Cost of Care

New Approaches for Improving the Cost of Care

August 17, 2018 3:05 pm

The requirements of a value-based healthcare system are making many providers look at their data in new ways with an eye toward cost improvement.

Cost performance in health care can be a tricky subject of discussion with providers. Most health systems have budgeting, purchasing, and financial systems in place to control the costs of the operations, and the technology deployments for monitoring those operational costs have become increasingly complex, matching costs with patient volume and demand.

Even with monitoring of operational costs, however, healthcare costs continue to rise. The problem is different definitions of costs that are driving expectations of the actions that providers should take to control them. 

When health plans, employers, and consumers refer to healthcare costs, they are defining the problem to be the cost of care. These purchasers are evaluating prices and inputs into the total cost of care that are driven not only by hospital utilization, but also by ambulatory and outpatient services. They want more efficient care and for the cost of care to be capped. Innovative approaches and technologies will be needed to help health systems and their providers step up to the plate on controlling costs.

The Changing Healthcare Landscape

 Historically, healthcare providers have had the advantage of being able to negotiate their own fees and get paid for every service they provide. If insurance won’t pay for it, in most cases they can charge the patient. Over the past few years, however, government, private health plans, and employers have laid the groundwork for paying healthcare providers fees that include at least some financial risk for costs that exceed targeted expenses.

The impact of the move to value-based care has been significant. Narrow provider networks include only lower-cost providers, cutting academic health centers out of benefit plans. Employers develop more options for narrowing provider choices, aligning or even establishing their own networks. Government payers provide differential payments to high- and low-cost providers, with the high-cost providers receiving less.

Financial risk will increase as government, employers, and health plans seek to cap their costs. It is likely that Medicare will be largely risk-based, either through increased Medicare Advantage risk plans or through risk-based provider accountable care organizations (ACOs), recently reinforced through new proposed rules for shared savings plans. a The appetite for “defined benefit” health plans is growing, with vouchers, risk plans, and health savings accounts (HSAs) to limit their cost and transfer the rest to employees. 

Finally, as patients increasingly are taking on the role of consumer, they are beginning to drive efforts to lower their costs, get price transparency, and engage in medical decision making. Many of those decisions could result in delays or decisions not to go forward with diagnostic testing or some treatments.

Providers that attract patients will be those that reorient their services, marketing, and value toward ensuring that consumers are protected from excessive costs and can make responsible decisions. 

Cost Ignorance Challenges Providers That Seek Change

Data for evaluating cost performance is virtually nonexistent among health systems, except for supply purchasing or, at best, provider-by-provider comparisons of costs that rarely incorporate adequate information about patient risks or differences. 

Health plans and government payers have robust but often one-sided cost algorithms that compare health systems and their total costs by condition or procedure and that look at practice variations among individual physicians or groups, including specialty groups, and total per patient costs generated by attributed primary care physicians. If providers are to get serious about improving cost performance and attractiveness to health plans and consumers, they require the same capabilities.

Can Providers Fill the Missing Cost Data Gap?

Providers can, indeed, access some data necessary to begin evaluating and comparing costs of care. In some cases, the data were provided to them already by Medicare, but in formats that require extra analysis. Likewise, health plan negotiations rarely include discussion about patient-level data that should be provided to the provider group to evaluate cost performance. Below are a few examples of specific reports can help provider organizations see beyond their internal systems and begin evaluating cost data. 

External data: data files and reports from Medicare and health plans. Annually and at other intervals, the Centers for Medicare & Medicaid Services (CMS) has released reports and data that show healthcare providers where they stand compared with other providers. These data, including patient-specific cases could be used by health systems to examine variations in cost. Most of these data are provided at the tax identification number (TIN) level rather than at the provider level, and  can provide both general and and detailed information about costs relative to specific episodes of care. Health systems also can integrate CMS’s patient data with their own source system data to take a more detailed look by provider. 

Unfortunately, in mid-2018, CMS announced that it would not be providing detailed data files for the previous year, leaving providers no information to interpret their relative cost scores or to take action. The agency left open whether future data would be forthcoming. Providers should still be able to access prior data and analyze their results.

Receipt of non-Medicare claims data or reports should be part of negotiations with each health plan. Identified patient claims data can be more difficult to obtain unless the health system has negotiated a commercial ACO arrangement with the health plan, but even de-identified data and aggregate reports can be useful for providers to diagnose areas of cost outliers that are seen in the industry.  

