The Center for Disease Dynamics, Economics & Policy

With the unveiling of the new Partnership for Patients Initiative last week and elements of the Affordable Care Act (ACA) kicking into full swing, there’s an optimism circling the U.S. healthcare system about our capacity to improve healthcare quality while also reducing costs.  Both goals have proven historically elusive.  A decade after the Institute of Medicine’s landmark report To Err Is Human: Building a Safer Health System, we have only seen modest gains in quality improvement and little to no gains in reducing disparities among population groups.  April’s aptly titled Health Affairs – Still Crossing the Quality Chasm – contains an excellent collection of articles on the ongoing journey towards quality improvement in a healthcare system where it is estimated that adverse medical events occur in up to 1/3 of hospital admissions.

Nested within this conversation about health reform is talk about widespread implementation of pay-for-performance programs, which provide financial incentives for providers and systems to meet certain quality benchmarks, ultimately as a means to achieve better patient outcomes. Pay-for-performance factors into the ACA, as the act shifts CMS payments towards values-based purchasing – i.e. paying based on quality metrics rather than based on quantity of procedures.  But as the nascent Partnership for Patients builds momentum, it’s worth considering several open questions about the structure and efficacy of pay-for-performance programs.

How effective are pay-for-performance programs and how do they fare in the long-term?
The design of effective pay-for-performance programs is open to much debate.   Some research even questions whether certain pay-for-performance programs have a positive effect on quality at all. A new study in Still Crossing the Quality Chasm finds that hospitals enrolled in a CMS-initiated demonstration pay-for-performance program did in fact see greater improvements in quality measures in comparison to a control group – but only for the first three years.  In years four and five, the groups achieved nearly equivalent gains.  

The research team then raises a series of important questions – exactly how big do incentives need to be in order to motivate change, particularly in the long-term? Should pay-for-performance programs consider dissolving once certain gains have been achieved in order to devote resources to improving a different quality measure?  Does pay-for-performance work best in non-competitive markets, where hospitals are not competing for market share and thus are unlikely to be as affected by something like a public reporting mandate? All questions to consider when structuring CMS payments under the ACA.

If reporting is linked to financial incentives, do we know that reporting mechanisms produce accurate results?
For many pay-for-performance programs, financial incentives might go hand-in-hand with public reporting of quality measures.  For example, the ACA requires hospitals to begin reporting rates of common hospital infections, including CLABSIs and SSIs, and in late 2012, CMS will begin paying hospitals based on publicly reported rates of healthcare associated infections (HAIs).  This stipulation provides incentives to hospitals to improve infection control for two reasons.   One is that incentives are tied to reducing HAI rates, the other is that making data on infection rates publicly available, such as through HHS’s Hospital Compare Website, could influence consumer choices.  A prospective patient who is able to weigh hospitals based on their quality ranking has the opportunity to seek care at a higher scoring hospital.

Linking incentives to reporting requires that we trust our reporting mechanisms, but is reporting accurate?  Using administrative codes – which are primarily for billing – for disease surveillance has been demonstrated to be highly problematic. But new research suggests that even “gold standard” reporting methods may be under-counting adverse medical events, including HAIs, pressure ulcers, etc. by tenfold.

This troubling assertion about the accuracy of reporting mechanisms for adverse events warrants further research into the design of reporting mechanisms.  The formidable challenge at-hand is to structure reporting programs so that they are clinically accurate but not so overly labor- or expertise-intensive to be infeasible.  Refining reporting mechanisms will likely be aided by integrating processes with electronic health records (EHR).

Are measures actually correlated with improved quality and reductions in healthcare disparities?
If the purpose of pay-for-performance is to reward hospitals and providers that provide higher quality care, aligning metrics with improved quality might seem like an obvious point.  But the connection is perhaps less straightforward than it looks.  According to Peter J. Pronovost at the Johns Hopkins University School of Medicine, "[t]here is bipartisan support behind efforts to start paying for value rather than volume. This is great, but we act as if there's a whole library of reliable outcome measures for us to use and the fact is that serious work needs to be done to create them."  Underlying this work is the question of whether we in fact have the research data to inform the design of optimal measures. 

Additionally, we should not forget about including reducing disparities, which persist despite modest gains in quality over the past decade, as an important component in the quality metric.  Incentives should be guided to ensure that the system does not end up penalizing hospitals that serve disadvantaged populations.  They could be structured to reward improvements, for example, or could offer larger rewards to hospitals that serve as social safety nets.  A brief from the Robert Wood Johnson Foundation suggests that there are indeed reasons to believe that pay-for-performance can be leveraged to reduce disparities.

Achieving the goals of Partnership for Patients
During the press conference announcing the Partnership for Patients initiative, Kathleen Sebelius, Donald Berwick and others were careful to note that there are hospitals and hospital systems across the country that are making some real progress in reducing adverse medical events.  One of the key components of the new program is recognizing these pioneering systems and initiatives and determining how to scale them up nationwide –defining the models enough to produce real change while keeping them flexible enough to fit across hospital contexts.  The goals Partnership for Patients has initially set – decreasing hospital-acquired conditions by 40% and hospital readmissions by 20% by the end of 2013 – will be a good test of widespread pay-for-performance as a motivator for progress in improving quality of care.

Imange credit: Flickr: Lidor