An ACO’s ability to effectively manage the risks associated withvalue-based payment depends on how precisely it can account for costs and whether it can apply predictive analytics to analyze clinical and financial outcomes.
At a Glance
- To succeed under value-based payment, accountable care organizations (ACOs) must be able to link, analyze, and compare clinical and administrative data from across their constituent organizations.
- ACOs require a precise costing methodology, such as activity-based costing, to be able to manage costs effectively and gain critical insight into which service lines are delivering value from a clinical and financial standpoint.
- To support informed strategic decision-making, ACOs also require ready access to integrated patient encounter data to be able to perform the sophisticated modeling of predictive analytics.
With the Supreme Court ruling on the constitutionality of the Affordable Care Act (ACA) and the presidential election behind us, accountable care organizations (ACOs) are solidly positioned to play a pivotal role in realigning our delivery system and driving down costs while improving outcomes. The same is true not only for the type of ACO described in the ACA, but also for the more generic concept of ACO that has gained prevalence, defined simply as a model of payment based on providers being held accountable for outcomes.
Providers embarking on an ACO strategy will face a new financial reality as well as a new set of information priorities and requirements. Clearly, the success of an ACO will depend, at least in part, on ensuring that the provider participants have strong incentives to effectively form links and coordinate care delivery to improve the quality of care while eliminating redundant and unnecessary services. Yet delivering high-value care in a way that maintains the organization’s financial solvency and/or profitability requires not only a strong focus on quality, but also a thorough understanding of costs and outcomes.
To obtain the clear picture of costs and their relationship to quality required in an ACO environment, organizations should consider:
- What drives profitability by case, service line, etc.
- Which activities and processes produce value, enhance quality, and improve outcomes
- How contractual payer prices compare with costs
- How much capacity is being used
- Whether equipment investments are justified
To support these considerations, ACOs must be able to link clinical and administrative data from multiple organizations in a way that facilitates a measurable and accurate analysis of quality, cost, and outcomes. Moreover, they will require new models and tools for quantifying and understanding the clinical or financial experience of their patients. Those that have yet to initiate this transition should be prepared to challenge the status quo, which may involve abandoning long-held processes and financial models and transforming infrastructures to gain the visibility and insight required to achieve success as an ACO.
In short, effectively accomplishing this transformation requires new approaches to analyzing both costs and outcomes. With respect to cost analysis, ACOs should work toward adopting activity-based costing (ABC) in lieu of more traditional costing methodologies. And with respect to outcomes analysis, ACOs will need to employ predictive cost-quality analysis, which will require much more than the historical data hospitals have relied upon for traditional analyses. The purpose of this discussion is not to provide specific guidance in how to apply these analytical methodologies, but to underscore their importance in realizing an effective ACO strategy.
Analyzing Cost: ABC
The well-known risks associated with an ACO model make it critical to account for all of the costs—e.g., for technology, human capital, medical devices, and drug treatments—associated with the highly coordinated, value-added care that an ACO is designed to deliver. ABC is a cost accounting methodology that provides the necessary detail by assigning costs to an organization’s activities (or cost drivers) related to its products and services in a way that reflects the extent to which the activities consume the organization’s resources. In this way, ABC provides a higher degree of specificity than can be achieved through traditional cost accounting methodologies, thereby providing an ACO with critical insight into which services are value-added from not only a quality-of-life, but also a financial perspective.
The theory of ABC is that a more accurate service line cost is derived when cost drivers are established based on assignment of a cost to an activity and the resources required to perform it. To properly implement ABC, separate measures must be established for each cost driver so that charge items that do not entail the use of a specific resource are not allocated any costs for that resource. ABC requires the use of complex and specific objective measures such as time studies and actual prices, which are challenging and costly to implement and maintain, but produce more precise results than can be achieved using other methods.
The service line connection to ABC. The term service line, more commonly known in manufacturing and retail as product line, is less precise in health care than in other industries because of the human nature of the profession. Nonetheless, effective use of ABC in health care requires the identification of cost drivers according to specific service lines so that healthcare leaders can ascertain which service lines are profitable and drill into the detail to analyze the reasons why.
