Applied Analytics for Business – Healthcare PDF

Applied Analytics for Business – Healthcare Dr. Ravi Shankar Prof – Data Science (AI -ML) Dean – Enterprise Solutions & Academic Collaborations https://www.linkedin.com/in/ravi-shankar-70584b1/ [email protected] Fourth Session AUG 30 , 2023 • Healthcare Data Governance Faculty Profile : Dr. R...

Applied Analytics for Business – Healthcare Dr. Ravi Shankar Prof – Data Science (AI -ML) Dean – Enterprise Solutions & Academic Collaborations https://www.linkedin.com/in/ravi-shankar-70584b1/ [email protected] Fourth Session AUG 30 , 2023 • Healthcare Data Governance Faculty Profile : Dr. Ravi Shankar INTRODUCTION • Background: PhD - Econometrics, Entrepreneur, Start-up Mentor, Social Impact Investor, Data Science Thought Leader, Senior Venture Partner • Current Focus: AI adoption in Business, Deep Learning, XAI, ML Operations, Design Thinking, OB & Strategy • Sectoral Exposure: multiple sectors straddling BFSI / Pharma / Manufacturing / Retail / Tech • Overall Experience : 30 Yrs. Applied Analytics for Business - Healthcare: Learning Journey Five sessions of 1.5 hours each First session Second session Introduction- HC Healthcare 4.0 + Ecosystems & HC - AMM Stakeholders Third session HC Data Governance Fourth session Fifth session AI- ML Real World Applications – I AI- ML Real World Applications – II Mortality Prediction Suicide Rate Trend analysis STRUCTURE HC Data Intro HC Data Dimensions HC Data Strategy https://youtu.be/_mXrZEIpNMw How do we define big data? • Big data refers to ways of analyzing, extracting information, and dealing with data-sets that are too large or complex for traditional data-processing methods and software. • Big data in healthcare describes the enormous amount of health information that has become available with the advent of modern technology. Big healthcare data analytics can inform tasks from reviewing health records to tracking epidemics, and for minimizing fraud. • More so, big data is more than just traditional documentation. The sheer amount of information, the speed at which it moves, and the diverse input sources are part of the significant data definition. PROS & CONS OF BIG DATA Pros • Personalization: Big data allows healthcare providers to personalize care by creating detailed patient portals, providing a consistency between each patient’s medical history. Today, patient portals are the most common example of utilizing big data in the healthcare industry. • Telemedicine: Big data is crucial for improving telemedicine quality to bring it on par with face-to-face physician visits. Telemedicine’s popularity has been growing for quite some time, and the demand for it has never been higher than it is now. • Improved communication: Patient portals, emails, phone lines, and mobile apps connect providers to their patients like never before. By using big data, all of these forms of communication will support each other to make the experience more convenient for the patient. Cons • Billing: The billing process can vary significantly from state to state, and it can also be complicated due to different insurance companies. As big data provides a broader picture of billing, clarifying billing information, and ensuring it makes sense for the situation is critical. • Inconsistency: Not all patients or providers fill out the required forms for healthcare visits, creating gaps in the analytics and big data. Even when all parties complete the relevant paperwork, the information required can vary according to the provider. Clinical data falls into six major types Electronic health records Administrative data Claims data Patient / Disease registries Health surveys Clinical trials data Clinical data is a staple resource for most health and medical research. Clinical data is either collected during the course of ongoing patient care or as part of a formal clinical trial program. HC Data Categories Clinical data • These are patient-level data pulled from electronic medical records (EMR) and patient registries that describe how patients are treated in the real world. They include lab values, diagnoses, notes, and other information from healthcare visits with physicians and other care providers. With more data from hospitals and entire health systems becoming digitized and more easily integrated across institutions, the power of these particularly rich datasets (for example, larger sample sizes, easier comparisons across systems) is increasing. Administrative/ claims data. • Detailed patient-level data is also collected for non-clinical purposes, primarily for billing by providers to insurers and other payors, which can include diagnoses, services provided, costs, and other data required for the reimbursement of healthcare services. Other more administrative sources of data can also include data collected for tracking purposes, such as patient or population surveys. Patientgenerated/reported data • This category covers individual data describing the patient’s experience and is typically both collected and shared/reported by the patient. Today this source of data is less prevalent than others but will likely expand due to the increased use of wearable devices that automate data collection and sharing. Online communities such as PatientsLikeMe encourage and enable sharing of patientgenerated data with peers and