealth #appliedstatistics #dataanalysis #healthcare ???? ??? ??????? ?????

Even seasoned researchers often mix up certain concepts. Here are 10 of the most commonly misused epidemiological or statistical terms: ???? 1?? Data vs. Datum ? Common Misuse: "The data was analyzed." ? Proper Usage: “Data” is plural, and “datum” is singular. So, it should be “The data were analyzed.” 2?? Univariate vs. Bivariate ? Common Misuse: People often think univariate analysis means a “crude” or unadjusted analysis. ? Proper Usage: Univariate refers to one variable (“uni-” as in “unilateral”). Bivariate analysis involves two variables (“bi” as in binary). ?? 3?? Confidence Intervals ? Common Misuse: All confidence intervals are treated the same. ? Proper Usage: Confidence intervals from descriptive statistics assess precision ?? but don’t convey statistical significance. Meanwhile, those in inferential statistics can indicate significance—e.g., non-inclusion of the null value suggests significance. 4?? Prevalence vs. Prevalence Rate ? Common Misuse: "The prevalence rate of diabetes was 50%." ? Proper Usage: Prevalence doesn’t involve time, so just say, “The prevalence of diabetes was 50%.” Let’s save “rate” for true rates, like incidence rates. ? 5?? Case-Control Studies ? Common Misuse: Any study with a control group is labeled as case-control. ? Proper Usage: It’s not just about having controls. Case-control studies start with the outcome and work backward to explore exposures. ?? 6?? Multivariate vs. Multivariable ? Common Misuse: Multivariate is often used when the term multivariable is what’s actually meant. ? Proper Usage: One variable, many predictors? It’s multivariable. Many outcomes? It’s multivariate! ???? 7?? Data Manipulation vs analysis ? Common Misuse: Using “data manipulation” interchangeably with data processing or analysis. ? Proper Usage: It’s better to use terms like "data processing," "data management," or "data transformation" to avoid the idea that we’re twisting the data to our will. 8?? Proportion vs. Percentage ? Common Misuse: Used as interchangeable entities. E.g., the proportion of smokers was 50 percent. ? Proper Usage: A proportion ranges from 0 to 1. A percentage ranges from 0 to 100. You can convert a proportion to a percentage by multiplying by 100. ?????? 9?? Causation vs. Association ? Common Misuse: People use synonyms of “cause” when referring to correlations (e.g., “impact”, “effect”) ? Proper Usage: Association is two-way; causality is one-way. ?? A is associated with B means B is associated with A, but A causing B doesn’t mean B causes A. ?? ?? Random Sampling vs. Simple Random Sampling ? Common Misuse: Treating these as identical. ? Proper Usage: Simple random sampling means equal selection chance. Random sampling is broader and means every participant has a non-zero chance, but probabilities don’t have to be equal. ?? By understanding these terms, we can communicate our science more effectively—and with fewer head-scratching moments! ?? Please reshare ?? #Chisquares #VillageSchool

Different Types of Sampling Methods in Statistics Applied to Data Science Here are five common types of sampling methods used in statistics and their applications in data science: 1. Simple Random Sampling - Description: Each member of the population has an equal chance of being selected. This method often uses random number generators or drawing lots. - Use Cases: A. Market Research: To gauge consumer preferences or satisfaction by randomly selecting a sample from a large customer base. B. Quality Control: In manufacturing, to randomly check the quality of products coming off the production line. 2. Stratified Sampling - Description: The population is divided into subgroups (strata) based on a specific characteristic, and random samples are taken from each stratum proportionally. - Use Cases: A. Healthcare Studies: Ensuring representation from different age groups, genders, or other demographics to get a more accurate picture of health trends. B. Educational Research: Assessing the performance of students across different school districts or grades. 3. Cluster Sampling - Description: The population is divided into clusters, usually based on geographical or natural groupings. Entire clusters are randomly selected, and all members of the chosen clusters are included in the sample. - Use Cases: A. Epidemiology: Studying the spread of diseases by selecting entire towns or regions. B. Public Opinion Polls: Surveying all individuals within randomly chosen neighborhoods or districts. 4. Systematic Sampling - Description: Selecting every k-th member of the population, where k is a constant determined by dividing the population size by the desired sample size. - Use Cases: A. Manufacturing: Checking every 10th item on a production line to ensure consistency in quality. B. Inventory Management: Conducting systematic stock checks in large warehouses. 5. Convenience Sampling - Description: Sampling members of the population who are easily accessible. This method is non-random and often used when quick, inexpensive data collection is needed. - Use Cases: A. Pilot Studies: Collecting preliminary data quickly to refine research hypotheses or survey questions. B. Social Media Research: Analyzing readily available data from users who interact with a specific brand or page. Industry Applications 1. Retail and Marketing: Random sampling and stratified sampling are commonly used to understand consumer preferences and market trends. For example, a retailer might use stratified sampling to ensure they get opinions from different age groups about a new product line. 2. Healthcare and Pharmaceuticals: Stratified sampling helps in ensuring diverse representation in clinical trials, while cluster sampling can be used in epidemiological studies to understand disease prevalence across different regions. 3. Manufacturing: Systematic sampling is often used in quality control processes to monitor product consistency and detect defects.

