Tiering Physicians and Steering Patients - Discussion Paper
The U.S. healthcare financing system is broken. In the U.S, healthcare expenditures in 2005 were $1.7 trillion with annual healthcare costs growing by double digits and over 46 million Americans were uninsured. What are healthcare financiers (employers and payers) doing to quell the healthcare costs trend? The answer to this question in the 1980s was the implementation of managed care which focused on aggressive contracting and limiting provider (hospital, physician, and other health care professional) choice. These activities experienced public backlash in the 1990s and payers responded by broadening provider networks and offering more choice. This has caused challenges to payers to contain growth in health care costs.
Now, one approach to slow rising healthcare cost is to treat healthcare more as a commodity, and to more actively engage consumers by providing them with choices and more financial obligations in the form of higher premiums, less health benefits, and greater out-of-pocket costs for health services. This shift in responsibility and engagement from the employer to employee is referred to as “consumer directed healthcare”. The basic concept is to provide consumers with tools upon which they can make/direct their own healthcare decisions as consumers’ financial responsibility for healthcare expenses increases. The health insurance industry and Government refer to this as “transparency” in healthcare cost and quality with the aim of providing a “value-driven” healthcare system.
The Government has been advocating for transparency in quality and efficiency in healthcare, which includes public reporting of physician performance. This focus was highlighted in President Bush’s Executive Order, “Promoting Quality and Efficiency Health Care in Federal Government Administered or Sponsored Health Care Programs” in August 2006, which includes data aggregation and public reporting of physician performance.
One strategy to deliver transparency is for health plans to rate physicians’ performance in terms of the cost and/or quality of the care and to share these ratings with their insureds and/or the public consumers. This is referred to as “physician profiling.”
Physician profiling is nothing new. In the past, it was primarily driven by cost and has also been referred to as “economic credentialing.” High-cost physicians were terminated from payer networks, which generated resistance from patients, physicians and physicians’ associations. More recently, payers apply sophisticated analytic software programs to physicians’ claims (administrative data) to rate cost-efficiency and/or quality. Based on these ratings, physicians are “tiered” and the insured are incentivized or “steered” to those designated as lower cost, higher-quality physicians through benefit design structures, e.g. lower co-payments or higher co-insurance levels.
One main difference today with physician profiling is the payers’ willingness to not only report physicians’ performance in the aggregate but to share actionable, meaningful patient-specific data with physicians on which physicians can empirically identify how their practice pattern deviates from peers and what they can do to improve. This has the potential for physicians to provide safer, higher-quality care.
What no one wants is for there to be physicians who want to do a hysterectomy simply because a woman still has her uterus, believe HbA1c’s are unimportant, think a blood pressure of 145/95 is acceptable, prescribes a third beta blocker for a patient, has a generic prescription rate of 14% because they like the perks extended them from the pharmaceutical representatives, or schedules patients for tubes and/or a tonsillectomy because they snored for a week. One way to determine these variations in the delivery of care are for payers to provide peer performance reports using consensus-based clinical performance measures. We all want the best care, which includes paying attention to patient safety.
Tiered networks are the stratification of physicians and other providers into two or more groups. Differences in the costs of care between physicians and their peers in the same specialty are typically the primary determinant used to define the tiers. Rating physician quality performance is more challenging, but is emerging in payers’ physician performance programs. Payers use evidence-based medicine guidelines embedded in software which analyzes physician claims data to determine a physician’s quality rating when compared to peers in the same specialty. There are significant limitations in using only claims data as the treatment’s clinical outcomes are not included in the performance assessment. This rating may then be incorporated into a pay-for-performance bonus program.
The tiered network pricing strategy in health benefits designs is similar to pharmacy benefit plan formulary tiers: scaling out-of-pocket costs from low to high for tier 1 generic drugs to tier 3 or 4 brand, non-formulary drugs. These networks are smaller, concentric networks within larger networks referred to as “narrow/high-performance/ select” networks and aim to support value-based purchasing of healthcare services. For purposes of this paper, the term used to describe these networks are “select networks.”
Primary Care Physicians (PCPs) are typically excluded from select networks because PCPs’ (a) claims volume for any certain class of patients does not meet statistical sufficiency to perform a valid performance assessment; (b) data integrity is often lacking (e.g. data for patients not under the care of the physician attributed, patient gender should have excluded patient from denominator for gender-specific measure, etc.); (c) plans want to avoid possible public discredit; and (d) health plan executives said they want to avoid disrupting patients’ established relationship with their PCP - a situation that occurs less frequently with sub-specialists.1 Even though PCPs are not typically tiered, some payers are identifying/designating PCPs as high-quality and/or efficient in the plans’ physician directories in the spirit of transparency in cost and quality of care information.