investigators. Non-traditional, health-related digital data sources • As digital becomes increasingly prevalent in our lives, new sources of patient-level health data are emerging. These span social media posts that have a rich trove of information, especially health-focused social media sites like Sage Bionetworks. Project Data Sphere is a pharmaceutical industry-sponsored platform to share, integrate, and analyze phase III comparator arm data from cancer trials to accelerate research. Some examples of data sets in healthcare World Health Organization Human Mortality Database U.S. Census Bureau National Center for Health Statistics National Cancer Institute Some examples of data sets in healthcare Research Studies and Clinical Trials Electronic Health Records Curated Sources Claims Data What Types of Databases Does Healthcare Use? TOXNET HealthData.gov PubMed PsycINFO Assignment : Prepare a five page note on the Databases – Overview, Search, Services etc Electronic Health Records Previous diagnoses Medical history Treatment plans and medications Laboratory and test results Immunization details and dates Radiology images Understanding the Types of EMR Reporting Wired Reporting Parameter Reporting DataIntensive Reporting Building A Data Driven Strategy What is a Data Driven Organization? Characteristics of Data-Driven Organization How to Create a Data Driven Culture Data-Driven Companies Examples What is a Data Driven Organization? Insights-driven organizations are a new kind of company that are hyper-attuned to their data and use it to inform all business decisions. Organizations that use this data driven approach are growing by more than 30% each year and are projected to earn $1.8 trillion by 2021. Why are these data-intensive companies growing so fast? McKinsey Global Institute has conducted extensive research into this question and has found that data driven organizations are 19 times more likely to be profitable, 6 times more likely to retain customers, and 23 times more likely to bring in new customers. These impressive stats aren’t the result of data alone. To tap into these results, an organization needs to know how to synthesize and interpret this data efficiently and accurately. It also needs to cultivate a data driven culture, where everyone, from the top to bottom of the organization, uses data and analytics as the foundation for all business decisions. • Characteristics of Data-Driven Organization Open-Access to Data • Organizations that are not data-driven tend to fall prey to knowledge silos. This is when company data is tied up within different parts of the organization. This handicaps decision makers as they don’t have access to the full picture of the organization’s data. • Successful data-driven enterprises ensure that knowledge silos are broken down and that teams are able to access each other's data. Additionally, intuitive software is put in place to make this data sharing more efficient. • Active Data Literacy Training • Data doesn’t do you any good if you don’t know how to read it. This is why companies with a data-driven approach ensure that decision makers are trained on how to interpret and use data. Data-Driven Companies Examples Netflix • With over 158 million subscribers, Netflix harvests watchtime data to strategically develop new content. Thanks to data-driven strategy, Netflix was so confident that House of Cards would be a hit, that it commissioned two seasons before the first • episode even aired. American Express • This banking giant uses data-driven solutions to forecast customer loyalty and churn. This enables them to act on at-risk customers, keeping them loyal for longer. Coca-Cola • Coca-Cola used big data to strengthen its digital loyalty program. They effectively used data to improve customer retention. How to Create a Data Driven Culture ? Ensure at least one member of your C-suite is a data leader. Make data broadly accessible to frontline employees. Support rapid testing and iteration based on data. Be tolerant of fast failure. When hiring for non-management roles, check for proficiency in data-related topics. What is Data Literacy and Why it is Important? Healthcare providers manage an average of 8.41 petabytes of data annually—that's 8,410 terabytes. And that number isn't going down anytime soon. Between 2016 and 2018, healthcare and life sciences data skyrocketed by 878%! Now more than ever, data literacy in the healthcare industry matters. Data Literacy Skills Lead to Healthcare Solutions Data Literacy Definition • Literacy Training Data Literacy Data Literacy Definition • Data literacy may seem like a daunting term, but it’s actually just a technical description of how people read and understand information. • Data is collected over time to see patterns. The correlations within that data can be used to predict future outcomes and achieve positive results. • Having the ability to analyze data correctly can help health care facilities draw conclusions that can lead to more efficient practices. • Data Literacy Skills Lead to Healthcare Solutions • Comprehending data is a skill that, like most things, can be learned, practiced, and mastered. Understanding data can sometimes mean the difference between success and failure when implementing new policies, procedures, or marketing campaigns. • Data is used widely for many purposes, and for that