A prototype Medicare program designed to improve the value of care by paying more to physicians who perform better on measures of health care quality and spending has failed to deliver on its central promise and, in the process, likely exacerbated disparities in health care delivery, according to findings of a study published November 27, in the Annals of Internal Medicine journal.

The Value-Based Payment Modifier program, which ran between 2013 and 2016, inadvertently shifted money away from physicians who treated sicker, poorer patients to pay for bonuses that rewarded practices treating richer, healthier populations. This unintended consequence stems from the program’s failure to properly account for differences across various patient populations in clinical and social risk factors for poor outcomes, the researchers note.

The researchers say that these findings bode ill for the program’s successor—launched in early 2017—because its basic design is similar to the failed, earlier iteration of the model.

Many health care reforms seek to simultaneously address three central concerns in contemporary care delivery: lowering costs, improving the quality of care and the health of patients and expanding access to care. The Value Modifier does not seem to have done anything to lower costs or improve care, and it may have made things worse for equity, the researchers said.

This latest study, led by investigators from the Department of Health Care Policy at Harvard Medical School and the Department of Health Policy and Management at the University of Pittsburgh Graduate School of Public Health, adds to a growing body of research questioning the merits of pay-for-performance and demonstrating unintended consequences of the approach for vulnerable populations.

“As long as these programs do not account adequately for patient differences, which is very difficult to do, they will further deprive practices serving low-income populations of important resources,” said Eric Roberts, assistant professor of health policy and management at the University of Pittsburgh Graduate School of Public Health and lead author of the study.

The Value Modifier was phased out in January, when it was replaced by the similar Merit-based Incentive Payment System, known as the MIPS, the largest pay-for-performance program for physicians and practices ever enacted by Medicare. The new program does not offer much hope that it will have better results, the researchers said.

“We’ve gone headlong into pay for performance despite study after study showing that it doesn’t improve quality or lower overall spending,” said J. Michael McWilliams, professor of health care policy at Harvard Medical School and senior author of the study. “We should expect more of the same from the MIPS because the MIPS is more of the same.”

The debate over the MIPS continues to gain steam in the wake of the Medicare Payment Advisory Commission’s recent recommendation to repeal it, the researchers said.

The Value Modifier offered bonuses for the top-performing physicians and practices, and penalties for practices that performed the worst. In theory, these incentives would encourage doctors and providers to find ways to deliver better care that improves outcomes. The measures included how often patients were hospitalized for preventable reasons, how often hospitalized patients were readmitted within a month of discharge, annual rates of mortality, and total Medicare spending per patient.

The researchers compared these performance metrics between practices that were exposed to the program and those that were not and found no differences in performance as a result of the incentive program. The researchers concluded that weak incentives for performance improvement in the Value Modifier might have contributed to its ineffectiveness, and noted that many of these weaknesses were not corrected in the MIPS.

What’s worse, the researchers said, is that not only did the program fail to deliver the hoped-for performance improvements, it also may have inadvertently worsened health care disparities, the researchers said.

To accurately measure care delivery across different patient populations, it is important to adjust for the fact that different populations have different health outcomes depending on a number of factors, such as preexisting chronic illness and socioeconomic status. The Value Modifier adjusted for only a limited set of risk factors. When the researchers recalculated practices’ performance measures after accounting for additional demographic and clinical differences, the performance gap between practices serving sicker or poorer patients and those serving wealthier or healthier patients narrowed by 9% to 68%, depending on the measure. If the program had accounted for differences in patient populations, up to a quarter of practices would not have been eligible for bonuses provided through the program, while a similar proportion of practices likely would not have been penalized, the study found.

Since the program was designed to be budget-neutral, this means that bonuses for physicians and practices that treated healthier, wealthier patients were paid for by penalties applied to those treating poorer, sicker people. Over time, such transfer of resources could make it more difficult for practices serving many low-income and medically complex patients to invest in resources and services that could improve care for disadvantaged populations.

“The current pay-for-performance approach is costly, ineffective and headed in the wrong direction,” McWilliams said. “If we want to improve quality, it is time to rethink our approach.”