Defining service lines. In healthcare enterprises, including those organizations that constitute an ACO, service lines have traditionally been organized by departments (e.g., radiology, surgery, pharmacy, lab, and nursing). The problem is that grouping departments as services rarely reflects the way patients typically use a healthcare facility, given that a patient may use services from several departments in one stay. The ACO will be paid based on the patient’s total stay, rather than for individual services purchased from each department. Most payments cannot be broken down into the granular department level of details.
What, then, is the most useful way to define service lines in an ACO environment? Organizations should consider basing service lines on patient encounter attributes. A data warehouse used for tracking patient encounters can provide a good source of patient attributes, such as inpatient, outpatient, and diagnosis category or grouping.
The first breakdown is typically inpatient versus outpatient services. Inpatient services then can be broken down into more detailed service lines in a variety of ways, including by major disease category (MDC), DRG, and ICD code. The availability of thousands of such possible codes or groupings can complicate this process. One could start with MDCs and then refine them based on DRGs and ICD codes. Outpatient services likewise can be defined in terms of current procedural terminology (CPT) or ambulatory payment classification (APC) codes.
The above approach is based on disease-or procedure-driven service line definitions. In an ACO environment, there are also business reasons to look at service lines from a payer or physician standpoint. For example, ACO payment contracts are often more complex, harder to administer, and less able to keep pace with rising costs than hospital payment contracts. Bridging the widening gap between care delivery and administrative costs requires ACOs to truly understand the cost of the care on a per-patient basis so they can negotiate realistic payment terms and stop-loss measures with their contracted payer groups. To effectively set a strategic and proﬁtable course for the future, internal planning and analysis groups should have insight into the volume and cost of care across functional departments (e.g., radiology, lab services, and nursing) and clinical service lines (e.g., oncology, cardiac services, and neuroscience).
It also is important to recognize that the cost impact of rendering patient care is not included in most clinical applications. Financial decision-support costing systems, however, can do a very credible job in assigning direct patient care (sometimes referred to as variable), direct overhead (or fixed), and indirect cost to the individual charge/service level and, ultimately, to the patient encounter. The challenge of achieving true ABC lies in the fact that the capture of nonbillable patient activity (transport and back-office administrative) is often very limited. Unless an ACO can allocate all of these costs directly to treatments, it will likely never gain the true cost of patient treatment care delivery. Averaging costs is not a solution as it destroys the viewable difference in the cost of operations or treatments—for example, between a quadruple heart bypass versus a tonsillectomy.
Ideally, costs should be attributed to all activities, whether chargeable or not, so that true variation in cost can be measured across individual patients. Patient acuity or procedure costing is a prime example of the potential in this area. If a patient needs to be watched because he or she is at high risk of a fall, the patient will require a higher level of supervision than will a stable patient in the same unit. The cost implications of additional supervision can be specifically identified if the assignment of expense is based on the resource level providing the additional care, such as a patient care aide.
Outcomes Analysis: Predictive Analytics
Further enhancements to cost accounting will likely result from better integration of encounter-specific data from various clinical and operational systems. By achieving this integration, ACOs can begin to develop and deploy predictive modeling capabilities to better understand cause-and-effect relationships and their impact on quality of care, and to proactively alter patient care practices based on the analysis of previous experience.
With its approach to service lines and structure for data capture and allocation in place, an ACO should initiate data analyses to support informed strategic decision-making. Traditionally, hospitals have focused such analyses primarily on historical data, leaving them to loosely extrapolate what the findings might mean for the future—whether in terms of costs or care outcomes. The advent of predictive analytics extends the horizon, not only building on historical data, but also enabling sophisticated modeling to accurately and flexibly project into the future, based on one or more modeling scenarios.
Analysis of historical data alone will be insufficient to improve outcomes and drive down costs. Within an ACO structure, organizations will require predictive cost-quality analysis to:
- Measure the impact of primary care on acute care costs across the continuum of care
- Analyze healthcare data in various dimensions to pinpoint cost-quality improvement opportunities across the continuum of care
- Create “what-if” scenarios to analyze various strategies (e.g., analysis of the impact on cost and quality of standardizing one leading practice [order set] over another across the system)
An important aspect to facilitating truly advanced analytics is complete integration of clinical data with financial support cost data at the patient-encounter level. To control costs and improve care, an ACO must be able to view and analyze care in various settings, from physician offices, to outpatient clinics, to acute inpatient stays. An ACO requires such capabilities to determine whether a patient’s stay was profitable and to understand the relationship between costs and outcomes.