What happens when the data used in medical studies is inherently biased? How can we ensure that under-represented groups are accurately reflected in health data? These are important questions that the latest study by Lamin Juwara, Alaa El-Hussuna and Khaled El Emam addresses through the innovative use of synthetic data augmentation. https://lnkd.in/gGVSfspf In their research, the authors evaluate various methods for mitigating covariate bias in health data, with a focus on Synthetic Minority Augmentation (SMA). The findings are compelling: SMA not only enhances the performance and fairness of predictive models in low to medium bias settings but also offers a robust solution to ensure that health data accurately represents diverse populations. The implications of this study are significant for both researchers and patients. By synthesizing under-represented groups, SMA helps reconstruct full and unbiased data samples, leading to more accurate and fair health outcomes. For example, in the Canadian Community Health Survey data, SMA provided parameter estimates and fairness metrics that closely matched the ground truth, outperforming traditional methods like random oversampling and propensity score matching. This approach ensures that the insights drawn from health data are inclusive and equitable, ultimately benefiting patient care and medical decision-making. As the authors recommend, adopting SMA can be a practical step in improving the generalizability and reliability of health data analytics, especially in scenarios where recruitment of diverse participants is challenging. #HealthData #DataBias #SyntheticData #MedicalResearch #Analytics #HealthEquity #DataScience #MachineLearning #BiasMitigation #PredictiveModeling #HealthOutcomes #PatientCare #Epidemiology #DataAugmentation #HealthStudies #clinicalscience #AIinMedicine

?? Transforming Data with History: How the Narrative Historical Approach Can Drive Social Change I recently stumbled upon the narrative historical approach while revisiting an old epidemiology textbook, searching for a reference. What began as a quick reference check became a deep exploration into how this method could bring a unique perspective to my work in data, particularly in advancing data equity. Here, I’ll share an overview of the narrative historical approach, when it’s most valuable, and why it’s such a compelling tool. ?? What is the Narrative Historical Approach? The narrative historical approach is a method that emphasises storytelling to convey the sequence, significance, and impacts of events over time. Unlike purely quantitative methods, it delves into the contextual richness, providing a layered understanding of how historical factors shape present realities. This approach is often used in fields like history and sociology, but it has potential applications in health, especially when understanding complex social determinants of health. ?? When Should You Use the Narrative Historical Approach? This method shines when the goal is to illuminate how history has shaped present-day conditions. In data equity work, it can be particularly valuable for: ???? Tracing Inequities: Understanding the root causes of data inequities, such as how certain groups are systematically underrepresented or misrepresented in datasets. ????Contextualising Findings: Situating modern data within the broader historical context to uncover patterns that purely statistical approaches may overlook. ????Framing Solutions: By understanding the historical context, we can frame solutions that are more targeted and respectful of communities affected by inequities. ?? Why Use the Narrative Historical Approach? Narrative history provides more than just context; it humanises data. For data equity advocates, this approach can help highlight overlooked narratives and ensure that the data we present honors the lived experiences of those behind the numbers. By weaving in history, we also foster empathy and understanding, supporting policy changes that acknowledge systemic issues and build towards fairness in how data impacts lives. How are you using history in your work? Could a narrative approach to data help drive the impact you’re working toward? #dataequity #publichealthrebels

