Multiple stakeholders are interested in improving the value of health care in order to achieve the Triple Aimii of better health, better quality, and lower costs. In particular, stakeholders are focusing on controlling health care costs—from Federal, State, and local policymakers to large employers paying high health care premiums for their employees, to consumers paying increasing out-of-pocket costs in high-deductible plans, plans with higher levels of co-insurance, and plans with fewer covered services. This environment has led to increasing interest in price transparency—access to information on variations in cost across health care providers—and innovative policies to improve efficiency and value.
In this context, there is a call for more comprehensive datasets to enable price transparency, improve quality, and assess the effects of health care innovations. The overall momentum to improve value by improving quality and controlling costs has resulted in State- and national-level interest to create and maintain all-payer claims databases (APCDs). These are large-scale databases that systematically collect medical claims, facility claims, pharmacy claims, and dental claims (typically, but not always), as well as eligibility and provider files from private and public payers.1
Some States have also established consumer-facing Web sites that support price transparency for consumers. Interest is growing in how to leverage APCD data for consumer-facing Web sites for price transparency, as well as for efforts to understand new innovations in care such as accountable care organizations.
The focus of this literature review and environmental scan was to map an approach to creating an inventory of measures of quality, cost, and utilization of care across settings for potential use with an APCD, noting potential and existing gaps and current barriers to APCD measurement.
To meet this objective, the project team explored various conceptual frameworks that can help interested parties choose measures for use with an APCD, enumerated potential sources of measures in the environment and in the published and grey literature, and examined gaps and potential barriers to measure use.
This report focuses primarily on ambulatory care, both because much less information is available about cost and quality in this setting and because it will leverage the unique outpatient data available in APCDs for the purpose of finding measures that may improve care and decrease costs.
Over the past decade, a growing number of States have adopted APCDs to meet the critical information needs of State agencies, inform health care and payment reform initiatives, and support price transparency initiatives to meet the needs of consumers, purchasers, and State agency reform efforts. Currently, 12 States have legislation mandating the creation and use of an APCD, with more than 30 States maintaining, developing, or having a strong interest in developing an APCD. Seven States have public reporting Web sites with cost and quality information either wholly or in part coming from APCD data.1
Multi-payer claims databases (MPCDs), which were included in the literature review in research papers that were potential sources of APCD measures, generally contain the same data elements as APCDs but do not cover all payers.
Because APCDs can be useful for improving quality of care, policymakers are interested in using them in specific ways. APCDs create a more comprehensive picture of outpatient and pharmacy care than is otherwise available in most States, including at the population level.3 These data are potentially quite powerful, as some of the best opportunities to improve care are in chronic disease management, before a hospitalization occurs, which often requires careful outpatient monitoring and adherence to drug regimens. Conceptually, then, APCDs could be used to improve patient outcomes, prevent hospitalizations, and reduce costs.
The other major impetus for APCD development is price transparency:
- To help contain health care costs, allowing purchasers to negotiate with providers more effectively;
- To allow providers to compare themselves with others in efforts to improve quality and value in shared risk and accountability payment models;
- To inform consumers’ health care decisions as they assume greater financial responsibility; and
- To address significant price variation in the system.4
If APCD data are used effectively for these goals, more efficient health care may be achieved, maximizing value with more high-quality, lower cost care choices.
In addition, local data aggregated into APCDs can be used to understand local market functioning and assess whether spending variations reflect pricing, utilization, or both. Understanding these patterns is important because multiple factors contribute to rising health care spending and differ across communities. These may include, for example, provider culture and supply, payer mix, regulation, and competitiveness of local markets.5
The impact of each factor may vary by market segment (e.g., outpatient, inpatient, home care, long-term care). Regional variation in spending in commercial insurance markets is due in large part to differences in markups by providers, but differences in utilization have been shown to explain more than 30 percent of regional variation in spending.6 Furthermore, Chernew, et al., found that the drivers of commercial spending are not correlated with Medicare spending across hospital referral regions.7
APCDs could also be used to engage local stakeholders in the often difficult tasks of managing the function of local markets. These tasks may include helping clinician leaders and others identify clinical areas of over- or underutilization or allowing regulators to identify geographic areas where unusual pricing patterns may be occurring.6
In addition, APCDs can provide data to help States develop strategic plans for public health legislation or to determine the impact of policy changes at the State level.8 The comprehensive nature of the data allows estimation of disease prevalence across a population, identification of utilization patterns and potential areas for targeted interventions, and planning and evaluation of health reform programs and legislation on cost, quality, and access to care.