Ultimate steering is done through the creation of a smaller, concentric network within a payer’s larger network - referred to as a narrow/select network because insureds do not have any coverage for services outside of the network. Thus, select networks are usually not distinct benefit plans, but rather an option for payers to apply across varying benefit plan platforms. Conceptually, this benefit plan design is the same as an HMO in terms of cost containment, minus the referral requirements. Employees are attracted to enroll in a select network benefit plan due to lower premiums and out-of-pocket costs. The number of these networks developed for employers who have demanded them are limited and offered in only few markets.
Tiered and Select Networks Concerns
A risk associated with tiered networks is whether the information used to differentiate providers is accurate. The potential for disrupting patients’ medical homes, threatening the economic survival of family physicians, and unjustly causing public discredit to physician reputations based on unreliable claims data is of significant concern. Patients may incorrectly interpret tier differentials to mean that physicians in higher-cost tiers are of poor quality, or that they misuse health care resources. Tiered networks may exclude physicians based solely on insufficient claims data for a single insurer, when these same physicians may meet the quality benchmarks and cost targets when measured across all payers. Patients may have limited access to sub-specialists creating access issues. If PCPs are tiered, access issues for enrollees living in rural/frontier areas could be significant. The argument for applying such programs to rural areas is that patients should receive care from the best physician possible, and driving to the city may be necessary.
Physician Profiling Issues
No process to assess physician performance is perfect. Payers are using episodes of care “grouper” software to profile physicians utilizing claims data to assess physician efficiency. “Grouper software” is so named because it groups a patient’s claims that are related to a given diagnosis over a set period of time (referred to as the “episode of care”.) All costs attributed to the diagnosis, including outpatient, inpatient, radiology, lab, and prescription drugs are totaled for a specified continuous period of time, or a series of independent intervals of care for a certain diagnosis using claims data. Expense lines from claims processed are assigned to and evaluated using certain diagnoses. Note, the payers may not have all of the claims data for lab and pharmacy encounters due to how lab tests are billed and the situation where an employer has “carved-out” the pharmacy benefit plan to a separate entity. This weakens the accuracy and robustness of the lab and pharmacy data when comparing physicians’ performance.
Another deficiency in the grouper methodology is that it will not reflect the number of episodes of care that may have been prevented by the comprehensive primary care delivered by the family physician. For example, if an intervention by the primary care physician prevents the complications of diabetes or the onset of congestive heart failure, the impact of such an intervention may not be measured by this methodology. A more detailed performance analysis explanation and list of concerns are available in Attachment 2.
Assessing physician quality performance is sometimes done in concert with a plans’ profiling assessment to designate physicians as high-quality, which is much more challenging than rating efficiency when plans utilize claims data to assess clinical outcomes. Payers use software that applies evidence-based medical guidelines and consensus-based quality standards to physicians to determine their adherence to such performance measures. Thus, the grouper software looks at costs of care and not quality of care. Additional software analysis is used for quality profiling. The same concerns with rating cost-efficiency being delivered exist with physician quality of care ratings.
Tiered networks are expected to constrain overall health plan costs by providing a foundation upon which health plans provide financial incentives or “steer” insureds/patients to lower-cost providers and place pressure on physicians to improve their efficiency, or to accept discounted payment rates in order to be placed in the preferred tier. This strategy is referred to as “tiering and steering” as physicians are tiered and insureds are steered using financial incentives within the benefit design.
Steering may take several forms within a payer’s network.
- Lower co-pays for a primary physician’s office versus other settings, e.g. sub-specialist office, urgent care, or emergency room.
- Lower premiums or reduced levels of cost-sharing (i.e., co-payments or deductibles) when using payer’s preferred/designated physicians.
Plans typically create co-payment differentials from 1.5 to 2.0 times the in-network physicians versus the high-performing physicians. For instance, the co-payment for a high-performing physician would be $20 compared to $30 or $40 for other in-network physicians. In some cases, payers apply the PCP co-payment amount for sub-specialists in their select networks available.
Presently, payment rates for physicians in select networks are the same as physicians in the payers’ larger networks. However, it stands to reason that eventually physicians designated as high-performing will be financially rewarded in bonuses or an increase in their fee schedules. More than likely, this will occur as assessing quality performance becomes more robust leading to a growth in pay-for-performance programs.
Currently, health plans believe that through designating physicians as high-performers and steering enrollees to them, these physicians will be able to at least maintain their market share and perhaps, increase their volume of patients. There is concern from the medical community that the preferred physicians may not be able to handle the increased patient volume incentivized to seek care from them. Undoubtedly, most PCPs are already at their patient panel maximums and are not prepared to accept more patients. Rather, they want to be able to spend more time in direct patient care and in care coordination activities.
