Partnership for Patients hospitals tout quality improvements but critics see a lost opportunity

When HHS unveiled its Partnership for Patients initiative in 2011, its $1 billion budget and ambitious improvement targets signaled to many in healthcare that the patient-safety movement was finally being recognized with a sweeping and long overdue national program.

More than 3,700 hospitals signed on to participate in the voluntary initiative, touting their success in tackling issues such as obstetrical harm and healthcare-associated infections. Consumer groups and large employers lent their support. Patient advocates offered enthusiastic praise. Federal officials called it a game-changer.

But three years later, some prominent thinkers in quality and patient safety are criticizing HHS for what they say is a poorly designed program that sheds little light on what actually works in making care better and safer. They say the program lacks scientific rigor because it didn't require hospitals to adhere to a set of standardized quality measures. And they argue that it's difficult to tell whether safety gains are actually due to the Partnership for Patients rather than other improvement initiatives.

“To launch a national program without agreed-upon measures is premature and it's not a good use of taxpayer dollars,” said Dr. Peter Pronovost, director of the Armstrong Institute for Patient Safety and Quality at Johns Hopkins Medicine.

“I see this as a lost opportunity to learn how to improve,” said Dr. Ashish Jha, a Harvard University associate professor of health policy and management. “I'm not saying we have to be slaves to the classical randomized trial, but what they've done has very little scientific validity.”

CMS officials, however, say the relaxed rules were necessary to encourage large numbers of hospitals to join the program, and they say they have been nudging hospitals toward use of uniform performance measures over time. And those on the front lines say the effort has significantly boosted hospital quality and fostered collaboration and shared learning. “The progress we have been able to achieve is nothing short of spectacular,” said Keith Kosel, who leads Partnership for Patients efforts for Irving, Texas-based VHA, which has reported improvements in areas such as early elective deliveries and ventilator-associated pneumonia.

Dr. Don Berwick, CMS administrator at the time the Partnership for Patients was launched and widely seen as its chief strategist, rejects the argument that using a number of different measures necessarily makes it difficult to judge hospitals' performance. “As long as they're tracking their progress over time, we can see if they're getting better,” said Berwick, who is now running for the Democratic nomination for Massachusetts governor. “It's nice to have some common metrics, but it's far more important to have local settings using their own approach.”

Launched in April 2011 with funds provided by the Patient Protection and Affordable Care Act, the partnership aimed to reduce hospital-acquired conditions by 40% and preventable all-cause 30-day readmissions by 20%, both by the end of 2013, using data from 2010 as a baseline. HHS effectively pushed the 2013 deadline to the end of 2014 when it awarded all 26 Hospital Engagement Networks, or HENs, a contract for a third year, a decision Kosel said was motivated in part by weak baseline data for 2010.

The program addresses specific types of preventable harm, including surgical-site infections, adverse drug events, pressure ulcers and central line-associated bloodstream infections. Federal officials said reducing HACs and readmissions by the targeted percentages could prevent as many as 1.8 million patient injuries and save approximately 60,000 lives, while saving Medicare as much as $50 billion over a 10-year period.

HHS made the 26 HENs its main engines of improvement, charged with leading efforts for their member hospitals. The HENs include large groups, such as Premier, VHA and the American Hospital Association-affiliated Health Research & Educational Trust; state hospital associations, such as the Minnesota Hospital Association; and large health systems, such as 37-hospital Dignity Health.

HENs' duties include helping hospitals focus their efforts on areas of harm, facilitating the sharing of evidence-based practices through in-person meetings and conference calls, and collecting and submitting monthly performance data to the CMS Innovation Center, which oversees the campaign.

When federal officials laid the groundwork for the program, they gave HENs and their member hospitals plenty of leeway on which process and outcome measures they used. Indeed, hospitals in many cases were encouraged to use measures they already were collecting. That decision has made it a challenge to compare performance within individual HENs and across the country as a deluge of data on many different measures pours in, some experts say.

But Dr. Patrick Conway, the CMS' chief medical officer and deputy administrator for innovation and quality, contends that some of the program's measures are indeed standardized. For instance, most hospitals report data on catheter-associated urinary tract infections and central-line infections using the Centers for Disease Control and Prevention's National Healthcare Safety Network, an online infection- tracking system. In addition, the CMS has developed other standardized measures during the course of the program, and is asking the HENs to urge—though not require—member hospitals to collect data on them.

“Some measures were standardized from the beginning, some we standardized over time and some are self-reported measures where we allowed flexibility,” Conway said. A number of data sources, including the 2013 progress report for HHS' National Quality Strategy, are showing promising reductions in harm over the past two years, he added.

Responding to the question of whether improvements are due to Partnership for Patients efforts or other programs, Conway said the Innovation Center is using regression analysis and other techniques to try to tease out the causative factors.

CMS officials say the agency has had to weigh the goal of collecting data in a uniform way with making policies lenient enough that hospitals would actually participate.

“It's a sensitive balance and we struggle with it every day,” Sean Cavanaugh, the CMS Innovation Center's deputy director of programs and policy, told Modern Healthcare last year. “In the long run, we need common metrics. But in the short term, we are trying to provide some flexibility.”

