Critics are pushing back against Medicare quality reporting, deeming it burdensome and time-consuming to meet confusing quality measures. One survey asserts that barely a majority feel knowledgeable about MACRA or prepared to achieve long-term success. Indeed, CMS is pulling back on program requirements, with the stated desire of making it easier for physicians.
So, here’s what should be examined—especially when discussing Value-Based Health Care: Does MIPS Quality Reporting meet the benefit test for the effort expended by physicians and their staff? If the point of Quality Measurement and Reporting is to improve care for patients, can it fulfill that potential?
For Quality Results to Demonstrate Clinical Quality, then Quality Reporting Will Need Redesign
Quality Measurement Results and actual quality are distinct concepts. For those outside the process, reporting results seem easy to interpret. Good results may appear as clinical excellence, while poor results may appear as if the physician isn’t doing a good job. But this interpretation is invalid. The measures themselves, along with variations in practices, populations, and other quirks introduce inconsistencies into reported results. For example:
1. Measures often require single, once-per year values to convey patient status.
The measurement system obtains some good information, but it falls short of achieving a focus on long-term outcomes or improvement over time. Quality Reporting was designed to involve physicians in examining patient data, and we should remember that we are still at an early stage in its development.
2. Reporting is easier for some measures than others because data is inconsistently available.
Actions like the selection of specific perioperative antibiotics are often based on standing orders, requiring little or no effort because collected data almost always includes them. Information systems are designed to drop these CPT Category II codes automatically for these measures and have done so for years. However, there is a downside to this, too—when the clinician’s software automatically uses a code that has not been valid for years. The results then appear as if some providers deliver excellent care (when the correct code is dropped), while others do not.
Other measures are written with the assumption that providers have easy access to specific data points. Reports from pathology, radiology and echo labs are often scanned documents, meaning that the EHR cannot simply scoop up the data and transmit it as needed.
For certain measures, it’s easier to document the quality action than it is to determine whether the patient is eligible for the measure. For instance, for MIPS Measure #8, used to report whether patients with heart failure have been prescribed beta-blockers, it’s easier to document that a patient is on a beta-blocker (and transmit that data) than it is to report that, in addition to a diagnosis of heart failure, the patient also has a left ventricular ejection fraction of less than 40 percent.
These discrepancies are not distributed evenly, and the groups with outdated technology—or none—are forced to commit team members’ time and effort to a manual chart review process in order to fulfill reporting criteria.
3. Not all measures account for real life patient scenarios, so reporting negative results is open to misinterpretation.
Measures are designed by different organizations or specialty committees, using different processes, and reporting options may not translate into practice. For example, many measures allow for only two possible responses: a desired quality action was performed or it was not performed. The provider does not have the opportunity to justify the reason for not performing the action—it was either done or not.
Take the case of the measure for pneumonia vaccination (MIPS measure #111), which allows only two responses: patient received/previously received the vaccine (the “performance met” option) or, the patient did not receive it (“performance not met”). Patients who have received hospice care are excluded, but not a new patient who does not recall a November 2017 visit, but brings records to a visit in 2018. Alternately, there is no allowance for coverage by a physician for temporarily treating a patient, such as a patient seeing a provider out of town on a one-time basis, but who plans to follow up later with his or her primary care provider. The temporary provider would not vaccinate the patient without knowing the patient’s full history. In both situations, the provider’s scores would suffer, while the provider acted appropriately.
4. Performance scoring is based exclusively on reported patients.
Reporting results can be misleading, but performance, even more so. Performance is graded only when the provider reported a measure. So it’s possible to have a performance rate of 100 percent based on only one patient, even if one hundred were eligible. That isn’t enough to meet MIPS performance scoring requirements, but the discrepancy between completion and performance detracts from meaningful analysis. It’s not possible to reasonably compare Provider A’s 100 percent performance with Provider B’s 80 percent performance, because the proportion of reported patients varies.
To add an additional layer of complication, the same measure may be scored differently depending on how it was reported. For measures with multiple allowed submission methods (e.g. EHR, Registry or claims), the number of points earned with a 90 percent performance rate on a measure may earn fewer points if submitted through a different mechanism. If the manner in which the information is submitted can affect quality scoring, this rewards the practice for finding the best way to earn the most points through a submission strategy, rather than for meeting quality standards.
While Quality Results are Inconclusive, the Reporting Burden Is Often Exaggerated
The fact is that the majority of practices have EHRs that produce data for reporting either through the direct EHR method or delivery of that data to a Registry or QCDR. There is little actual work—or notice paid—by most clinicians to quality data capture. And, clean-up work is usually handled by staff without involving the physician. It can be argued that the process changes required to ensure that quality measures are met are part of the point of reporting. However, no one can argue that there is no work involved, but it has become easier and significantly less time-consuming.