reason, it’s imperative that top hospital executives are able to make sense of any valuable information they receive. • Misinterpreting or failing to realize the importance of data is like letting precious jewels slip through your fingers. Increasingly, health care facilities are pouring major resources into gathering data. Being able to read and communicate effectively can help hospitals to achieve both short-term and long-term goals. • Many hospital executives use data to analyze patterns about: • Marketing • Patient Information and Care • Employees • Business Practices and Quality Improvement • Risk Assessment • The uses and implementations of data analytics are endless. Finding intelligent and forwardthinking ways to utilize information is the key to success, and it all starts with data literacy. Everyone starts somewhere, including seasoned executives, which is why it's crucial to invest in getting proper data education. Literacy Training • Excellence in patient care and associated services depends on high-quality data. Healthcare data must also be accurate and contain no duplications or errors. Review incoming data thoroughly to ensure accuracy and prevent duplicate records. • Finally, you must keep the data consistent. Finding critical records quickly can mean saving lives. Format incoming information according to a logical system, making every piece of data correlate with the correct database field. • Obstacles to achieving proper data literacy can be overcome with proper training. Hospital employees and executives often need further assistance and education on how to use data tools properly. • Learning about how to create and read reports is a basic data skill that is imperative to that process. Reports can be customized to include information about finances, operations, and other important topics. • Data analysis techniques are another area where many employees require training. Reports can often be complicated and heavy with numbers and jargon, which dilutes the reader’s ability to zone into target information. Step one of reading any report is understanding what each data set means. With a proper education in terminology and report layout, it can become easier to decipher the numbers. • Dashboard design is another helpful tactic to allow users to see the information they need quickly and easily. • Dashboards can often be customized to show the information most relevant for an organization. It assists data leaders • within the company to spot opportunities quickly and identify problems before they become urgent. The best practices for dashboard use can be learned through training sessions online or in person. What is Data Quality in Healthcare? What Is Data Quality and Why Is It Important? What Is Data Quality Management? What Is Good Data Quality? What Is HighQuality Data Doing for Your Organization? Data quality in healthcare is vital to administrative, patient care, and records management processes. Where you get your healthcare data, as well as how you filter, organize, and update it, will all factor into your organization's success. What Is Data Quality and Why Is It Important? • Healthcare data encompasses all information gathered pertaining to a patient, their care, and their healthcare coverage. Data collection in a healthcare setting typically begins with personally identifying information (PII). Most healthcare systems use a combination of physical files and digital databases to gather this information. • Healthcare data can include but is not limited to: • Personally Identifying Information (PII) • Medical, family, and work environment history • Preexisting conditions and past hospitalizations • Past and current medications and supplements • Allergies and sensitivities • Insurance coverages and billing information • Associated healthcare providers • Other persons authorized to access patient data • Without accurate healthcare data, patient services suffer. Data comes from many different sources, including patients, family members, other healthcare providers, and third-party databases. Healthcare organizations must practice diligence to ensure data quality control. What Is Good Data Quality? • Good data quality in healthcare means that you can find information relating to a patient quickly and easily. Identifying vital information before gathering data ensures relevance and minimizes unnecessary data storage that takes up valuable space. • Excellence in patient care and associated services depends on high-quality data. Healthcare data must also be accurate and contain no duplications or errors. Review incoming data thoroughly to ensure accuracy and prevent duplicate records. • Finally, you must keep the data consistent. Finding critical records quickly can mean saving lives. Format incoming information according to a logical system, making every piece of data correlate with the correct database field. • What Is Data Quality Management? Data quality analysts were the original gatekeepers for data quality and management. Today's intelligent data quality management systems can source, clean, filter, and organize data to create a searchable database. • Quality-controlled data makes it easy for anyone with database authorization to search for, locate, and retrieve data at any time. Enforce data quality control at each point of data entry to maintain data integrity. What