Self-regulation Through Appropriate Use
In 2009, the American College of Cardiology (ACC), along with other societies, updated and expanded guidelines for appropriate use of cardiac nuclear imaging with a practical document designed to facilitate sound decision-making. The goal was to optimize patient care efficiently, and by doing so, say good riddance to regulatory entities such as radiology benefits managers. Two years later, the strategy seems to be working, at least on some fronts.
In the U.S., use and spending on SPECT myocardial perfusion imaging (SPECT MPI) and other imaging services burgeoned after 2000. According to an analysis by the Government Accountability Office (GAO), Medicare spending on imaging services more than doubled between 2000 and 2006. The 2008 report called out independent cardiologists in particular, claiming that cardiologists racked up the largest share of in-office expenditures of all specialties—$3 billion of the total $14 billion spent in 2006. Proportionally, more cardiologists had skin in the game, with 43 cardiologists per 100 physicians billing for in-office imaging services in 2006, up from 24 per 100 physicians in 2000.
The report recommended that the Centers for Medicare & Medicaid Services (CMS) adopt a front-end approach, such as instituting the use of radiology benefits managers (RBMs). Its authors noted that some private health insurance companies already required prior authorization through RBMs as a way to manage utilization and costs.
Guidelines as a framework for determining what indications fit testing may help chip away at overutilization, but only if physicians know of and follow them. And that is a challenge.
"Expecting busy physicians to read guidelines and appropriate use criteria [AUC] and incorporate them into practice, never happens," says Brahmajee K. Nallamothu, MD, MPH, a cardiologist at the Center for Healthcare Outcomes and Policy at the University of Michigan in Ann Arbor. "You have to re-engineer how practices flow and how patients receive care to have an impact."
While not comprehensive, the 2005 guidelines helped cardiologists identify gaps and strategies for future directions. In a pilot study, Hendel and colleagues used the SPECT MPI recommendations as a template to track appropriate use in clinical practice. The results, based on 5,928 patients enrolled at six clinical sites between 2008 and 2009, were consistent with other AUC assessments: 71 percent of cases were appropriate; 15 percent were uncertain and 14 percent were inappropriate (J Am Coll Cardiol 2010;55:156-162). The findings allowed the researchers to identify patterns of inappropriate use and target them for improvement.
"If you eliminated just one inappropriate indication, you would drop volume by 6 percent," Hendel says, referring to what they found was the most frequent inappropriate indication: detection of coronary artery disease in asymptomatic, low-risk patients. "However, more importantly, you would get rid of about half the inappropriate studies."
The 2009 guidelines were purposely formatted to be embedded into decision-support algorithms, Hendel says. Early adapters included Northern California Kaiser Permanente in Oakland, which integrated the guidelines into its EMR to help clinicians order tests efficiently and appropriately, says Ralph G. Brindis, MD, MPH, senior adviser for cardiovascular disease at Northern California Kaiser Permanente and chair of the 2005 guidelines writing committee.
"There is no way you can manage patterns of care unless you actually measure them," Brindis says. "Data feedback looking at different centers' utilization and individual physician's utilization of tests can help change practice patterns."
In another pilot study, Hendel and colleagues tested whether use of a feedback and education process improved adherence to the 2009 AUC. The participating 23 sites entered patient cases beginning in April 2010 to establish baseline usage and then incorporated the Performance Improvement Module (PIM), a clinical support tool and an educational initiative, to their programs. A subsequent evaluation found a relative increase of appropriate use of 6 percent and overall decrease in inappropriate studies of 50 percent.
"This is a voluntary community and quality initiative aimed at the most motivated practices, but it is starting to sound like we can actually get a handle on this," Hendel says. "We are very bullish on this approach as a substitute for radiology benefits managers, which provide no feedback and no education."
The cardiology community scored several victories in 2011. In February, CMS launched a two-year demonstration project to study the impact of imaging decision support systems such as the PIM on Medicare patients' utilization and access to care. Modalities studied included SPECT MPI. Also, the Delaware insurance commissioner ruled in September that Blue Cross Blue Shield of Delaware's nuclear cardiac imaging pre-authorization process was flawed and ordered the insurer instead to use the ACC's decision support tool, FOCUS (Formation of Automatic Utilization Strategies in Cardiac Imaging) PIM. The instrument is an outgrowth of the pilot study that includes webinars and online interactive participation.