Finally, APCDs can support research that may be of interest to State policymakers, such as comparative effectiveness studies or the development and evaluation of targeted interventions to improve chronic disease care.
APCDs represent a new approach to providing information about care. Before the advent of APCDs, datasets tended to be limited to certain populations or to particular components of care. For example, Medicare and Medicaid claims cover important populations but leave out a large portion of the overall population. Similarly, hospital discharge abstract databases cover only inpatient care. It is also difficult or expensive to follow patients longitudinally in these databases.
A few databases have been created for research or reporting purposes on a national level (e.g., Medical Expenditure Panel Survey). However, these are only samples of a small percentage of the patients in any given locality, so they cannot be used to make precise statements about care at the State or regional level, much less for individual providers or for specific populations (e.g., those with a particular disease).3
The idea behind APCDs is to address these limitations, at least on a State or regional level. Collecting data from all payers includes all patients, and care is captured not just in the inpatient setting, but all settings. This approach facilitates having adequate sample sizes to make precise statements about patterns of care in small areas or for individual providers. In addition, patients can, at least in theory, be followed over time, even if they change providers or payers. Using APCDs could facilitate, for instance, assessing how often a patient receives care from the same provider over time (continuity of care) and other previously unmeasurable but potentially important aspects of care.9
Lastly, APCD data usually include useful data on commercially insured patients that are otherwise difficult to access, medication use in the outpatient setting, and patient payments. Medication information, integrated with patient-level medical claims, can be used to assess performance on process measures of care, showing whether patients are getting the medications they need for specific diagnoses (e.g., antithrombotic prescriptions for those with cardiovascular disease). Patient payment information is needed for consumer price transparency efforts, and APCD data are a feasible source for this often difficult to obtain information.
While APCDs offer a number of advantages over other databases, like any data, they also have limitations. We delineate some of these more fully below in the section “Problems With Data Quality, Standardization, and Access and Strategies for Overcoming Them.”
Some notable limitations include:
- Lack of data on certain populations (uninsured patients who, because they pay entirely out of pocket, do not have claims; some behavioral and mental health populations; HIV patients; worker’s compensation patients; Tricare or Veteran’s Affairs patients, Federal Employees Health Benefit Plan patients, and Indian Health Services patients).
- Lack of access to data due to feasibility of gathering them from certain sources (e.g., small private insurers who do not meet minimum data thresholds, staff model health maintenance organizations [HMOs], pension plans [Employee Retirement Income Security Act, or ERISA], electronic health records, health insurance exchange plans, public health data, and aspects of hospital care that are part of a bundled payment, such as specific medications given).
- Lack of clinical detail (e.g., laboratory values, biometric details).
To set the stage for the rest of the report, the definitions of a few key terms as they are defined in the literature on APCDs and in the environmental scan are described below. Given the very technical nature of APCD data and variations across databases, defining a common vocabulary is important rather than defining specific variables that might be found in an APCD data dictionary.