An ACO will receive the full claims history of its patients. Integration of these data with source system data will create a rich database that can be used to analyze costs, service usage, and out-of-system referrals. 

Internal health system data. Healthcare providers can develop their own system data for comparing costs within and between specialties, diagnoses, and types of care. Ultimately, they can be useful for making physician comparisons, but physicians should be careful to validate detailed information and provide feedback on a case level. The reports can be created if the systems are patient-centric or can be integrated to identify patients across multiple information systems, and if structured data can be captured to the extent necessary. These reports also may help healthcare leaders identify the extent of work to do on the systems, operations, and care programs to fuel future analyses.

This information creates the baseline for the most important analyses of cost: overall cost of care by condition and procedure, leakage from the provider system to outside services (and why), and generation of costs for technology and other key cost drivers.

Cost Performance Measurement and Improvement: Technology Requirements

Information systems in use by healthcare organizations to address cost are designed to manage operational costs and budget tracking, identify revenue cycle and accounting issues, and in some cases develop targets for negotiations with health plans. Although essential to the business of health care and ensuring survival, these systems are distinct from systems that can measure the costs of care relative to benchmarks or can assess variations in care costs. And these systems are not equipped to create cost measures or manage performance improvement. 

To address the cost of care, existing systems must fuel targeted technology that aggregates internal and external data sources to measure cost and manage (and test) multiple interventions to improve the cost of care. The systems embed algorithms and logic to establish medical cost measures by patient, condition, treatment, and specialty or provider. To best support efforts to modify the cost of care, this technology must enable the organization to evaluate the inputs of all services into each patient’s care and associate the resulting costs with the clinical outcomes of those patients.

These performance measurement and improvement tasks are well beyond the core competencies of most health systems. Even the largest and most sophisticated systems have well-designed, internally focused operational cost monitoring but virtually no technology to address medical costs from a purchaser’s perspective. For example, market comparison of costs for specialty services is often episodic, but health systems lack the ability to integrate disparate provider billings and organize services into episodes of care. Healthcare organizations must focus first on measuring, then on improving, performance.

Measures of cost performance are key indicators, and until recently, there have been only a few actual cost performance measures in use. These measures include the percentage of readmissions, admissions for ambulatory care sensitive conditions, and cost per attributed patient, all of which have been part of Medicare’s Value-Based Modifier and are now scored under the Merit-based Incentive Payment System of the Medicare Access and CHIP Reauthorization Act (MACRA). Also emerging as cost measures are total costs of episodes of medical and surgical care. 

Existing cost -related technologies used by healthcare organizations are insufficient to explain the variation of patient populations and car specialties. They do not address the fundamental drivers of the cost of care: medical decisions made by providers and patients. One of the most important contributions providers can make is to develop more accurate measures of cost and methods of accounting for variations.

As with quality measures, there are flaws when such cost measures are used as scores or benchmarks for care. There are hundreds of reasons why variation can and should occur. The task of determining why variation occurs should be part of the measurement process, and feedback from clinicians is necessary to ascertain why. Higher cost in individual cases often is completely justifiable. High aggregate costs across all providers or individuals should raise questions. But judgement without feedback and further analysis of patient circumstances invariably is premature. 

It is necessary to adjust for risk, evaluate care by different diagnoses or patient attributes, and permit feedback at the patient level of data. A clinical data registry (CDR) will perform these functions.

Qualified clinical data registries (QCDRs) are clinical data registries approved by Medicare under MACRA to evaluate long-term outcomes and provide performance improvement capabilities to providers. QCDRs have patient-identified data aggregation capabilities, the capacity to develop and use cost measures, and the functionality for including physician or other feedback to explain the measure results. However, not all QCDRs have these capabilities, and providers should investigate functionality along with expertise in performance improvement for these functions.

Cost performance technology also should be able both to distinguish measures for primary care and specialty physicians and to integrate source data with external health plan and Medicare data.

In addition to cost measures, episodes of care can be constructed to evaluate the types and costs of services that go into major service lines, such as diabetes care, and major procedures, such as percutaneous transluminal coronary angioplasty. The technology should be able to compile and integrate multiple data sources (including clinical and financial systems as well as claims), attribute care type, and create analyses of variation among the episodes by specialty and provider. 