But Pronovost, who has headed a number of safety initiatives, including a successful effort to curb central line-associated bloodstream infections in Michigan intensive-care units, strongly disagrees with that approach. He initially drafted a HEN application with the Maryland Hospital Association but grew apprehensive when he learned more about the program's design, which he said greatly resembled the Institute for Healthcare Improvement's 100,000 Lives campaign designed by Berwick when he led the IHI before going to the CMS. In the end, Pronovost decided not to participate.

He wasn't the only one to have early reservations about the partnership's methodology. In May 2012, Dr. Don Goldmann, chief medical and science officer for the IHI, co-authored a perspective piece in the Journal of the American Medical Association urging a more rigorous measurement strategy. “Delay in measuring the effect of best practices on harm rates will postpone the day when the nation can celebrate significant improvement in patient safety,” he and his colleagues wrote.

Other quality experts are more ambivalent. They share the concerns about widely disparate measurement strategies, but praise the Partnership for Patients programs for encouraging public-private cooperation and energizing hospital improvement efforts.

Dr. David Nash, founding dean of the Jefferson School of Population Health in Philadelphia, called the campaign a major advance that never would have happened without the Affordable Care Act. He praised the CMS for allowing flexibility on performance measures, despite acknowledging the potential for murky results.

Leaders of the HENs say the CMS made the right call. Kosel said the VHA's HEN would not have been able to recruit 191 hospitals if the partnership's protocols were more rigid. “They made a judgment: Get more hospitals in the program and take whatever data they can give us,” he said. “I think that was one of the smartest things they did.”

During 2012 and 2013, VHA's HEN hospitals achieved overall reductions in all nine of the initiative's preventable harm types, including a 58% drop in early elective deliveries, a 48% drop in ventilator-associated pneumonia and a 30% drop in falls with injury, according to an annual report. Kosel acknowledged that it's difficult to say whether the Partnership for Patients was the direct cause of those improvements when other quality-improvement programs were operating at the same time, including Medicare value-based purchasing. “Peter Pronovost is right that this is not designed as an experiment, but the reality is we've seen tremendous improvement,” he said.

One of VHA's HEN hospitals is Jennings (La.) American Legion Hospital, a 60-bed rural facility. Through the Partnership for Patients, the hospital has achieved success in improving venous thromboembolism prophylaxis rates through the use of a scorecard that tells each unit how it's doing, said Phyllis Theriot, the hospital's director of clinical systems improvement. Compliance rates have jumped from 50% to 93%, she added.

Jennings CEO Dana Williams said resource-strapped rural facilities like hers reap particular rewards from the collaborative approach of HENs. “We have the same compliance requirements as larger hospitals, but we have to rely on a smaller subset of resources,” Williams said. “Networking and sharing with peers through the HEN is so beneficial to us.”

Urban public hospitals also benefit greatly from sharing best practices and lessons learned, said Thomas Holton, patient safety officer and director of education and training at 460-bed San Francisco General Hospital. His facility is part of America's Essential Hospitals' 22-member HEN. The hospital focused primarily on pressure ulcers and falls with injury. Pressure-ulcer rates have fallen significantly, from eight or nine a month down to zero or one, though progress on falls has been more uneven, Holton said.

Dignity Health, San Francisco, which runs its own system-wide HEN, has reported significant safety gains in early elective deliveries. The system halted nonmedically indicated early elective deliveries before 39 weeks gestation and reduced rates from 7% to .03%, said Barb Pelletreau, Dignity Health's senior vice president of patient safety. The system also has seen big drops in central line-associated bloodstream infections and catheter-associated urinary tract infections.

Like other HEN leaders, Pelletreau said the CMS' flexible measurement policy was needed to get hospitals on board. She argued that it has not hindered Dignity's ability to generate data that can be used to compare performance both within its own HEN and across other HENs. “We don't all need to agree on the same numerator and denominators,” she said. “If CMS had taken that approach, I think we would still be arguing about metrics right now.”

Dr. Maulik Joshi, president of the Health Research & Educational Trust, agreed that stringent rules can lead to “data collection paralysis.” HRET's massive HEN has nearly 1,600 hospitals across 31 states. In an annual report released in December 2013, HRET said its HEN-related efforts have prevented harm for 69,072 patients over two years, equivalent to an estimated $200 million in savings.

“We have an encyclopedia of process and outcomes measures posted on our website,” said Charisse Coulombe, HRET's vice president for clinical quality who oversees the HEN. “If we told hospitals exactly what to submit, they might not have joined.”

That kind of logic is exactly the problem, Pronovost said. He argues federal officials could have decided from the start on a smaller, standard set of valid outcomes measures that weren't necessarily too burdensome to hospitals, while allowing for a localized approach to implementation. In his Michigan central-line project, for example, he and his colleagues omitted a range of process measures, focusing instead on a small set of outcomes measures. “Hospitals will perform to the bar that you set, especially if they believe the standards you're setting are worthwhile,” Pronovost said.

The government skimped on the science and took the “feel-good approach,” he said. “I worry that at the end of 2014, we're going to realize we didn't really learn anything about how to prevent harm. That's a tragedy.”

Follow Maureen McKinney on Twitter: @MHMMcKinney