Share this article on:

POLICY WATCH: Why Stroke Specialists Are Questioning Two New CMS Stroke Measures

Butcher, Lola

doi: 10.1097/01.NT.0000438144.99817.47


Stroke specialists are not happy with the quality of care measures that the Centers for Medicare & Medicaid Services (CMS) will introduce next year. But they hold out hope that the federal government will adjust course before unintended consequences hurt patient care.

“We have to prepare for the worst and hope for the best,” said Irene Katzan, MD, director of the Center for Outcomes Research and Evaluation in Cleveland Clinic's neurology department.

At issue are two hospital quality measures that CMS intends to add to its Hospital Inpatient Quality Reporting Program: 30-day all-cause readmission rate and 30-day all-cause mortality rate following acute ischemic stroke.

CMS will begin reporting those measures on its Hospital Compare website in July 2014, reflecting each hospital's performance for the three years ending June 30, 2013. Beginning in the government fiscal year that starts Oct. 1, 2015, CMS will include the two measures in its payment formula so hospitals with high stroke readmission and/or mortality rates will be punished financially for their performance.

Neurologists told Neurology Today they are alarmed because neither measure is severity-adjusted to reflect the patients treated at a given hospital. Thus, tertiary and quaternary medical centers that specialize in the treatment of the most severe stroke patients are at risk of financial penalties because they accept the difficult cases.

“By doing that, they put themselves in a tough situation because, in many cases, despite their best efforts, the patients do not have a good outcome,” said Mark J. Alberts, MD, vice chair of clinical affairs in the department of neurology and neurotherapeutics at University of Texas Southwestern Medical Center. “We — meaning society — do not want to penalize these particular medical centers that reach out and bring in the sickest stroke patients.”

Back to Top | Article Outline


As part of its push to improve quality, CMS started publicly reporting 30-day mortality measures for acute myocardial infarction and heart failure in 2007 and pneumonia in 2008. Since then, CMS has expanded its Hospital Compare website to include 30-day readmission rates for those conditions along with a long list of patient-safety indicators, hospital-acquired condition measures, and process of care measures.

To incentivize hospitals to improve the quality of care they provide, CMS more recently started adjusting hospital pay to reflect a hospital's performance on certain outcome and process measures. It intends to continually expand the list of measures used in its payment calculation, and stroke is a priority.

That is understandable to Marilyn M. Rymer, MD, who chairs the Academy's Stroke Systems Work Group.

“Stroke should be a focus,” said Dr. Rymer, vice president of neurosciences at the University of Kansas Hospital. “It represents a huge financial challenge for the US because in addition to being a leading cause of death, stroke is the leading cause of adult disability.”

The wide variation in outcomes for stroke patients supports the idea that some hospitals can improve the care they deliver. CMS reported that the median 30-day risk-adjusted readmission rates for Medicare patients hospitalized for stroke in 2007 was 14.7, but the rate ranged from 11.6 percent to 19.4 percent across 4,242 hospitals.

Meanwhile, the 30-day mortality rate for that same patient population was 15.3 percent — and the rates ranged from 10.7 percent to 23.5 percent across 4,288 hospitals.

Public reporting of readmission and mortality rates will shine light on the variability of outcomes from one hospital to the next. But unless the rate calculations reflect the severity scores of patients treated at a given hospital, the data will not clarify which hospitals deliver the highest quality care, Dr. Alberts said.



“To properly interpret that quality data, folks need to know the types of patients that a given facility is taking care of,” he said. “That is going to impact the bottom-line results that are being reported by CMS.”

Dr. Alberts has been a leader in the movement to designate primary and comprehensive stroke centers and establish protocols that ensure stroke patients are transported to the center that can treat them most effectively. He and others worry that the CMS use of stroke measures that do not take severity into effect might undermine the stroke center concept.