Accurate reporting, however, does require ongoing review and assessment. Many have fallen into the trap that their EHRs would automatically capture quality data and report it accurately on their behalf, without looking at the details of how the EHR could accomplish this task, and what may be required. As a result, many EHRs have been used differently in practice than the manner intended, with data not captured into the record as data fields. Rather than creating a patient-centric, searchable database, some groups have allowed charts to include scanned (and not searchable) documents and free-text notes. The data needed for quality metrics is visible to providers, staff and administrators, but does not contribute to measure data or even important clinical data captured in the EHR database. The EHR could store the information, but could not report it.
Specialty-Centered Measures—Intended to be Fair to Physicians—Instead Complicate Reporting
Reporting challenges have been amplified by the vast increase in measure volume. The number of measures in play has significantly increased, especially under MACRA. In 2007, there were only 74 measures for the Physician Quality Reporting Initiative (PQRI), and only three (any three) were required. Since then, we’ve seen as many as nine measures required, broken down in different categories. Each year, new measures are added, some measures are deleted, and changes are made to others. The process that’s being used in one year may not work in the next year, so results can’t be consistently compared, either.
If the information needed to report is not readily accessible, the burden falls on those tasked with finding those measure responses: organizations’ administrative and technical teams. This drives the process toward finding the information for reporting, rather than improvement. The best performers are the ones who can find the information, since inconsistent EHR use may only shift the burden away from clinicians.
How To Make Quality Measurement and Reporting More Meaningful—and Less Burdensome
Quality reporting can be meaningful, but providers and CMS both have to change how the program works and how it is implemented to make that happen:
1. CMS should clarify and mitigate formulas where comparisons may result from underlying data or measure flaws. If quality measurements cannot realistically compare physician quality through equivalent numeric scores, then formulas that punish providers for not meeting benchmarks should be relaxed. At the same time, however, providers should not get a total pass for failure caused by bad data, without documenting a corrective action plan with their EHRs (or their usage of these systems) to improve quality reporting in the future. Both providers and consumers must get a better explanation of what quality scores represent and how best to use them.
2. Providers must take action to improve their data and systems for performance measurement and improvement. Reported results that assume everyone can provide the same level of detail are unrealistic and will result in inequitable scores. MIPS and Alternative Payment Models give practices the freedom to participate in the manner that they choose, even allowing for variance between groups who are participating in the same program.
3. CMS should bring Cost scoring back into MIPS. Currently, quality reporting accounts for more than half of the total MIPS score, and this overvalues quality scores in a merit-based payment adjustment. By returning Cost to the MIPS scoring algorithm, CMS can combine quality reporting with an alternate quality determination—expenditures per patient, and per episode of care. This brings balance to quality scoring, by looking at what happened on an episodic basis. Quality can then be measured using a combination of quality reporting and costs. Neither cost/utilization nor quality results, alone, provide a complete picture , but taken together, they help to to identify potential improvement opportunities, which may be achieved through the Improvement Activities portion of MIPS.
4. CMS and Providers should be rewarded more for Performance Improvement. One of the historical problems of quality reporting is that clinicians removed themselves from the process. Thus quality reporting often failed to involve providers in review of data and benchmarks, stifling initiatives for improvement, as well. Providers must recognize that while quality reporting is not going away, the review of outcomes and processes is critical.
Cost data alone does not sufficiently guide improvement initiatives, but the addition of accurate, comprehensive quality information will facilitate the full discussion of value. Population-based care and specialty episodes of care are models of how both cost and quality/outcomes can be measured simultaneously—and eventually compared to others.
5. CMS and Providers should concentrate their quality reporting on a streamlined core of quality measures where improvement is evaluated over time. The overwhelming number of quality measures now in use represents a focus on physicians rather than patients. They also measure single-point-in-time values as opposed to outcomes over time, which present a better view of patient status. Unlike MedPac recommendations in its June 2017 report to Congress, we believe that this data should include clinical EMR data and not just claims data.
6. Providers should focus on creating value from their EHRs through appropriate use, standardization and better education and training. This will mean commitment of resources to investigating how the EHR stores data, ensuring that codes are updated, and that regular education and training are provided to users. Good use of technology will remove reporting burdens from providers. By standardizing what is collected, and how, organizations can shift quality reporting from a burdensome administrative task into a method of quantifying outcomes over time.
7. Consumers and patients must be part of the direct measurement process. Providers need to consider how to incorporate direct patient feedback, as well as outcome data, into quality measurement and reporting. CMS also needs to adopt a strategy that includes these elements in quality measures.
Founded as ICLOPS in 2002, Roji Health Intelligence guides health care systems, providers and patients on the path to better health through Solutions that help providers improve their value and succeed in Risk. Roji Health Intelligence is a CMS Qualified Clinical Data Registry.
Image Credit: Robert Lukeman