PDM-DECISION: Data-Driven Decision-Making to Control Medical Costs, Promote Effective Medical Management and Enhance Quality of Care Introduction Data-driven decision-making to control medical costs, promote medical management and enhance quality of care. A great premise. You have the data, but do you have the means to manipulate this data and produce actionable information to address critical issues that impact your business? Which members are likely to require a disproportionate share of the health plan’s resources in the near future? How much does it cost to manage plan members with different diseases and complex combinations of diseases? What are the cost components of care? Which members should be in disease or care management programs? How effective are the disease management programs? Are my members and providers using prescription drugs in the most efficient way possible to promote health? Can I identify members/providers exhibiting polypharmacy and/or poor compliance and intervene to improve their health and reduce costs? What are my medical and pharmacy costs now and what are they likely to be next year? What are the key drivers of my medical and/or pharmacy costs? How can the efficiency and effectiveness of the provider network be improved? How can the plan identify and communicate opportunities to improve the delivery of care to our members? Plan Data Management, in association with the Johns Hopkins Bloomberg School of Public Health, has developed an enterprise clinical decision-support system designed to meet the needs of all functional areas. The system integrates the Johns Hopkins ACG system (Adjusted Clinical Groups) — a risk adjustment and predictive modeling system now used by more than 200 health care organizations. PDM-Decision marries medical claims, pharmacy data, membership demographics, and other clinical and financial data into a comprehensive data warehouse and provides meaningful person and population oriented profiles of utilization and resource consumption. Sophisticated data models categorize data into clinically-meaningful and financially relevant groups. Our grouping algorithms employ the following systems as well as others: The Johns Hopkins ACG Predictive Model (ACG-PM) Johns Hopkins ACG risk adjustment methodology Disease classification systems developed by both Johns Hopkins and the Agency for Health Care Research and Quality (AHRQ), to group diagnosis codes into clinically relevant “buckets” suitable for analysis. Prescription drug classification systems developed by Cerner/Multum, to facilitate analysis by therapeutic class, active ingredients and routes of administration. A service classification system developed in a collaboration with researchers from CMS, John Hopkins and leading “think-tanks” that take the 15,000 CPT-4, HCPCS and revenue and creates clinical meaningful service utilization categories. The ability for Medicare-Advantage plans to model using the CMS-HCCs. Inpatient hospital utilization categorized by DRGs and All-Patient-Refined DRGs (APR-DRGs). Hospital outpatient services categorized by proprietary methods incorporating the Ambulatory Patient Category (APC) system. Indicators of inadequate primary care management of hospitalized patients using the Ambulatory Care Sensitive Conditions (ACSC) Unlike many systems that just produce a mass of "canned" reports without the ability for the user to customize them, PDM-Decision allows users, with the click of the mouse, to develop the reports they need and to quickly and easily drill down to the member level by a variety of dimensions and variables. For example, PDM-Decision allows users to analyze costs and utilization by any combination of the following dimensions: Type of Service - (e.g., inpatient care, E&M visits, interventional radiology, home health care, specialist visits, etc.) The presence of certain diseases - in a myriad of combinations (e.g., how much do my members with diabetes and heart disease cost?) Johns Hopkins ACG Group - (how much do providers spend on members with similar levels of illness burden?) Johns Hopkins ACG Predictive Model - (who are likely to be the highest utilizers next year and what are their expected costs?) Time - (what is the cost trend from 2003 to 2004? How about 2003-Quarter 1 vs. 2004 Quarter 1)? Prescription Drug Use — generic drug episodes and active ingredient episodes of use Outreach/Tracking — Which members have not had an E&M visit within the last 12 months? Which chronically ill members (e.g., patients with CHF) have not had an E&M visit within the last 12 months? Hospital Use — Are readmissions post-MI related to the proper or improper use of beta blockers and ACE inhibitors post-MI; what is the distribution of inpatient episodes by DRG; do readmissions cluster around selected providers or patients)? Predominant Providers - How do providers that serve as “predominant providers” for cohorts of members perform, in terms of financial efficiency and clinical outcomes? PDM-Decision's comprehensive data warehouse, elaborate and novel uses of clinical grouping, user-driven reporting and point & click analytical capabilities can make data-driven decision making in healthcare a reality for your organization. A System Overview PDM-Decision is a clinical and financial enterprise decision-support system. It empowers users to access, analyze and present data into actionable information for use in clinical and quality management, financial decision making and provider care delivery. PDM-Decision joins a variety of clinical data describing the illness burden (morbidity) of health plan members and marries these data to the financial measures associated with members’ utilization of