Publicly Reported Health Outcomes: A National Initiative To Improve Care

In this article, we describe the rationale and design criteria for a public-private approach to improving the quality and value of health care in the United States: a Publicly Reported Health Outcomes (PRHO) program. This long-term, aspirational initiative, focused on results and transparency, will require both public- and private-sector action.

Rationale

The US currently lags behind most Organization for Economic Cooperation and Development (OECD) countries in life expectancy, infant and maternal mortality, and many other measures of mortality and morbidity. While many factors contribute to this quality gap, huge variations in health plan and hospital mortality rates suggest that US health outcomes could be substantially improved. In fact, recent research demonstrates that hospital mortality rates can vary by a factor of 3 to 1 and health plans by 4 to 1, and publicly available quality ratings of health care plans and services do not correlate well with outcomes. What is more, lack of transparency among health plans and provider organizations prevents purchasers from making informed choices based on relative quality of provider organizations and health plan networks. Currently available outcomes data are limited to Medicare fee-for-service.

We believe, therefore, that an effective method of reporting comparative health outcomes would motivate health plans and provider organizations to improve their performance.

Changing The National Conversation

For at least five decades, the public health care conversation has focused primarily on rising costs, while the US has lagged behind other OECD countries in numerous measures of mortality and morbidity. Yet, while the US reports jobs and gross domestic product growth, inflation, and unemployment on a regular (monthly or quarterly) basis, no effective reporting system holds health plans and provider organizations accountable to the public for health outcomes. But if the US published a few key risk-adjusted measures of mortality and morbidity for payers and provider organizations, purchasers, government officials, consumers, and voters would have the necessary information to pinpoint responsibility.

External Motivation

As management consultant Peter Drucker is reputed to have said, “If you can’t measure it, you can’t improve it.” Consider, for example, how management of most publicly traded companies focus on quarterly earnings per share. A system that regularly reports mortality and morbidity to the public could bring the same kind of focus to health care organizations and purchasers. A 2013 study of quality improvement efforts in Wisconsin found “that large group practices will engage in quality improvement efforts in response to public reporting, especially when comparative performance is displayed.”

Changing The Purchase Paradigm

Today, individual and corporate purchasers of health care are flying virtually blind when it comes to quality. The industry standard for health plans, the Healthcare Effectiveness Data and Information Set (HEDIS), only tracks process measures, not outcomes. Indeed, no rating system for health plans and provider groups provides outcomes information in a usable way. If purchasers, consumers, and government officials had timely data on mortality and morbidity of procedures, hospital-acquired infections, and medication errors, they would have the tools to choose best performers. This would also drive poor performers to improve, thereby improving the overall quality of care in the US. A robust outcomes reporting system is needed to change the purchase paradigm.

Power Of Competition

One of the founding principles of the US economy and the current US health insurance system is competition. Enabling health plans and provider organizations to compete on the basis of outcomes would align well with a national goal of improving measurable outcomes compared to other nations. Giving purchasers of health care the ability to select plans with low mortality and morbidity rates could create a sea-change in the way they make their health coverage decisions.

High Performance

A 2020 National Bureau of Economic Research working paper found a fourfold difference in mortality rates among Medicare Advantage health plans within the same county. A 2018 review of teamwork in health care pointed to a study of surgery teams’ behaviors in which, “Patients receiving care with poor teamwork are almost five times as likely to experience complications or death.” Despite these trends, consumers, especially in employer-sponsored plans, tend to value choice over the benefits of team care inherent in group and staff model health maintenance organizations, accountable care organizations, and other integrated systems. Comparing mortality rates could enhance the popularity of high-performing plans and counter the excessive emphasis on choice.

A National Initiative?

In addition to benefitting purchasers of health care, public reporting of outcomes could be a powerful tool to improve the overall quality of health care in the United States. It would aid government payers, regulators, economists, and planners in allocating resources. By measuring and aggregating a few outcomes measures, the country could establish goals for quality of care at the federal, the state, and county level and identify problem areas that need attention.

Design Criteria

A PRHO program is intended to provide health quality comparisons of health plans and provider organizations to purchasers, government officials, and the general public. Health plan and provider organization executives may also find PRHO helpful in comparing their health quality performance with that of other organizations.

To succeed, the PHRO program would need to incorporate the following design principles:

• A scoring system based on outcomes rather than processes
• Comparative risk-adjusted outcomes with breakdowns by race, ethnicity, and related variables
• Clear communication of the significance of those outcomes to non-health care professionals
• Comparisons of mortality and morbidity among health plans and provider organizations
• Availability at the point of purchase, such as health insurance exchanges and provider selection sites
• Functionality allowing users to drill down to greater levels of detail, comparing plans and providers using multiple criteria
• Incorporation of advances in outcome measures as they become available over time

What To Measure

Health plans and other provider organizations can and should be measured and scored using outcomes. Existing quality reporting bodies use process-oriented measures that, according to an analysis conducted by one of us (Wadsworth), do not correlate well with comparative performance based on outcomes. We suggest developing a starter list of outcomes beginning with the following sources: the Centers for Medicare and Medicaid Services (CMS) hospital outcomes (deaths, readmissions, and infections) augmented with maternal and infant mortality and morbidity measures; and OECD or the World Health Organization (WHO) measures of mortality and morbidity. Patient reported outcomes can be incorporated when a methodology has been developed to ensure comparability among organizations.  CMS has targeted 2030 as a completion date. To be useful and credible, outcomes must be risk adjusted for age, sex, comorbid conditions, and health equity determinants.