???? Our ‘Covid19 in MS global data-sharing initiative (GDSI)’, launched in 2020 to investigate the effects of COVID-19 on people with Multiple Sclerosis (MS), has achieved a remarkable (and last) milestone with the acceptance of our recent manuscript in the high-impact Journal 'Scientific Data' (Nature). Back in 2020, the GDSI quickly resulted in the largest federated data network on Covid10 in people with MS, aggregating data from >20 sources across over 80 countries, delivering crucial insights that have helped update global advice for people with MS. ??? A major triumph of our work was the ability to collate and analyze data from existing registries and cohorts, ensuring the long-term storage and governance of this invaluable data. With the closure of our central platform in 2022, a challenge arose: some data, directly shared by clinicians and individuals with MS, risked being lost forever. ???? In alignment with our core value of open science—"to be as open as possible and as closed as needed"—we took a decisive step to preserve and share this unique dataset. This action underscores our commitment to transparency and allows for ongoing research and educational activities, ensuring that the insights gained from this data continue to benefit the global MS community and beyond. ???? I would like to thank Hamza K. and the other co-authors Lotte Geys, Peer Baneke and Giancarlo Comi ?? Hamza, your expertise in real-world data handling and anonymization strategies played a pivotal role in the success of this paper titled "Patient level dataset to study the effect of COVID-19 in people with Multiple Sclerosis." Your dedication not only safeguarded this critical dataset but also exemplified the spirit of open science and collaboration. ????? This work ensures that the data-driven insights we've gathered will continue to empower research and inform global strategies for people with MS in the face of COVID-19. ???? Here's to the lasting impact of your work and to more achievements in the future! ???? Link to the paper: https://lnkd.in/eBsjvp9v ----------------------- ???? Who am I? ?? Data Scientist - Entrepreneur - Idealist ?? I envision a world where every single person gets the treatment they deserve in a timely matter. I are convinced that #DataSavesLives. Therefore, my research focuses on investigating new methods to handle and analyze #BigData in #Health & #Care. ?? I am an assistant professor in biomedical data sciences, affiliated with Biomedical Research Institute (BIOMED-UHasselt) and the Data Science Institute (DSI_UHasselt), within the Faculteit Geneeskunde en Levenswetenschappen UHasselt at UHasselt. ?? More details on our website: https://lnkd.in/eGWyx2Da

Start Asking Data Why | Causality Intro| Eyal Kazin Start Asking Your Data “Why?”?—?A Gentle Intro To Causality A beginner’s guide to thinking beyond correlations. “Causation is not merely an aspect of statistics?—?it is an addition to statistics”?—?Judea?Pearl Newton’s Cradle. Credit:?Pickpik. Correlation does not imply causation. It turns out, however, that with some simple ingenious tricks one can, potentially, unveil causal relationships within standard observational data, without having to resort to expensive randomised control?trials. This post is targeted towards anyone making data driven decisions. The main takeaway message is that causality may be possible by understanding that the story behind the data is as important as the data?itself. By introducing Simpson’s and Berkson’s Paradoxes, situations where the outcome of a population is in conflict with that of its cohorts, I shine a light on the importance of using causal reasoning to identify these paradoxes in data and avoid misinterpretation. Specifically I introduce causal graphs as a method to visualise the story behind the data point out that by adding this to your arsenal you are likely to conduct better analyses and experiments. My ultimate objective is to whet your appetite to explore more on causality, as I believe that by asking data “Why?” you will be able to go beyond correlation calculations and extract more insights, as well as avoid common misjudgement pitfalls. Note that throughout this gentle intro I do not use equations but demonstrate using accessible intuitive visuals. That said I provide resources for you to take your next step in adding causal inference to your statistical toolbox so that you may get more value from your?data. The Era of Data Driven Decision?Making “In [Deity] We Trust, All Others Bring Data!”?—?William E.?Deming In this digital age it is common to put a lot of faith in data. But this raises an overlooked question?— Should we trust data on its?own? Judea Pearl, who is considered the Godfather of causality, articulated best: “The collection of information is as important as the information itself “?—?Judea?Pearl In other words the story behind the data is as important as the data?itself. This manifests in a growing awareness of the importance of identifying bias in datasets. By the end of this post I hope that you will appreciate that causality pertains the fundamental tools to best express, quantify and attempt to correct for these?biases. In causality introductions it is customary to demonstrate why “correlation does not imply causation” by highlighting limitations of association analysis due to spurious correlations (e.g, shark attacks ?? and ice-cream sales ??). In an attempt to reduce the length of this post I defer this aspect to an older one of mine. Here I focus on two mind boggling paradoxes ?? and their resolution via causal graphs to make a similar?point. Paradoxes in?Analysis To understand the importance of the...