At a minimum, the episodes of care in current and proposed use by Medicare are a start for internal provider review. The point is not only to validate Medicare information, but also to create a knowledge base that will set the platform for lowering costs.

Improving cost performance is a more complex process that will take time and more sophisticated technology. Improvement in both quality and cost is an iterative process; the interventions for cost improvement are rarely clear cut. Improving performance should have the objective of identifying interventions that are or could be effective. 

Performance improvement technology should have the capability of distinguishing individual improvement projects and each of their interventions, so that cause and effect of interventions or changes on costs can be measured.

An important feature of most performance improvement systems is the sharing of results among providers involved, with feedback into the system’s notes functionality or other structured data fields.  This facilitates peer review and the education of providers, both of which are essential to engaging physicians on cost reduction.

Likewise, excessive costs can come from a patient not fully understanding the condition or being in agreement with the treatment plan. Shared decision making, as an intervention, is carried out by clinicians and patients to establish a plan that is then decided by the patient. The technology should generate registries of patients qualifying for the intervention according to project criteria (for example, patients with congestive heart failure who have more than one admission in six months) and track activities such as patient outreach, shared-decision-making appointments, and provider or patient feedback. Both the feedback and the results can be used to tweak the project for greater effectiveness.

Condition- and Procedure-Specific Technologies to Curtail Costs

Because chronic disease now accounts for a large portion of healthcare spending, providers must adopt more-focused programs in working with specific patient conditions. As part of their larger focus on improving outcomes providers should undertake such programs to track disease-specific outcomes improvements—for example, testing approaches that work improving hypertension control with patients having diabetes or in addressing other major chronic conditions.

Unique performance improvement technology is necessary because project management tailored to specific populations and conditions is beyond the scope of either electronic health records or even other population management technologies. 

First Steps to Innovation in Cost Performance

Health systems have grown from hospitals into hospital-plus-physician enterprises. This evolution carries with it some legacy issues that must be resolved as these organizations move forward with innovation.

Simply accessing a single list of physicians by correct business and provider identification numbers often is a challenge. The enterprise also is characterized by many different billing and clinical systems. Legacy systems also may have data integrity issues stemming from variable or incorrect usage. All these realities complicate the initial steps to improving cost performance.

Nonetheless, health systems seeking to achieve more efficient operations should take the following four steps.

First, they should choose technology that will measure cost of care through both global and episodic measures, can customize functionality for both performance measurement and improvement, and can execute and measure improvement initiatives along with their specific interventions and results. 

Second, they should focus that technology on performance measurement and improvement. The  fuel for achieving value starts with having the tools to measure performance and variations in outcomes and costs. But the execution and leadership comes from the healthcare organization itself, to implement a cycle of continuous improvement initiatives. Each initiative should be focused on patient populations that are customized on the basis of risk, condition, and other criteria, and on improvement through measurable projects and interventions.  The role of the technology is to contain such project and intervention settings, filter the data to create the population, track execution of activities, and measure the results.  This will require the integration of Medicare and other external claims data with internal systems’ patient data.

Third,  they should specifically address problem areas that are revealed in cost data, including patients whose services exceed cost parameters.  In particular, these are likely to focus on high risk patients with chronic disease, and will require the implementation of disease-specific projects or those identify  barriers and individualized care plans.

Finally, healthcare organizations should focus on customer service needs, especially regarding cost and price transparency and financial services.  

Medical costs have been a difficult area for providers to address because their systems limit the view of patients’ medical care to what they provide, but value-based health care is making those providers accountable for the medical costs of every patient attributed to them. Provider systems need to enter this new territory with initiatives and technology that will expand their vision of medical costs and how they compare with competitors. This effort will require increased data aggregation and negotiations with health plans for claims data, adoption of performance-oriented technology, and involvement of providers in performance improvement interventions. 

By addressing the issue of the cost of care and how it is reflected in pricing presented to health plans and consumers, providers can improve their competitive position within their markets and with purchasers of care. 

Theresa Hush is the CEO and co-founder of Roji Health Intelligence, Chicago.


a. Kunkel, C., Osborne, D., Dong, L.F., Polakowski, M.J., Champagne, N., and Mills, all, MACRA and Medicare Advantage Plans: Synergies and Potential Opportunities ,” Milliman, Feb. 21, 2017.


googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text1' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text2' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text3' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text4' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text5' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text6' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text7' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-leaderboard' ); } );