Welcome to Population Health, Population Health IT, and Data Systems. This is lecture f, this component, population health, discusses the application of informatics and informatics methods in population health management. This unit, population health IT and data systems, explains the challenges and opportunities of using different data types, data sources, and data systems for population health IT. The objectives for this lecture are to describe the advantages and disadvantages in using various data sources for population health management and analysis. Explain the added value of non-traditional data sources for population health IT. Explain the features of a number of available population-wide health data sources. This lecture discusses the factors affecting population health data sources and also reviews a select sample of data sources with a wide population coverage. A number of factors have empowered the collection and development of new population-wide datasets. The end-user adoption of health IT solutions has increased dramatically among healthcare providers and consumers. The data variety generated by different population health data sources has also expanded by emerging data types such as electronic health records, EHRs, personal health records, PHRs, and even mobile health data, mHealth. The continuity of data flow among health IT solutions has increased across various health related domains, which is mainly due to higher adoption of mature interoperability standards. And finally, the population coverage of health IT solutions has become wider. Thus, making larger population-wide data repositories more accessible to population health analysts. As depicted by this diagram, the adoption of EHRs has increased dramatically among hospitals. Indeed, the adoption of basic EHR systems, which include more features than certified EHRs, is increased from less than 10% in 2008, to more than 75% in 2014. This has enabled eligible hospitals to develop local population health repositories and advanced their population analytics. Similar to the increased adoption of EHRs among hospitals as depicted by this diagram. The adoption of EHRs has also increased among office-based positions. Indeed, the adoption of basic EHR systems which include more features than any EHRs is increased from almost 10% in 2006 to more than 48% in 2014. This rate has been increased from 18% in 2001 to 78% in 2013 for any EHR systems. Note that this has enabled analysts to collect local EHR data from office-based positions and develop all patient population health repositories. Not only has the adoption of certain health IT systems increased over the last decade but also the variety of growing data sources has expanded rapidly. This diagram shows physician and patient interaction and how each works within a larger context. The physician, with his or her practice team, and eventually with the larger integrated delivery system, IDS, or accountable care organization, ACO, and the patient with his or her family, and the larger community. In this larger context, a long list of data sources have emerged that can be useful for population health analytics. Some of these current and potential data sources include insurance claims and management information systems, MIS, health information exchanges, HIEs. EHRs, computerized physician order entry, CPOE, and embedded clinical decision support systems, CDSS. National datasets collected by various federal, state, and non-profit agencies. Web portals for patients and families. Personal health record, PHR systems. Biometric and tele-health systems, mobile health apps and ecosystems, social networking data, and others such as marketing, human resources and geo-bound information. It should be noted that these data sources are also covering a continuum of patient and information flow among various health care entities. Ranging from private businesses to patients, providers, populations, and public health. This diagram depicts the spread of various data sources across the continuum. Note that population health analytics can benefit with having an expanded view of this continuum. Multiple population-wide data sources are currently available that can be used to train and develop population health predictive models. These data sources include both clinical and non-clinical data sources. Large population-wide clinical data sources include consolidated insurance claims such as the statewide all-payer claims databases and large insurer data bases such as centers for Medicare and Medicaid services or private insurers. Centralized or distributed EHR clinical research data warehouses, such as EHR-based data warehouses developed by large Health Maintenance Organizations, HMOs, integrated delivery systems, IDSs. ACOs, or the Veterans Health Administration, VHA, and large clinical registries of the National Patient Center Clinical Research Network, PCORnet, or the National Institutes of Health's, NIH, Commons. Mobile health data, which are collected and shared by various smartphone platforms. And also large-scale population-wide surveys and registries such as CDCs, Behavioral Risk Factor Surveillance System, BRFSS. The all-payer claims databases, APCDs are large-scale databases that systematically collect medical claims, pharmacy claims, dental claims, and eligibility and provider files from private and public payers. APCDs usually cover an entire state and often include commercial insured claims only. The first statewide APCD system was established in Maine in 2003. Currently, more than 30 states have, are implementing, or have strong interest in APCDs. According to the APCD council, information typically collected in an APCD includes diagnosis, procedure, and medications. Information on service provider, prescribing physician, health plan payments, member payment responsibility, type and date of bill paid, facility type, revenue codes, and service dates. This map shows the coverage of statewide all-payer claims databases. As shown, a considerable number of states have existing APCD's, or in the process of implementing one. Another population wide database, is the insurance claims reported to. The Centers for Medicare and Medicaid Services, CMS. Medicare insurance claims, include almost 44 million patients, mainly seniors, on a national scale. Medicare data is maintained by CMS, and includes sections such as. Medicare Provider Analysis and review, MEDPAR file, referred to as Part A. Outpatient claims which contains Part B, home health agencies, hospice, durable medical equipment, DME, and medications. Also known as Part D. Medicare data, can be acquired through various CMS programs, for research and population health analytic. Medicaid insurance claims are statewide, and the number of members varies by state. Medicaid data include the following categories, Children's Healthcare Insurance Program CHIP, non-disabled adults such as low-income parents. Pregnant women, individuals with disability, and low income seniors. Medicaid data's often controlled by the states, and each state may have different mechanisms, to share the data with researchers. A number of Commercial Insurance Claims data warehouses, are also available for population health analysis. These large claim databases, often have a national scope. Some of these large commercial databases, include non claims data sources as well. A sample list of data sources, often used by population health analysts include. IMS Health, which contains a mix of claims, EHRs, labs, and other data sources. Optum