Confounding efforts is a reimbursement system in flux as it attempts to shift to rewarding quality rather than quantity. "We are skating to where we think the puck will be in terms of reimbursement," Brindis says, adding that it is a challenge to get physicians to fully embrace AUC if they lose money in the process. "We need to have financial alignment in a way that encourages the adoption of appropriate use criteria."
A recent study linked overutilization of inappropriate cardiac stress testing to the ordering physicians' financial interests (JAMA 2011;306:1993-2000). Nallamothu, co-author of the accompanying editorial, pointed out that cuts to Medicare reimbursements for imaging services between 2005 and 2010—the period following the GAO report—put the brakes on the rising rates of use in Medicare patients' tests (JAMA 2011;306:2028-2030).
The regulatory changes may have altered delivery of care, he wrote, prompting cardiologists to abandon once lucrative single specialty groups for multispecialty and hospital-based practices. But that may be passing the buck if hospitals are production oriented, possibly at the patient's expense.
"You can nuance this for the umpteenth time, but you can't regulate every incentive," Nallamothu states. "Either we forbid physicians from performing these services in offices, which may lead to problems with access, or we need to value these services more fairly and set the right price upfront."
In the U.S., use and spending on SPECT myocardial perfusion imaging (SPECT MPI) and other imaging services burgeoned after 2000. According to an analysis by the Government Accountability Office (GAO), Medicare spending on imaging services more than doubled between 2000 and 2006. The 2008 report called out independent cardiologists in particular, claiming that cardiologists racked up the largest share of in-office expenditures of all specialties—$3 billion of the total $14 billion spent in 2006. Proportionally, more cardiologists had skin in the game, with 43 cardiologists per 100 physicians billing for in-office imaging services in 2006, up from 24 per 100 physicians in 2000.
The report recommended that the Centers for Medicare & Medicaid Services (CMS) adopt a front-end approach, such as instituting the use of radiology benefits managers (RBMs). Its authors noted that some private health insurance companies already required prior authorization through RBMs as a way to manage utilization and costs.
Build it and they will comply?
Even before 2008, ACC advocated another approach: identifying and publishing guidelines on the appropriate use of SPECT MPI, and in the next iteration in 2009, other nuclear cardiology modalities. The first guidelines were merely rules and regulations, says Robert C. Hendel, MD, director of cardiac imaging and outpatient services at the University of Miami Miller School of Medicine in Florida, and chair of the 2009 guideline writing group.Guidelines as a framework for determining what indications fit testing may help chip away at overutilization, but only if physicians know of and follow them. And that is a challenge.
"Expecting busy physicians to read guidelines and appropriate use criteria [AUC] and incorporate them into practice, never happens," says Brahmajee K. Nallamothu, MD, MPH, a cardiologist at the Center for Healthcare Outcomes and Policy at the University of Michigan in Ann Arbor. "You have to re-engineer how practices flow and how patients receive care to have an impact."
While not comprehensive, the 2005 guidelines helped cardiologists identify gaps and strategies for future directions. In a pilot study, Hendel and colleagues used the SPECT MPI recommendations as a template to track appropriate use in clinical practice. The results, based on 5,928 patients enrolled at six clinical sites between 2008 and 2009, were consistent with other AUC assessments: 71 percent of cases were appropriate; 15 percent were uncertain and 14 percent were inappropriate (J Am Coll Cardiol 2010;55:156-162). The findings allowed the researchers to identify patterns of inappropriate use and target them for improvement.
"If you eliminated just one inappropriate indication, you would drop volume by 6 percent," Hendel says, referring to what they found was the most frequent inappropriate indication: detection of coronary artery disease in asymptomatic, low-risk patients. "However, more importantly, you would get rid of about half the inappropriate studies."