Clinical systems contain the data elements necessary to know what procedures were done, the diagnosis of the patient, results of tests, quality outcomes measures, and more. If the care provided in various settings is captured in a single database, a healthcare organization can perform comparisons of outcomes across varying scenarios. This capability also provides the basis for predictive modeling from the perspective of generating data from which algorithms and scenario analysis can be derived. For example, obese, hypertensive diabetic patients may avoid admission to acute care settings when they are seen at least monthly in a physician’s office. Currently, that knowledge would be based on hypothetical extractions, rather than factual data.
Creating the environment for predictive cost-quality analytics needed in an ACO environment typically requires four primary components:
- Accurate and detailed data from core systems
- A unified data warehouse with a healthcare-specific data model
- A cost accounting system
- Analytical applications
Although many organizations have not yet implemented all of these components, many have invested in creating a solid foundation, and there are numerous commercial off-the-shelf, ready-to-deploy solutions available for purchase today.
Data from core systems. Poor data equate to poor outcomes. This statement is just as true for assessing costs and outcomes of care as it is for diagnostic medicine. Today, many clinical and operational source systems capture data at levels of detail never before possible. For example, new wireless capabilities assist organizations to track labor expended in seconds as opposed to minutes or quarter hours.
To support predictive cost-quality analysis, ACOs require the ability to capture a wide range of data reliably, including:
- Detailed clinical data
- Clinician, support staff, and administrative salary and benefit data
- Billing and administrative cost data
- Data on equipment, supplies, and medications leveraged per case
- Data on square footage costs involved in care delivery
These detailed data, which are essential for effective micro-costing, have the potential to drive more accurate and insightful analysis of true costs, how and where the costs are expended, and their relationship to outcomes. With the advent of technology to track episode-specific consumption of resources, micro-costing models now attempt to leverage this actual consumption data for allocation purposes. Micro-costing models allocate indirect costs to the lowest possible level—the encounter or episode of care—based on actual consumption drivers.
Centralized data warehouse. An ACO can achieve more efficient integration by implementing a centralized data warehouse to facilitate managing data in a quick and precise format. The data warehouse should contain all relevant information about each patient encounter. Many healthcare organizations are capable of capturing much of this information, but they face a challenge in integrating or aggregating the information in a single, unified data warehouse. This challenge is particularly acute for an ACO, which must consolidate data from multiple provider entities. Nonetheless, consolidating and integrating these diverse data are necessary steps to providing a central hub for both reporting and analysis.
The data warehouse requires data from general ledger and patient accounting sources to allow for costing. Data elements required for service line reporting include encounter demographics, charge, payment, diagnoses, procedure, physician information, insurance information, and more. General ledger data should include account and department level details as well as statistical information such as square feet and number of employees.
Within the data warehouse, the data should be organized in an intuitive and well-documented manner, and the data model should support highly scalable extractions and views for analytical applications to facilitate the acquisition of valid, accurate, and standardized data from source systems. One of the most time-consuming and expensive parts of building a data warehouse from the ground up is creating data models that enable effective data mining and analysis.
Today, ACOs have the option of purchasing models that can be tailored to a variety of specific requirements, including, for example, direct and full costing by department; activity-based profit and loss statements specific to patient and DRG level; and clinical indicators related to DRGs, case mix indices, and average length of stay metrics.
The warehouse data model should be structured under a relational star or snowflake schema, whose names reflect how the architectures look when presented in a schematic showing the relationship of the data tables to the various attribute or dimension tables populated by the main data store. These standard architectural approaches are designed to ensure high levels of query performance and ease of use. Online analytical processing (OLAP) can be used for faster run times when business requirements are well-defined, and can support slice-and-dice and drill-down capabilities into specific data—for example, enabling an ACO to look at financial performance and/or outcomes of a specific department across multiple facilities. In addition, building multiple OLAP applications based on specific business purposes is a good approach as the relational warehouse will provide all of the data feeds in one place.
Cost accounting solutions. A cost accounting solution is required to interface with the data warehouse. An effective application can support all of the methodologies required for cost accounting analytics by taking costs from a data model and allocating them based on specific methodologies for various cost centers or service lines.
Analytical applications. These applications are required to make 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—ACOs must be able to apply data to facilitate consistent improvement.