For the bottom 10-15% physicians, there may come a time where they are either pressed to answer the payer’s question of why they should be included in payers’ network. Some health plans are sharing actionable data upon which physicians have a better opportunity to improve their patient care. Another strategy to thin the networks may be for plans to offer only aggressive discounted physician fees making it extremely challenging for a non-designated physician to be financially successful.
Physicians want to and believe that they provide the best patient care. Their sentinel concerns with health plans performance assessments include: use of sufficient sample size; data integrity; case-mix adjustments; physician attribution; use of consensus-based performance measures; automatically excluding physicians based upon the use of standard deviations; and timeliness of the data included in the assessment. Aside from the analysis activity itself, physicians frequently complain about health plans’ dearth in sharing information about the tiered networks existence, including the physicians’ designations, their comparison to their peers, and details about the performance assessment using the grouper methodology – opening the proverbial “black box.” Physicians are also concerned that many plans do not have a formal and fair process for physicians to appeal their designations prior to their release.
With respect to consumers’ health literacy and ability to assess physician performance reports, physicians are concerned that consumers cannot make sound medical decisions. For instance, how will they be able to determine when it is appropriate, from whom, and where to seek care?
Although a limited number of large, national self-insured employers and health benefits consultants have driven the creation of select networks, not all are prepared to deal with potential employee backlash. Consequently, actual adoption of these networks has not yet matched interest. Employers that have pushed these products aggressively reported feeling as if they are out there alone and everyone else is “window shopping” as they wait and observe the experiences of early adopters. More than likely, the adoption of select networks might be higher if employers’ premium savings were greater than the estimated 3 to 5 percent.1 Sharing cost of care and quality information for physicians and other health care professionals is accelerating and principally being displayed in the payers’ online physician directories.
Employers are grappling with how to deal with variance in how plans build their select networks and how to have a consistent health benefit strategy with similar health benefits across multiple states. Further, employers are concerned about employee access. For example, larger employers with employees who are located across a wide area are concerned that high-performing providers may not be evenly dispersed. Thus, they are less likely to include these networks in their health benefit strategy. However, employers are interested in using benefit design in tiered networks to provide employees broad network access and provider choice unlike the narrower, select networks. Smaller employers, who are more price-sensitive and less deterred by such issues as network disruption, may find the select networks more appealing due to their affordability.1
In the beginning, payers only utilized crude costs of care analysis to create select networks. They quickly learned that tiered networks must be based on costs and quality of care as value cannot be defined by costs or quality alone. Who wants to receive care from the lowest cost physician without any idea of that physician’s quality? The goal is to have the most efficient use of health services to reach the desired health outcomes.
Health plans have realized that their physician designation efforts will not be successful if the provider community is not engaged in the development and implementation processes. Just as payers are being transparent with the results of their physician designations to consumers, they must be transparent in the physician performance analysis and provide actionable data upon which improvement can be actualized. These educational efforts should benefit not only the physician, plan, and employer but the most important party - the patient.
Data aggregation across multiple payers – both public and private – is essential in providing statistically significant data upon which a robust analysis can be performed. Also, payers must use consensus-based quality and efficiency performance measures so that the analysis is consistent across payers, e.g. Ambulatory care Quality Alliance (AQA) measures. However, in the litigious environment anti-trust concerns arise when collaboration across payers exists. However, the problem with a physician being rated by Payer A as high-quality and efficient and not designated identically by Payer B, creates patient confusion and physician concern.
Payers financial success with physician performance reporting is largely dependent how employers structure their health benefit plans to steer patients to the lower-cost, higher-quality designated physicians. If done correctly, the less efficient physicians will lose patients to the more efficient ones reducing health care expenses. This shift should serve as a motivator for the lowering performing physicians to actively engage in quality improvement activities and attend to cost-effective treatment plans.
Accurately rating physician performance is important, but to date the methodologies used have contained many flaws. Until a better method is developed, efforts to reduce healthcare costs will continue to drive payers to utilize episode grouper and/or evidence-based medicine software methodologies to assess and report physician performance. Physicians will be “tiered” based on these physician ratings, and insureds will be ”steered” to less expensive, supposed higher-quality physicians or find themselves in a select network.
Plans will continue to increase their cost containment strategy to develop select networks, provide transparency in physician quality and efficiency and steer patients to these designated physicians. It is advisable for practices to evaluate their practice using its own formal performance measurement process, and to be ready to engage in continuous performance improvement. It may go without saying, but effective use of a fully-functional electronic health record system will assist in this effort of identifying recommended care for physicians’ panel of patients
AAFP related policies and positions
Performance Measures Criteria
The sentinel New England Journal of Medicine article by Elizabeth McGlynn in June 2003 revealed that patients only received the evidence-based recommended care 55% of the time. This quality gap must be addressed by the physician community. Employers and payers cannot wait for medicine to respond, which has led them to develop and utilize available technology to assess physician performance.
