Don’t Let Perfect Be the Enemy of Good: Performance Measures for Stroke Rehabilitation

In the era of value-based care, patients and hospitals are rewarded for high quality care, as value = quality/cost. The institute of medicine defines quality as “the degree to which health care services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge¹.” We need to be proactive as a medical community in creating feasible, reliable, and meaningful metrics that are based in the best available evidence, as performance metrics are increasingly driving reimbursement for quality. It is on this basis that Stein and colleagues² have derived performance metrics for Stroke Rehabilitation based on the American Heart Association/American Stroke Association Guidelines for Adult Stroke Rehabilitation and Recovery³.

So what makes a “good” performance metric? In the development of performance measurement, it is important to address several key questions.

1. Is the topic important? While this may seem obvious, we cannot develop metrics for all processes or disease states. Stroke is a high severity and prevalent condition with geographic and ethnic disparity and variation in quality across the continuum of care. The intensity of comprehensive, coordinated, continuous services required for stroke rehabilitation is considerable. Stakeholders in stroke rehabilitation have reason to support improvement in numerous dimensions. Over half of the costs associated with stroke care occur after the acute hospitalization, including indirect costs such as unemployment⁴, premature death making this a topic of importance to commercial and government payors. Stroke is a top cause of disability, which greatly impacts patients and their families.
2. Is the performance metric based on strong scientific evidence? It is the goal of the stroke performance measures oversight committee to create performance metrics in close timeframe to publication of new AHA/ASA guidelines, focusing on Class I and Class III evidence to quickly implement and disseminate evidence-based quality care to the stroke community.
3. Is the measure valid and reliable? Performance metrics should measure what they are supposed to measure, and this should be reproducible across various settings.
4. Is it feasible to collect the data? Data collection should occur at a reasonable effort and cost, and should also lead to meaningful results that are actionable by care providers. Clearly, it is easier to collect data already available in reports, but there may be cases where abstraction of the medical record may be necessary to gain the appropriate data to translate into meaningful change. Performance metrics may lead to information technology changes that more efficiently and effectively abstract data, and additional effort may be required to educate the clinical team to improve documentation. While a measure itself may be valid, the reliability of the practice metric across various settings will depend on the feasibility of data collection.
5. Is the measure attributable to a particular provider? If a metric does not clearly define which provider is responsible for the measurement, data collection may prove to be challenging, and this may not translate into meaningful data for clinicians to improve quality of care for our patients.

Considering the above, how do the new practice metrics measure up?

Overall, this was a carefully selected group of performance measures created from the rehabilitation guidelines with thoughtful consideration of included/excluded patient populations, measurability of metrics, and challenges to data collection. The chosen metrics span topics important to patients across the continuum of care including activities of daily living, depression screening and treatment, and coordination of care as patients transition across various settings from inpatient to outpatient. They also address topics associated with readmissions, including fall prevention, DVT prophylaxis, dysphagia screening, and urinary retention, making these metrics⁴ meaningful to payor stakeholders. The metrics are valid and based on scientifically sound guidelines³.

As with any performance metrics, there will be challenges to implementation, many of which are addressed by the authors. Ensuring reliability of the data will be complicated. It is recommended that patients who qualify for, would benefit from, and have access to inpatient rehabilitation care should go to inpatient rehabilitation. Several questions arise. Who is the source of truth for this clinical recommendation? Is it the physical therapist, occupational therapist, speech therapist, stroke physician, or physiatrist? And how do we also account for shared team decisions that include the patient and family? We have all been part of discussions where team members may disagree about a patient’s candidacy, so this may be challenging to measure within and across different inpatient hospital settings. Second, defining access to care and geography is difficult as needs including finances, insurance, contact isolation or dialysis, and family visitation often dictate whether the patient truly has access to a particular inpatient rehabilitation facility. While this is addressed in the denominator, it may be difficult to measure accurately from chart review.

Feasibility of data collection will also be complex. It is recommended that patients who are candidates for post-acute rehabilitation receive organized, coordinated, and interprofessional care. What defines organized, coordinated, interprofessional care and how is this abstracted from a medical chart in a reliable manner?

Attribution will also be a challenge. Post-stroke depression screening and treatment is recommended annually for outpatients. While some patients may only follow up with their primary care physician, others may be followed by stroke neurology, physiatry, primary care, and who is responsible and accountable for screening and treatment of depression?

I commend the authors for developing the first set of performance metrics for rehabilitation after stroke. As Donald Berwick said, “You cannot improve what you don’t measure.” What we measure is what gets attention, and these performance metrics will allow us to concentrate on significant aspects of post-stroke rehabilitative care. It is also essential to remember that everything that is important cannot be measured and everything that is measured is not necessarily important. Don’t let perfect be the enemy of good!