which contains claims, EHRs, surveys and other data sources. Truven Health Analytics MarketScan, which mainly contains claims. And FAIR Health, which also contains claims. One of the largest EHR data warehouses in the US, is the VHA's Corporate Data Warehouse, CDW. The VHA uses the Veterans Health information systems and technology architecture, VistA, a unified EHR system, across the entire national health care system. CDW includes the data collected, across the entire VistA system. The VHA CDW, supports administrative and research objectives of the VHA. And includes historical data from 1999. Data are added to CDW, on a daily basis. And additional domains will be added to CDW, in the near future. CDW includes the following, population health analytic data types. Demographics, consults, health factors, immunization, mental health, primary care management, vital signs, inpatient data, laboratory or chemistry orders and results. Outpatient encounters and pharmacy data. Another source of population health data, are the data sets generated by the Healthcare Costs and Utilization Project, HCUP. HCUP is directed and maintained, by the Agency for Healthcare Review and Quality, AHRQ. HCUP datasets, are mainly concerned with, inpatient and emergency department ED discharges. HCUP data includes both, national and statewide datasets. HCUP also includes related software tools and products, that can help analysts better understand and analyze the data. According to the AHRQ, HCUP databases are derived from administrative data. And contain encounter level, clinical and non-clinical information. Including all listed diagnoses and procedures, discharge status, patient demographics, and charges for all patients, regardless of payer. HCUP data include a variety of payers, including Medicare, Medicaid, private insurance, and even the uninsured. HCUP data go back, to the beginning of 1988. The HCUP databases, enable a broad range of health policy research. Including cost and quality of health services Medical practice patterns, access to healthcare programs. And outcomes of treatments at the national, state, and local levels. HCUP contains multiple databases, that are developed for different purposes. A sample list of HCUP databases from the AHRQ, includes National Inpatient Sample, NIS. The largest publicly available all-payer hospita, inpatient care database in the US. Kids' Inpatient Database, KID, composed of hospital inpatient stays for children. Nationwide Emergency Department Sample, NEDS, which captures information on ED visits that do not result in an admission. As well as ED visits that result in an admission, to the same hospital. HCUP databases also include Nationwide Readmissions Database, which is a unique and powerful database. Designed to support various types of analyses, of national readmission rates for all payers and uninsured individuals. State Inpatient Databases (SIDs), which are a set of hospital databases containing the universe, of the inpatient discharge abstracts. From participating states, translated into a uniform format to facilitate mult-state comparisons and analyses. State emergency department databases SEDDs, which are a set of databases, that capture discharge information. On all emergency department visits, that do not result in an admission. Another population-wide data source is the national Patient-Centered Clinical Research Network, PCORnet. PCORnet, is the Patient-Centered Outcome Research Institute's, PCORI, flagship initiative. To build an efficient and patient-centered platform, for research. PCORnet, is intended to lower the cost, of clinical trials and create a reusable national infrastructure for research. PCORnet center's often harness the potential, of EHR data for this purpose. The long-term goal, is that PCORnet Center's will share their data, in a large national data warehouse. PCORnet includes 29 networks and a coordinating center. PCORnet networks, include 11 Clinical Data Research Networks, CDRN, which are system-based networks. Managed by IDSs, academic medial centers, and federally qualified health centers, FQHCs. PCORnet networks, also include 18 Patient-Powered Research Networks, PPRN, which are the result of patients working together. To discover, propose, and answer relevant research questions. There are 155 organizations and more than 3,000 collaborators across the US, who are involved in PCORnet. The number of patients receiving care, in the participating systems is in the millions. Thus, making it a suitable data source, for population health analysis. PCORnet's data sources include data collected in or by EHRs, patient powered registries, clinical and translational science awardees, CTSAs, FQHCs, HIEs, IDSs, pharmacy vendors, payers and others. There are platforms that facilitate the collection of patient generated data from mobile or app-based programs. Here are two examples. SMART Health IT quote, SMART Health IT is an open, standards based technology platform that enables innovators to create apps that seamlessly and securely run across the healthcare system using an electronic health record, EHR system or data warehouse that supports the SMART standard. Patients, doctors and healthcare practitioners can draw on this library of apps to improve clinical care, research and public health, end quote. HealthKit is a tool that allows health and fitness apps on iPhones to work together. Many apps have already been developed using HealthKit. HealthKit utilizes the user's consent process provided by the ResearchKit, another iPhone app. HealthKit collects a variety of data, such as daily step counts, calorie use and heart rates. There are also similar platforms on other smartphone vendors. There are also large scale national and statewide surveys that can be used for population health analytics. Here is a sample list of surveys administered by the CDC. Behavioral Risk Factor Surveillance System, BRFSS. National Health Interview Survey, NHIS. National Health and Nutrition Examination Survey, NHANES. National Health Care Surveys, NHCS. National Vital Statistics System, NVSS. National Survey of Family Growth, NSFG. National Immunization Survey, NIS. And of course, there are ample data sources that are not primarily considered health related, but can be matched to various health determinants and be used for population health analytics. Some examples of these non-health data sources are social and administrative data sources derived from federal and state departments, such as the National Death Index. Environmental data sources, such as OpenMaps and other datasets created by federal, state or local agencies. Marketing data sources, such as consumer review and purchasing datasets. And finally, financial data sources, such as income levels, credit history and others. This concludes lecture f of Population Health IT and Data Systems. In this lecture, we discuss factors that affect population health data sources and provided some examples of large scale population data sources. Factors affecting population health data sources include the increasing adoption of health IT solutions among end users. The increasing variety of population health data sources and the expanding continuum of data sources that can be used for population, health research. Data sources with the wide population coverage include All-Payer Claims Database, APCD. CMS's Medicare and Medicaid data. Large commercial insurance claims databases. VHA's Corporate Data Warehouse, CDW. Healthcare Cost and Utilization Project, HCUB datasets and others.