“If you're getting a disproportionate number of the most severe strokes, your mortality rate is going to be higher, and therefore you are going to do poorly on this CMS measure,” Dr. Rymer said. “The unintended consequence of that could be — and we hope this will not happen, but it could — that the hospital decides to back off accepting these severe stroke patients and those people will not get to the place where they can get the best treatment.”

“Even if techniques could effectively adjust for the effect of stroke severity on mortality, there are other factors that have a significant impact on early stroke mortality that may be difficult to account for,” said Adam Kelly, MD, assistant professor of neurology at the University of Rochester Medical Center in New York. For example, there is considerable variation in the use of do-not-resuscitate orders and the utilization of life-sustaining interventions such as gastrostomy tube placement, tracheostomy, and decompressive hemicraniectomy following stroke, he said.

Patients may have advanced directives or strong preferences about the intensity of care that they want in the setting of a severe stroke; the use (or non-use) of higher intensity interventions is likely to impact 30-day stroke mortality, Dr. Kelly explained. Unfortunately, these preferences are not typically included in the large administrative databases from which risk-adjusted mortality rates are generated.

“If a patient with a severe stroke has clearly defined wishes to avoid mechanical ventilation or artificial hydration and nutrition, then these directives need to be followed even if mortality is the outcome,” Dr. Kelly said. “Medical care in this setting can still be evidence-based but should also be preference-sensitive and patient-centered.”

Back to Top | Article Outline


In general, CMS uses quality measures that have been endorsed by the National Quality Forum (NQF), but the stroke measures are an exception.

The NQF's neurology steering committee initially voted to endorse both measures but, after receiving considerable feedback during the public comment period, the committee rejected the readmission measure and issued an 11–11 tie on whether to approve the mortality measure. Another public comment period ensued, but the committee opted not to re-vote the readmission measure, and CMS withdrew the mortality measure from consideration before the committee met.

In issuing its final rule, CMS said it was unable to find other feasible and practical measures for stroke, which is why it proceeded despite the concerns of the neurology community.

One problem is that obtaining a severity rating is not currently standard practice for all clinicians who see stroke patients. While stroke specialists use the National Institutes of Health Stroke Scale to measure severity, staff at community hospitals where most stroke patients initially go for treatment generally are not trained to use the scale.



“Some people think it would be too difficult to require the NIH stroke scale to be used or that (clinicians who are not stroke specialists) couldn't learn it,” Dr. Rymer said. “I happen to disagree. We have been training nurses to use it for over a decade and they do it very well. But one of the contentions is that, if we can't get an accurate severity rating at the onset of the stroke, then how can we even talk about using that as a risk adjustment tool?”

Another factor in the CMS decision is its established protocol for measure development. The agency's risk-adjusted outcomes measures for myocardial infarction, pneumonia, and heart failure are based on hospital-level administrative data and do not include a severity indicator.

“They are using their usual approach, which has worked for these other diseases, and I think they believe it is fine to use in stroke,” Dr. Katzan said.

Back to Top | Article Outline


The stroke measures have moved beyond the proposal stage and were finalized in the CMS final rule for the Hospital Inpatient Prospective Payment System published in the Aug. 19 Federal Register (

Despite that, Dr. Alberts thinks the measures may be refined to address neurologists' concerns.

“In the discussions we have had with them, they do seem to be open and receptive of the concept of doing some pilot testing with some type of stroke-severity measure included in the formula, which I think is a reasonable path forward,” he said.

Dr. Katzan said neurologists need to advance their argument in two ways: an analysis of the impact of not including stroke severity on risk adjustment for the two measures and the development of a process whereby stroke severity scores can be universally collected, regardless of the care setting.

“Maybe this battle didn't go the way the neurology community felt would be best, but there are many things we can do to provide more information and hopefully use to modify the score in the future,” she said.

Back to Top | Article Outline


•. CMS Hospital Inpatient Quality Reporting Program:
    •. National Quality Forum:
      •. Neurology Today archive on quality measures initiatives:
        •. AAN resources on quality measures:
          •. Neurology Clinical Practice archive on stroke & quality measures:
            © 2013 American Academy of Neurology