Usability

Many provider selection services today list large numbers of provider organizations with little or no differentiating criteria. And even when that information is available, it isn’t presented in a way that the user can understand it and “filter” based on their preferences. To illustrate, we have to look beyond health care sites. Cars.com, for example, provides huge amounts of data and choices in a way that need not overwhelm users. The site enables the user to limit their search to new, used, or certified cars of a specific make and model and refine the search by specifying additional criteria, such as colors and various features to narrow the choices. Many retail shopping sites follow a similar approach.

Summarization And Granularity

While some health quality reports are too simplistic, others provide too much information. A better approach is to summarize and enable the user to drill down and filter the results to narrow their choices. HospitalSafetyGrade.org, for example, provides granularity but not in a way that enables the user to compare hospitals based solely on outcomes.

Scoring

Letter and star grades (for example, A through F or 1 to 5 stars) tend to downplay the substantial differences in mortality rates (3 to 1 or greater) that exist among hospitals and health plans. Numerical grades, such as FICO credit scores, are a much more effective way to highlight the differences that exist.

Accessibility And Public Awareness

To be useful and used, quality ratings must be widely available to the intended users at their decision points—such as health insurance exchanges, employer intranets, and public sites—and become as widely known as Consumer Reports, Amazon, Google, and Facebook.

State, Health Plan, And Provider Rankings

The PRHO program would aggregate outcomes data for states, health plans, and provider organizations. By using data-mining and artificial intelligence to collect, aggregate, analyze, and report quality performance data, it would minimize the administrative burden on health plans and provider organizations. The program should first be tested at the state level or among a group of health plans and provider organizations before going national.

Potential Obstacles

The reader may ask, why isn’t public outcomes reporting being done yet? The answers include structural and motivational factors and inertia. Health outcomes have been reported internationally for years by the OECD and the WHO. America’s Health Rankings annual state rankings demonstrate the feasibility of PHRO at the state level. CMS collects hospital outcomes, but the data are typically one and a half to two years old before they are made public and only account for Medicare fee-for-service patients, a population that continues to shrink as Medicare Advantage grows in popularity. Medicare and Leapfrog both dilute the CMS outcomes with process measures so that CMS mortality outcomes represent less than 20 percent of the basis of Leapfrog’s Hospital Safety Grade and 22 percent of Medicare hospital ratings.

The National Committee for Quality Assurance’s HEDIS database does not include a single outcome. Care Compare’s Quality Payment Program suffers from similar problems. The data are two years old and consist mainly of processes not outcomes. Many of these approaches are undermined by the lack of a national all-payer database that includes outcomes, but at least 11 states already have all-payer databases, a prerequisite for an outcomes database.

Siloed electronic health record (EHR) data have also been an obstacle. But the two largest vendors, Epic and Cerner, are now sharing data, and interoperability is a high priority within the industry and at the Department of Health and Human Services. A PRHO program would require a data collection and reporting system that builds on existing EHR systems. We believe that with a push from CMS, interoperability is inevitable within the next few years—with or without PRHO.

Finally, state and federal policy makers could overcome the reluctance of provider organizations and payers to furnish outcomes data by enacting mandates.

The Way Forward

The PRHO program will require an integrated real-time, all-payer information system that brings together patient data from all sources with a minimum of human intervention. Some of the key initiatives needed to make PRHO a reality include:

Defining What Will Be Measured

Convene a multistakeholder group of health care purchasers and consumers advised by health professionals with expertise in quality measurement to determine how to measure specific patient outcomes that include risk-adjusted mortality, patient harm measures, such as iatrogenic infections and medication administration errors, as well as patient-reported outcomes.

Enacting Mandates At The State Level, Federal Level, Or Both

Ensuring all provider organizations and payers comply with data collection and reporting requirements will require federal or state mandates. Such compliance could be a condition for any health plan listed on an insurance exchange. The Joint Commission and other accrediting bodies might also require PRHO program participation as a criterion for accreditation.

Developing The Information System

The PRHO program will require an information system that extracts and anonymizes all patient outcomes from existing claims, regardless of payer, and from medical records with little or no manual intervention. EHR companies have already developed robust observational databases that could possibly serve this purpose. New and rapidly developing technology such as blockchains and artificial intelligence may also facilitate and accelerate the process. What is more, to minimize the administrative burden on health plans and provider organizations, the PRHO could rely on data-mining and artificial intelligence to collect, aggregate, analyze, and report quality performance data.

Testing And Refining

No information system can be implemented without rigorous testing and modification. The early failures of the Affordable Care Act’s federal health insurance exchange was an unfortunate case in point and an important reminder for future rollouts: Testing on a small scale at county or state level would help facilitate implementation on a national scale.

Publicizing

Without broad public usage, a PRHO program would be of limited usefulness. A PR program led by state health exchanges and other appropriate conduits will be needed to inform purchasers of health care of why, how, and where to access the PRHO program.

A Long-Term Commitment

Public reporting of mortality and morbidity outcomes offers the potential to transform the national conversation and move the country’s health care results closer to those of other OECD countries. But development and implementation of a PRHO system will require a long-term commitment.