Feedback and focus
The pilot study also let researchers develop tracking tools and feedback mechanisms to educate physicians and possibly improve adherence. After analyzing its data and educating staff, one site reduced its inappropriate use rate from 22 percent to 13.3 percent.The 2009 guidelines were purposely formatted to be embedded into decision-support algorithms, Hendel says. Early adapters included Northern California Kaiser Permanente in Oakland, which integrated the guidelines into its EMR to help clinicians order tests efficiently and appropriately, says Ralph G. Brindis, MD, MPH, senior adviser for cardiovascular disease at Northern California Kaiser Permanente and chair of the 2005 guidelines writing committee.
"There is no way you can manage patterns of care unless you actually measure them," Brindis says. "Data feedback looking at different centers' utilization and individual physician's utilization of tests can help change practice patterns."
In another pilot study, Hendel and colleagues tested whether use of a feedback and education process improved adherence to the 2009 AUC. The participating 23 sites entered patient cases beginning in April 2010 to establish baseline usage and then incorporated the Performance Improvement Module (PIM), a clinical support tool and an educational initiative, to their programs. A subsequent evaluation found a relative increase of appropriate use of 6 percent and overall decrease in inappropriate studies of 50 percent.
"This is a voluntary community and quality initiative aimed at the most motivated practices, but it is starting to sound like we can actually get a handle on this," Hendel says. "We are very bullish on this approach as a substitute for radiology benefits managers, which provide no feedback and no education."
The cardiology community scored several victories in 2011. In February, CMS launched a two-year demonstration project to study the impact of imaging decision support systems such as the PIM on Medicare patients' utilization and access to care. Modalities studied included SPECT MPI. Also, the Delaware insurance commissioner ruled in September that Blue Cross Blue Shield of Delaware's nuclear cardiac imaging pre-authorization process was flawed and ordered the insurer instead to use the ACC's decision support tool, FOCUS (Formation of Automatic Utilization Strategies in Cardiac Imaging) PIM. The instrument is an outgrowth of the pilot study that includes webinars and online interactive participation.
Getting the price right
AUC still has a ways to go, says Brindis, who is helping to deliver the next iteration before 2013. Those guidelines will include a multimodality component that provides recommendations for choosing the best option or the best order of options based on indication. The authors also are wrestling with ways to evaluate outcomes and access to care, including underutilization and how that may contribute to racial and socioeconomic disparities.Confounding efforts is a reimbursement system in flux as it attempts to shift to rewarding quality rather than quantity. "We are skating to where we think the puck will be in terms of reimbursement," Brindis says, adding that it is a challenge to get physicians to fully embrace AUC if they lose money in the process. "We need to have financial alignment in a way that encourages the adoption of appropriate use criteria."
A recent study linked overutilization of inappropriate cardiac stress testing to the ordering physicians' financial interests (JAMA 2011;306:1993-2000). Nallamothu, co-author of the accompanying editorial, pointed out that cuts to Medicare reimbursements for imaging services between 2005 and 2010—the period following the GAO report—put the brakes on the rising rates of use in Medicare patients' tests (JAMA 2011;306:2028-2030).
The regulatory changes may have altered delivery of care, he wrote, prompting cardiologists to abandon once lucrative single specialty groups for multispecialty and hospital-based practices. But that may be passing the buck if hospitals are production oriented, possibly at the patient's expense.
"You can nuance this for the umpteenth time, but you can't regulate every incentive," Nallamothu states. "Either we forbid physicians from performing these services in offices, which may lead to problems with access, or we need to value these services more fairly and set the right price upfront."
| Most Common ‘Inappropriate’ Indications | ||
| Indication | % Inappropriate Indications | % Total Studies |
| Detection of CAD; asymptomatic, low CHD risk | 44.5% | 6% |
| Asymptomatic, post-revascularization; < two years after PCI, symptoms before PCI | 23.8% | 3.2% |
| Evaluation of chest pain, low probability; interpretable ECG and able to exercise | 16.1% | 2.2% |
| Asymptomatic or stable symptoms, known CAD < one year after cath or abnormal prior SPECT | 3.9% | 0.5% |
| Pre-operative assessment; low-risk surgery | 3.8% | 0.5% |
| Total: | 92.1% | 12.4% |
| Source: Robert C. Hendel, Multicenter Assessment of the Utilization of SPECT MPI Using the ACC Appropriateness Criteria Pilot Study |