A wide range of analytical solutions are currently available, including patient, physician, costing-and-allocation, planning, supply chain, and operating-room analytics solutions. When choosing a solution, ACOs 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)
- Ease of use for modeling multiple scenarios for detailed analysis and insight into numerous “what-if” scenarios
ACOs should consider adopting an enterprisewide tool that provides:
- Clinical, financial, and operational data consolidated in one reporting and analytical view
- Surveillance reporting to improve patient outcomes before discharge
- Predefined indicators and measures for retosperctive analysis
- Data extraction and online analytics processing designed specifically to promote data integrity, without the need for additional data transfer interfaces, resulting in lower total cost of ownership
- A healthcare-specific business intelligence too
- An advanced web-based portal with a shared dashboard solution
To allocate services into various service line categories, ACOs can use a rule-based service line grouper, which can assist in defining alternative service line definitions based on multiple criteria. Once the service lines are defined and costing is completed, an ACO can create many different service line performance reports, including inpatient service, physician, and dashboard views.
Organizations moving to an ACO model should start early to evaluate their preparedness to accurately identify costs and assign them to patients. They should begin by assessing their existing infrastructure to determine which systems are in place and what additional support or solutions might be required. Finance managers should identify and thoroughly analyze all cost centers (including administrative costs). Building on this foundation, the organization should then assess its current proficiency in accurately assigning these costs to service lines, and ultimately, to patients.
Healthcare organization employees—clinical, administrative, and service—are an important part of the assessment equation and should be engaged and surveyed with regard to the time they spend on specific activities, as well as where they see opportunities for savings and improvement.
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, improved outcomes, and profitability. ACOs that take these critical steps will be more nimble in the era of healthcare reform, making better resource allocation decisions that lead to better financial results.
Harry Greenspun, MD,is senior advisor, Deloitte Center for Health Solutions, Washington, D.C.;
William Bercik is director of healthcare, Oracle, Atlanta, and a member of HFMA’s Georgia Chapter (firstname.lastname@example.org).
Transitioning to Activity-Based Costing
Many healthcare organizations use standard costing tools dating back to an assembly line model, in part because they are simple to implement, manage, and maintain. A June 2011 HFMA study, Value in Health Care: Current State and Future Directions, reported that 69 percent of healthcare organizations are still using legacy methodologies. These standard costing models are not well suited for capturing the complex cost and service structure of modern healthcare organizations, especially ACO models, which assume and manage greater financial risk as part of the model. The problem is that the legacy methodologies yield rudimentary cost allocations that fail to match resources and related costs to their services, making it virtually impossible to make informed decisions that will improve financial results and meet ACO targets.
For example, under legacy methodologies, hospitals often allocate the cost of equipment to surgical procedures that do not necessarily use it, because they cannot capture the detail required for such precise allocation. These hospitals also are unable to account for cost of internal support services, such as registration and materials managed, except as an allocation to a charge on the chargemaster. Many costs that are triggered by the requirements of customers or payers are allocated instead to services.
A well-defined, organizationwide program to build cost transparency using activity-based costing (ABC) can identify areas where costs are not well understood or are unproductive or excessive.
Narrowing the definition of the cost objective adds granularity to the cost information presented, but it also increases the time and expense of collecting the data.incremental approach to adopting ABC therefore may prove most feasible. Healthcare organizations should consider the costs and benefits of initially adopting the simpler ratio of cost to charge (RCC) method, looking at more directional cost, and then moving on to a relative value unit (RVU) method and payment for a unit of service, before advancing to full implementation of ABC to determine the impact of specific performance improvement activities.
RCC is commonly used because it is easy to implement, maintain, and understand. Under RCC, charge amount or charge code level is the measure by which costs are disseminated. Typically, when cost models are initiated, all revenue-producing departments can be quickly set up with the RCC method.
Hospitals using RCC methods eventually strive to implement the RVU method, which can be set up to stimulate the standard costing approach. The RVU and RCC methods are similar in that both employ weights to determine proper allocation. RVUs, however, can use measures other than charge amount to determine resource consumption. The RVU process is more complex than the RCC method, and it requires more maintenance. However, it produces more reliable and predictable cost estimates.
Ultimately, the more sophisticated data inputs of the ABC approach can be satisfied by using a variation of the RVU approach.
Publication Date: Friday, February 01, 2013