?? Proudly Announcing the Completion of Omdena Kitwe, Zambia Chapter - COVID-19 Predictive Model for Kitwe! ?? I am thrilled to share the success of our latest project at the Omdena Zambia Chapter: a cutting-edge machine learning model developed to predict COVID-19 dynamics in Kitwe, Zambia. This achievement is a pivotal step forward in our fight against the pandemic and our commitment to protecting our community's health. Project Overview: In the battle against COVID-19, accurate forecasting is key to making informed decisions, managing resources, and implementing effective public health measures. Our predictive model uses historical data on COVID-19 cases, recoveries, deaths, testing rates, and other essential factors to offer valuable insights into the pandemic's course in Kitwe. Key Highlights: Data-Driven Insights: Leveraging data from reliable sources, our model provides precise predictions of COVID-19 trends, aiding strategic decision-making. Community Benefits: The model's insights enable local authorities to allocate healthcare resources more effectively, implement timely interventions, and raise public awareness. Collaborative Innovation: This project exemplifies the power of teamwork and innovation in solving real-world challenges. Personal Growth and Learning: Working on this project has been an incredible journey of learning and growth in the field of data science. I have gained hands-on experience in data collection, preprocessing, and the development of machine learning models. Collaborating with a talented team has not only enhanced my technical skills but also broadened my understanding of how data science can be applied to address pressing public health issues. Acknowledgements: A project of this scale requires dedication and expertise. I would like to extend my heartfelt thanks to the following team members for their exceptional contributions: Samrawit Mewa: For leading data collection and preprocessing, ensuring our dataset's accuracy and completeness. NITESH KESHARWANI & Denis Surnin: For their outstanding work in developing and fine-tuning the machine learning model. Sawsan Abdulbari & Arpit Sengar: For designing and deploying the user-friendly web application that makes our model accessible. Maira Shiki, Priyanka N & Antara T.: For their invaluable domain expertise and guidance on public health implications. David Hartsman & Amal Altalhi: For facilitating communication between the project team and local stakeholders, ensuring the project's relevance and impact. I am incredibly proud of what we have achieved together and am excited about the positive impact this project will have on the Kitwe community. This success highlights the power of collaboration, innovation, and a shared commitment to making a difference. #Omdena #MachineLearning #COVID19 #PublicHealth #Zambia #TeamWork #Innovation #CommunityImpact #DataScience #PersonalGrowth

Career Paths in Public health *Biostatistician* Biostatistician applies statistical techniques to analyze data in fields like medicine, public health and other similar fields.. #OLAJUMOKE #askmeanythingaboutpublichealth #yourhealthconsultant #healthforall

?? Excited to share my recent SQL-based project on analyzing global data for Trachoma, a Neglected Tropical Disease (NTD) affecting millions and causing preventable blindness. Using data from Our World in Data, I explored key trends in at-risk populations, treatments, and surgical interventions across multiple countries over time. Through exploratory data analysis (EDA), I discovered that 5.27 billion people are at risk for trachoma worldwide, with the highest burden in Africa. Despite a decline in cases, my analysis highlighted a 2018 spike, emphasizing the need for continued global investment in public health. This project reflects my passion for patient data management and analysis, driving actionable insights to improve health outcomes. I'm excited to continue leveraging data to create a positive impact in global health! To see complete project flow, please click below: https://lnkd.in/g2jj3EQk #HealthcareData #NTDs #PublicHealth #DataAnalysis #Trachoma #SQL #GlobalHealth