- Charges: Amount of money a provider would seek unless another amount has previously been negotiated. This amount is often charged to patients without health insurance. Health plans typically negotiate the charge down to the allowed amount on behalf of their members.10
- Allowed amount: Maximum amount a health plan will pay for a covered health care service. Beneficiaries may have to pay out of pocket for some or all of the amounts not covered in the allowed amount. The allowed amount is sometimes also called the “cost” of care.10,11
- Cost: Sometimes used interchangeably with allowed amount but can often have a broader definition that includes out-of-pocket payments for the consumer.10
- Out-of-pocket payments: Amount paid by the consumer for care. The same unit of care may have very different out-of-pocket payments for different consumers depending on whether the patient is insured and the benefits design for the insurance the patient has (including what is covered and what the levels of copay and deductible are). Some organizations also call this the “price” of care.4,10
- Value: Relationship between cost and quality of care, with higher value providers delivering lower cost, higher quality care. Value and efficiency are sometimes used interchangeably.10
- Episode of care: A series of temporally contiguous health care services related to the treatment of a given bout of illness or provided in response to a specific request by the patient or other relevant entity.12,13
- Cost measure: A financial measure of cost, charge, reimbursement, payment, or out-of-pocket expenses associated with a visit to a health care provider or facility.2
- Resource use measures: A general term for utilization of health care services. These measures reflect the amount or cost of resources used to create a specific product of the health care system, which could be a visit or procedure, all services related to a condition, all services during a period of time, or a health outcome.14
- Relative resource use measures: Refers to a specific set of measures from NCQA that quantifies relative resource use across providers such as physician visits, hospital stays, and other resources to care for patients having one of five chronic diseases (cardiovascular disease, chronic obstructive pulmonary disease, diabetes mellitus, hypertension, or asthma).
- Current Procedural Terminology (CPT) codes: Numeric codes used to refer to services rendered on a billing claim. CPT codes are licensed by the American Medical Association. They are similar to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes but describe services rather than diagnoses. CPT codes include codes for evaluation and management, medical visits, anesthesia, surgery, laboratory tests, and radiology services. Thousands of CPT codes refer to all aspects of medical care delivery, providing a common codebook nationwide for referring to these services.15 CPT codes are sometimes used as the basis for public reports of cost, with reports describing cost data for individual CPT codes.
- Consumers: Any actual or potential recipient of health care services and their families or advocates who act on their behalf.2
- Purchaser: An individual or organization that buys health care services. The Nation’s largest health care purchaser is Medicare, which spent $425 billion on health care services in 2007.16 A purchaser might also be a large self-insuring employer such as Wal-Mart.
- Professional fees: Separate fees for the physician components of care delivered.
- Facility fees: Fees for the facility (e.g., hospital or ambulatory care center), outside the fees for billable physician components of care. Hospital-owned physician practices may have a facility fee as part of their affiliation with the hospital.17
- Public reports: Online public reports of cost, quality, or utilization that report comparative provider-level metrics and that are often interactive and consumer facing. Examples include reporting on the CalQualityCare.org Web site and APCD public reporting Web sites such as CO Medical Price Compare. For this report, the phrase does not include the broader category of public reports compiled for State policymakers and released publicly in static form.
The team used several guiding questions to focus the search strategy and data collection. The guiding questions, which speak to the objectives of the report, follow:
- What measures or outcomes (quality, utilization, safety, price, etc.) that leverage the unique data in APCDs have been reported in the scholarly literature or in online public reports using APCD data?
- What measures or outcomes have been proposed for use with APCD data or claims data that are episode based or longitudinal in nature?
- What important measure gaps have been noted in relation to transparency initiatives? Have APCD-specific measure concepts been proposed to fill these gaps (even if no fully specified measures yet exist)?
- What potential barriers to using and reporting measures with APCD data have been identified in the peer-reviewed or grey literature, including issues around availability and access to data elements? What strategies for overcoming these barriers have been proposed in the literature?
- What are some of the methodological considerations pertaining to using APCD data for measurement that have been discussed in the peer-reviewed or grey literature?
These questions reflect our findings from a preliminary review of the literature and environmental scan, as well as feedback on AHRQ’s priorities from the Task Order Officer.
Because this is the first step in creating a measure inventory, the literature on creating an APCD, creating a public reporting Web site, or getting people to use a public reporting Web site are not covered here, except to the extent that the literature is relevant to finding and prioritizing APCD measures.
ii Visit the Institute for Healthcare Improvement Triple Aim Web page for more information.