There is a wealth of data that could help hospitals cut costs while still providing high-quality service for patients, if physicians were willing to join forces with administrators to truly understand how much their services cost, according to a new article by Stanford researchers.

The Centers for Medicare and Medicaid Services (CMS) has been pushing physicians and providers toward population-based payment, which requires that providers reduce their internal costs below payment levels.

In this effort, the beleaguered health-care payer for the elderly has been undertaking innovative payment models, such as accountable care organizations (ACOs) and bundled payment that require providers to better coordinate care and reduce reimbursements and unnecessary or redundant patient procedures.

“However, it has proven challenging for the models, which focus on costs from the payer perspective, to achieve the desired effect of reduced Medicare spending,” writes Merle Ederhof, PhD, in this Health Affairs Blog. The researcher who focuses on issues at the intersection of health-care and accounting is with Stanford’s Clinical Excellence Research Center.

Her co-authors, Alexander L. Chin, MD, MBA and Jeffrey K. Jopling, MD, MSHS, are also at the center, which is dedicated to discovering, testing and evaluating cost-saving innovations in clinical care.

Changing old patterns at hospitals and among physicians

“Highly detailed cost data generated by internal cost accounting systems already exist in a large, and growing, number of health-care organizations,” says Ederhof. 

As Ederhof wrote in this New England Journal of paper last year, the data collected by the Healthcare Information and Management Systems Society shows that more than 1,300 U.S. hospitals have adopted sophisticated internal cost accounting systems.

The authors argue that the cost data produced by these accounting systems can be used in hospitals internally to lower their costs of providing services to all their patients, both within and outside the Medicare system. But physicians must get on board.

“The high adoption rate of these cost-measurement systems is not surprising, considering that the systems are designed around the existing data infrastructure that providers must have in place for billing purposes,” the authors write. “However, while provider administrators have used such cost accounting systems for some time, we are only now beginning to see them being used by interdisciplinary teams involving physicians to restructure clinical processes.”

Some large health-care systems have already started using these accounting systems alongside teams of physicians.

Partners HealthCare in Boston has started to use this approach to analyze costs for a set of services, for example, in a recent project a team of spine surgeons reviewed and discussed unblinded comparisons at the episode and cost-category levels. 

“Analysis of the costs in the individual categories revealed variation in clinical processes across surgeons, which was very illuminating to the team,” the authors wrote.

Leaders at NYU Langone Health have also started to use the cost data in the organization’s “Value-Based Management” initiative. A key feature of the initiative, the authors write, is a dashboard that is accessible to all physicians. For each specific diagnosis-related group (DRG), the dashboard shows cost averages for each physician performing the procedure, at the procedure level and at the level of individual cost categories, such as the ICU, laboratory, operating room and therapies.

“Physicians have been highly engaged and interested in the dashboard since it allows them to compare their costs to their peers and external benchmarks, and to learn how they can restructure clinical processes to lower their costs,” the authors write.

This Value Based Management initiative at NYU, which incorporates cost savings targets, development-level incentives and quality components, has apparently resulted in substantial cost savings for the organization.

Stanford Health Care has also joined the movement to promote value-based care, recently launching its Cost Savings Reinvestment Program. 

Compare, for example, the average cost for a hip replacement surgery among five surgeons who perform the surgery in the same hospital. Then take the “positive outlier,” or the surgeon with the lowest cost for the surgery.

“Once positive outliers are identified, detailed analysis that combines physicians’ clinical expertise and administrators’ insight can uncover ways in which clinical processes can be restructured to deliver high-quality care at lower total episode cost,” the authors wrote.

Then the interdisciplinary team of physicians and administrators must try to understand why that surgeon’s costs are lower and what he or she does differently. Did she order physical therapy sooner after the hip-replacement surgery? Did he use a different anesthesia approach that resulted in a shorter recovery for the patient? 

But you still have to get those four, more expensive surgeons to adopt the less-expensive treatments. And that can go to the heart of a physician’s identity.

“Even just a few years ago concern for the cost of providing health-care services still heavily clashed with physicians’ professional identity,” Ederhof said in an interview. 

The authors believe there is no turning back.