health care services. Health status is measured using a variety of proprietary tools, including the ACG System developed by Johns Hopkins Bloomberg School of Public Health. The PDM-Decision data warehouse not only provides a single location to house data from a variety of sources (medical claims, mental health, pharmacy, health status information, member demographics, etc.) but it collects claims to form clinically-meaningful “medical events”. Our proprietary methodologies eliminate duplicates, even where several different bills are submitted (such as many ER claims where two distinct providers submit claims). After removing duplicates we assign procedures to a hierarchical type of service classification system and standardize around a fixed time interval. Outpatient care provided by hospitals is rolled up to the level of a “medical event”, using the Ambulatory Patient Category system, eliminating much of the difficulty and lack of specificity associated with analyzing revenue code data. Prescription drug episodes of care are constructed using a proprietary methodology. Users of PDM-Decision have the option to analyze costs of care using their own claims experience (typically allowed or paid charges) or to normalize the claims data using standardized fee schedules which are imbedded into PDM-Decision. Some types of analyses call for using actual claims costs (e.g., financial modeling that is compared to the entity’s financial statements), while others are better served by using normalized costs (e.g., provider profiles should not be confounded by differences in contract terms or unit prices). Once PDM-Decision creates the medical events, they are rolled up into comprehensive clinical and financial person-level profiles. Although medical and pharmacy services are purchased and accounted for at the claim or service level, effective analysis and planning must occur at the person-level. Through PDM-Decision’s incorporation of tools from Johns Hopkins, Cerner/Multum, CMS and the Agency for Healthcare Research and Quality (AHRQ), data can be easily organized & sliced based on accepted industry groupings and is “normalized” for accurate and meaningful comparisons across populations and individuals. This provides a powerful tool to perform numerous analyses. The product is organized into business function modules. Each comes with a set of standard report templates with an integrated point & click analysis tool allowing users to quickly and easily drill down and analyze standard reports further, plus the ability to analyze and customize your reports and output. 1) Medical Management & Predictive Modeling: Disease Management Targeting Member Level Case Management Reports Disease-based health risk assessment, utilization and cost analysis Outreach/tracking 2) Efficiency Profiling: Hospital Use (Utilization, Ambulatory care Sensitive Conditions, Readmissions) Person Level Utilization by Type of Service Provider Level Efficiency/Utilization by Patient conditions or Type of Service Pharmacy Use/Overuse (Prescription drug profiles) Report Cards 3) Financial Modeling and Product Development: Cost Trends Product Pricing Underwriting Provider Reimbursement Impact 4) Quality Profiling Disease Management Compliance Provider Prescribing Patterns – Pharmacy Use Inappropriate Service Use Standardized Reporting PDM-Decision offers a series of pre-packaged parameter-based dashboard reports for each Module in order to provide quick, easy access to clinical and financial information commonly used by key decision-makers and analysts. These same reports can serve as a starting point for detailed analysis, with the ability to drill down to the service, provider, member or other unit of measure level by clicking on key variables within the report. The majority of the standard reports are directly linked to the interactive analytic tool whereby the user can click on certain dimensions and fields within that report and immediately drill down on those results. Analytical Tool and Ad-Hoc Reporting PDM-Decision also allows users, with the click of a mouse, to develop the reports they NEED and to quickly and easily drill down on populations to the member level on a variety of dimensions and variables. Selection of dimensions and parameters is facilitated through drop-down lists and standard industry terminology. Users can generate and store their own analyses and create tabular and graphical report outputs from that analysis. Data Inputs PDM-Decision utilizes the same basic file types and formats as other PDM Products. Files are flat, fixed length text formats: Claims Data Extract file Pharmacy Claim Extract Member file – member identification and demographic information Span File - Date sensitive member enrollment and demographic information Technology Platforms and Features The application uses the most modern database technology – data warehousing and online analytical processing (OLAP) to provide easy and fast access to data. Unlike relational-only models which have to resolve each query from "scratch" each time, the OLAP data in PDM-Decision is "pre-aggregated", lending itself to high performance and flexibility. Reports can run in "canned" format, ad-hoc format, or a combination of the two. You can run a report with parameters, and click a button to be able to perform further analysis on those results using the integrated OLAP/query reporting tool. Web-based application SQL Server 2000 relational database Data warehouse employing multi-dimensional Online Analytical Processing (OLAP) OLAP query/reporting tool Relational Reporting capabilities For more information, contact Stephen Jackson. Copyright © 2008 Plan Data Management, Inc. All rights reserved.