Accountable Care Organizations
Setting the stage for ACO performance
The key is streamlining clinical, financial and operational information.
By Don Pettini, January 2013
Accountable care organizations (ACOs) stand to transform the delivery of healthcare as we know it by arresting spiraling costs while improving outcomes. The model, however, is not without risks to providers.
To thrive as an ACO, healthcare organizations must understand not only their true costs of care delivery, but the correlation between therapies, costs and outcomes. The ability to turn these silos of disparate data into actionable insight is no longer nice to have – it’s a must have.
Enterprise data warehouses and analytical applications are the keys to unlocking this insight. Building a business intelligence environment from the ground up, however, has historically caused trepidation even among the most seasoned IT organizations, requiring huge investments of time and money. Times are changing.
ACOs, which as of July 2012 were serving 2.4 million patients in 40 states and the District of Columbia, are expected to gain momentum as healthcare reform rolls out in earnest. The success of the ACO model depends on the ability to incentivize healthcare organizations and providers to form links and facilitate coordination of care delivery to eliminate redundant and unnecessary services. There are risks and opportunities associated with this model. If costs fall below a set budget, ACOs, which today largely serve Medicare patients, share in the profits. If costs exceed the budget, they bear some of the losses.
The stakes are high, and ACOs are looking to their data infrastructures to help them strike the right balance.
Getting a complete picture
To navigate risk and fulfill their potential for delivering high-quality care and outcomes at a lower cost, ACOs require a new level of business and clinical intelligence. They must be able to definitively answer these questions:
- Who are our patients, and how is their health as a population (including chronic and acute problems)?
- What services do we deliver to them?
- What is the true cost of these services and procedures?
- Is there an alternate path that can improve outcomes and reduce costs?
- Can we identify gaps in care and the impact addressing them will have?
The questions appear deceptively simple. Most healthcare organizations, however, cannot answer them precisely for themselves, let alone when coming together with other providers as part of a larger group.
To understand true costs and outcomes, ACOs must bring together several critical pieces of internal and external data, including:
- Patient information, such as demographics and family history;
- Clinical data, such as lab and test results, medical history, medications, procedures and other information typically found in electronic medical records (EMRs) and electronic health records (EHRs);
- Financial information about the costs, direct and indirect, of services provided;
- Outcomes, including improvement in vitals, quantitative disease measurements (such as HbA1c), reduced hospitalizations, etc.; and
- Claims or health information exchange (HIE) information, as patients often go outside of the ACO for care. Organizations have historically faced roadblocks in integrating these disparate internal and external data sources. Those that have made progress on the integration front often continue to struggle with the ability to analyze the data in a timely manner to yield actionable insight.
New times call for new approaches
How can healthcare organizations build the data and analytical foundation that they require in the age of ACOs? The good news is that many already have at least some of the building blocks in place. Core to any effort is data governance. Particular aspects of data governance can help with topics such as terminology mediation and standardization, along with master data management. In addition, providers no longer need to develop these solutions alone as there is a growing set of commercial off-the-shelf (COTS) and software-as-a-service (SaaS) offerings available today.
Consolidating and integrating patient, services, claims and cost data in a single data warehouse for the multitude of source systems provides the necessary centralized hub for both reporting and analysis. However, any healthcare organization that has attempted to build a data warehouse from the ground up knows that one of the most complicated, time-consuming and expensive parts is the extraction and integration and ongoing maintenance of clean, complete data for analysis. Today, healthcare organizations have new options with COTS models and data integration infrastructures built specifically for the industry, which can be tailored to organization-specific requirements. This approach saves both time and money, due to the collective expertise of the vendors developing these solutions.
Ideally, the warehouse data model should be structured under a relational third normal form (3NF) model, with marts created as star/snowflake schema for optimum performance and ease of use. Near-real-time extract, transform and load (ETL) can be used for more rapid and optimized run times when business requirements are well defined.
The framework must be able to handle many different sources of data and varying levels of quality and completeness and also be able to connect the two. For example, the system may have a patient record without any encounter history and an encounter history without a patient. It must be able to bring this data together and handle the normal case of these coming in at two different times from different systems.
It is also imperative to capture and leverage data at different levels of granularity. Today, many clinical and operational source systems enable providers to capture data at levels never before possible. For example, new wireless capabilities allow organizations to track vitals during a procedure, or over a long period in the outpatient setting, for chronic disease management. To enable predictive cost-quality analysis, providers also should reliably capture the following: detailed clinical data; clinician, support staff and administrative salary and benefit data; billing and administrative costs; equipment, supplies and medications leveraged per case; and even square-footage costs involved in care delivery. This more granular data, which is important to the activity tracking and activity-based costing that many organizations seek, will likely help to drive more accurate and insightful analysis of true costs, how and where costs are expended, and their relationship to outcomes.
The data warehouse environment also must be extensible. Healthcare organizations are very dynamic, and so is the data they collect. For example, a provider may feel that it is important to begin to record data on the arm used to capture blood pressure readings. Traditionally, an addition like this would require a change to the data model and ETL infrastructure, which can divert considerable time and resources from other priorities. As the range and type of data captured grows, so does the maintenance burden. Extensible solutions enable organizations to accommodate new data categories without the need to completely rework the data model.
A final consideration when creating a viable data warehouse is whether the environment will become more cost effective to operate and maintain as it grows, thereby encouraging organizational stakeholders to centralize information. An enterprise data warehouse asset would support clinical, financial, operational and research uses. Focusing on an initial set of business needs with a central asset keeps scope reasonable and allows for a quick time to value. The timeframes rival other mart-focused approaches, but you gain the advantages of an enterprise warehouse.
Next stop: data analysis
The data warehouse alone cannot deliver the insight that ACOs require. Analytics brings it all together, making sense of data in a way that is both meaningful to business users and actionable to care providers. It is no longer enough to simply measure and understand data. Prudent organizations should be able to leverage data – in essence creating a closed-loop system – to facilitate consistent improvement.
Once again, options abound, including patient, physician, costing and allocation, planning, supply chain and operating room analytics solutions. When choosing a solution, healthcare organizations should keep in mind several important criteria, including scalability (as data volumes increase dramatically), flexibility (to accommodate changing requirements and the need for new metrics), as well as the ability to easily model multiple scenarios for detailed analysis and insight into numerous “what-if” scenarios.
Considerations should include an enterprise-wide hospital tool that consolidates clinical, financial and operational data in one reporting and analytical view; surveillance reporting to improve patient outcomes before discharge; predefined indicators and measures for retrospective analysis; data extraction and online analytics processing designed to specifically promote data integrity – with no additional data transfer interfaces required – resulting in lower total cost of ownership; a healthcare-specific business intelligence tool; and an advanced Web-based portal with a shared dashboard solution.
Achievement in the era of ACOs will require an efficient alignment of internal resources to streamline clinical, financial and operational information to more effectively manage costs across the continuum of care. Setting the foundation for accountability, and ensuring the appropriate IT infrastructure is in place to deliver insights when and where needed, can drive innovation, and provide for achievement and profitability for forward-looking organizations.
About the Author
Don Pettini is senior director, healthcare product strategy, Oracle Health Sciences. For more on Oracle Health Sciences: www.rsleads.com/301ht-206