Is High-Quality Data Doing for Your Organization? What is data quality capable of bringing to the table? When integrated into a business intelligence (BI) solution, high-quality healthcare data leads to better patient experiences. Improving data quality means enhancing patient outcomes, minimizing redundancies, and reducing the risk of medical errors. • Consolidating healthcare data analytics in a single business intelligence solution provides almost unlimited potential for organizational improvement. Leveraging demographics data to drive more effective healthcare marketing empowers providers to take ownership of medical data to improve patient care. Patients with personalized access to protected databases can remain informed about test results and doctor recommendations. Research teams can connect remotely using anonymized data to inform new protocols. Improving the relationship between patient and provider becomes possible only with relevant, accurate, and consistent data. Combining the five capabilities is important for any HST vendor, regardless of whether its market is payers, health systems, life-sciences companies, or all three. A vendor could use these capabilities to develop new revenue pools by going deeper into its existing client base or broader into adjacent market segments, or both BIG DATA IN HC : A NEW VALUE FRAMEWORK Right living Right care Right provider Right value Right innovation • Patients must be encouraged to play an active role in their own health by making the right choices about diet, exercise, preventive care, and other lifestyle factors. • Patients must receive the most timely, appropriate treatment available. In addition to relying heavily on protocols, right care requires a coordinated approach, with all caregivers having access to the same information and working toward the same goal to avoid duplication of effort and suboptimal treatment strategies. • Any professionals who treat patients must have strong performance records and be capable of achieving the best outcomes. They should also be selected based on their skill sets and abilities rather than their job titles. For instance, nurses or physicians’ assistants may perform many tasks that do not require a doctor. • Providers and payors should continually look for ways to improve value while preserving or improving health-care quality. For example, they could develop a system in which provider reimbursement is tied to patient outcomes or undertake programs designed to eliminate wasteful spending. • Stakeholders must focus on identifying new therapies and approaches to healthcare delivery. They should also try to improve the innovation engines themselves—for instance, by advancing medicine and boosting R&D productivity. 3 + 5 = 88 DATA GOVERNANCE PROGRAM CONSISTS OF THE INTER-WORKINGS OF 7 CORE COMPONENTS: Strategy Organization Technology Directives Communication Measurement Change Management Consider them as building blocks, consisting of the processes, policies, organization and technologies required to manage your data asset and to ensure the availability, usability, integrity, consistency, auditability and security of your data. THE STATISTICS ON RARE DISEASES ARE STAGGERING: • 7,000+ known rare diseases in existence • only 5% have an approved treatment • over 400 million people affected by rare diseases • 50% of those people are children • 3 out of 10 of those children won’t live to see their 5th birthday • that’s 60 million children who face an early death due to rare disease Based on these numbers, rare diseases don’t seem so rare. • CREATE A “GOOGLE” FOR BUSINESS DATA Consider an end-to-end platform that allows you to catalog, control, and consume data in one place, like the Zaloni Arena DataOps platform. “Maturity takes hard work, time, and consistent focus. It’s not easy, but you reap the rewards of being on top of your data game.” – Says Venky, and when that “game” is creating new drugs to save the lives of children with rare diseases, enterprise-wide data governance is of utmost importance. Data governance is necessary for compliance with current regulatory expectations for data integrity in pharmaceutical R&D and manufacturing organizations. Data Governance in Healthcare - An Alexion Pharmaceutical Story (zaloni.com) ASSIGNMENT • • Write a three page note on how leading private payors operate stand-alone analytics divisions. Choose any three. Develop a three page note on “TransCelera te BioPharma” Collaboration Empowering life sciences and health care organizations with real-world data and insights Analyzing data and data sources towards a unified approach for ensuring endto-end data and data sources quality in healthcare 4.0 ArgyroMavrogiorgouaAthanasiosKiourtisaKonstantinosPerakisbDimitriosMiltiadoubStamatiosPitsiosbDimosthenisKyriazisa Computer Methods and Programs in Biomedicine Volume 181, November 2019, 104967 • Identification of heterogeneous data sources, recognizing the ones of unknown type. • Dynamic mapping of data sources’ APIs of known type with those of unknown type. • Correlation of data sources’ quality and their corresponding data quality. • Data collection from data sources through a dynamic data acquisition API. • Efficient results, ensuring end-to-end both data sources and data quality. Assignment : Read, Digest & Produce a Executive Summary (one pager) Thank you