To rate physician performance, payers take claims data and apply proprietary, analytic software to analyze claims data to determine physician performance. Some payers combine a type of quality assessment in tandem with rating cost of care for a defined set of diagnoses. It is important to define what payers refer to as “efficiency”. Some refer to their programs as ones which use efficiency measures, when in actuality the measures are only a reflection of physician-attributed costs of care. The Ambulatory Quality Alliance (AQA) has three (3) definitions of costs of care related to performance measures for AQA:
- “Cost of care” is a measure of the total health care spending, including total resource use and unit price(s), by payor or consumer, for a health care service or group of health care services, associated with a specified patient population, time period, and unit(s) of clinical accountability.
- “Efficiency of care” is a measure of cost of care associated with a specified level of quality of care. “Efficiency of care” is a measure of the relationship of the cost of care associated with a specific level of performance measured with respect to the other five IOM aims of quality.
- “Value of care” is a measure of specified stakeholder’s (such as an individual patient’s, consumer organizations’, payers’, providers’, governments’, or society’s) preference-weighted assessment of a particular combination of quality and cost of care performance.
Episode Grouper Overview
The methodology used to profile physicians on costs is referred to as “grouper methodology.” It is provided by three (3) main organizations: Ingenix’s Episode Treatment Groupers (ETGs), CAVE Consulting Grouper (CCG), and MedStat Episode Groups (MEG). The approach of these entities is to sell their software to payers, which can integrate and customize it for use in their claims system to assess physicians’ cost-efficiency. Of note, the Center for Medicare and Medicaid Services (CMS) is testing and assessing how the grouper methodology could be used to rate physician performance for Medicare beneficiaries enrolled in traditional Medicare. This is part of the industry push to support transparency in health care by sharing physician and hospital quality and efficiency information.
Grouper methodology software includes a table of diagnosis and procedure codes, with a diagnosis being associated with certain procedure codes. For example, the Symmetry Episode Treatment Groups® (ETGs®) combines information from inpatient, outpatient, ancillary and pharmacy claims data as inputs, captures data on the relevant services during a patient's treatment, and "groups" claims information into meaningful, defined episodes of care. The ETGs combine claims to define a comprehensive treatment episode from the onset of care until treatment is complete. The duration of each episode is based on a continuous time period in which there is an absence of treatment, called a "clean period".2
The basis of the clinical logic for the ETG methodology is a series of diagnosis and procedure code "eligibility" tables. Each and every ICD-9 (diagnosis), CPT-4 (procedure), NDC (drug) and HCPCS (procedure) code has been mapped to each of the ETGs. As each claim record is considered by the ETG grouper, the procedure and diagnosis codes are evaluated with respect to each other and in turn to each of the eligibility tables. Those "matches" are then considered with respect to the ETGs3.
Following this, relative weights are determined based on gross charges per episode, resulting in the ETG costs per episode of care. A cost of care per physician who provided services during the episode of care is rated, with the physician delivering most of the care being attributed/assigned these costs. This mathematical calculation becomes the unit of analysis upon which physician performance is compared.
When claims data cannot be clearly attributed to a physician, the data are not included in the episode of care treatment for the patient’s identified treatment group. However, attribution rules are difficult to define.
An actual episode cost is calculated by adding together the allowed payment amounts for all claims included in a specific episode (e.g., hospital, physician, pharmacy, etc.). An expected or average cost of an episode is then calculated based on the average actual cost of all episodes of the same type. This average cost of care is used as the benchmark against which each physician in the same specialty in the same market is typically compared to their peers in a pass or fail comparison. Statistically, such a comparison will mathematically cause a certain percentage of physicians to fail, even though they may be near the average market mean.
For payers that rate physicians across quality performance measures, physicians’ claims data is compared to a set of consensus performance measures (NCQA, HEDIS, RAND, AQA, Physician Consortium) with a benchmark performance being set, i.e. 70% across all performance measures. It is clear that payers do not want to invent their own performance measures; rather, they want to adopt consensus-based performance measures in their data analysis. By doing this, not only are the performance measures less controversial (because physicians are being rated using evidence-based, scientifically valid performance measures), but the payers also avoid the costs of developing and applying their own independent measures.
For any grouper performance analysis to be meaningful, there are key criteria that must be satisfied, including:
1Debra A. Draper, Allison Liebhaber, Paul B. Ginsburg, High-Performance Health Plan Networks: Early Experiences, Issue Brief No. 111, Center for Studying Health System Change, May 2007.
2Symmetry Episode Treatment Groups® (ETGs®) information from its web site, http://www.symmetry-health.com/downloads/ETG_WhitePaper.pdf, accessed on December 13, 2006.
3Symmetry Health Episode Treatment Groups white paper: A condition classification and episode building system, http://www.symmetry-health.com/downloads/ETG_WhitePaper.pdf, accessed on December 13, 2006.