“In my view, the shift in recent years is attributable to the fact that physicians are starting to realize that the rising costs of the U.S. health-care system are no longer sustainable and that things will have to change — with or without their collaboration,” Ederhof said.

Rural areas of China have made remarkable progress in reducing adult mortality within the past 15 years yet broadened health insurance was not a casual factor in that decline, according to a new study by an international research team that includes Asia Health Policy Program Director Karen Eggleston.

The New Cooperative Medical Scheme (NCMS), a government-subsidized insurance program that began in 2002-03, expanded to cover all of rural China within a decade. Examining NCMS and cause-specific mortality data for a sample of 72 counties between 2004 and 2012, the researchers found that there were no significant effects of health insurance expansion on increased life expectancy.

The study, published in the September issue of Health Affairs, showed results consistent with previous studies that also did not find a correlation between insurance and survival, although much research confirms NCMS increased access to healthcare, including preventive services, and shielded families from high health expenditures.

Commenting on the study, Eggleston said population health policies remain central to China’s efforts to increase life expectancy and to bridge the gap between rural and urban areas.

Eggleston also noted that multiple factors beyond the availability of health care determine how long people live, and anticipates the research team will continue to explore the impacts of NCMS by extending the study to look at infants and youth.

Read the study (may require subscription) and view a related article on the Stanford Scope blog.

Stanford Health Policy’s newest faculty member, Joshua Salomon, believes that one urgent need in global health research is to improve forecasts of the patterns and trends that are the major causes of death and disease.

Salomon, who is leaving leaving his position as professor of global health at the Harvard T.H. Chan School of Public Health to join Stanford on Aug. 1, works on modeling of infectious and chronic diseases and their associated intervention strategies, as well as methods for economic evaluation of public health programs and ways to measure the global burden of disease.

And he looks at the potential impact and cost effectiveness of new health technologies.

“Projections of future trends in health are crucial to formulating policy,” said Salomon, who has a PhD from Harvard. “To think strategically about the technologies and policies that would make the biggest impact on health over the next 20 to 50 years, we really need to start by understanding the range of likely trends in major health challenges over the coming decades.”

Stanford, he said, offers him a “rich collaborative environment” to better learn from advances in forecasting across a range of other disciplines, such as economics, political science, and environmental science.

“With a better picture of what the world is likely to look like over the next 50 years — and what are going to be the most pressing health problems — we can invest wisely and put ourselves in a position to respond more effectively.”

Salomon is also the director of the Prevention Policy Modeling Lab, which is funded by a five-year award from the Centers for Disease Control and Prevention. The consortium represents the collaborative research of experts from Massachusetts General Hospital, Boston Medical Center, Dana Farber Cancer Institute, Yale School of Public Health, Brown University School of Public Health, and the Massachusetts Department of Public Health and.

He will continue directing the lab from Stanford and intends to bring in new research threads from his colleagues here on the Farm. The lab works on a wide range of projects dealing with policy analysis for hepatitis, sexually transmitted infections and diseases such as HIV, and tuberculosis.

“It’s a rewarding grant for me to work on because, unlike a lot of modeling projects, the work that we do really starts from urgent public health questions that policymakers have,” he said. “All of the questions that we are working on are questions that originated directly from discussions with CDC and other public health partners.”

With Salomon’s move to Stanford, the university gains a dynamic duo.

Grace Lee joins Stanford as the Associate Chief Medical Officer at Lucile Packard Children's Hospital in the fall, 2017.

His wife, Grace Lee, MD, MPH, joins in the fall as the Associate Chief Medical Officer at Lucile Packard Children’s Hospital. As a professor of population medicine at Harvard Pilgrim Health Care Institute & Harvard Medical School, Lee has led research in vaccine safety in the FDA-funded Post-licensure Rapid Immunization Safety Monitoring (PRISM) program and the CDC-funded Vaccine Safety Datalink, which monitors the safety of vaccines and studies rare and adverse reactions from immunizations.

She has also examined the impact of financial penalties on rates of healthcare-associated infections, as the principal investigator of an AHRQ-funded study, as well as developed novel surveillance definitions for ventilator-related events in neonates and children.

While at Stanford, Lee said, she intends “to find opportunities to enhance the learning health system approach to improve patient outcomes and population health.”

Salomon has spent his entire career as a collaborator on the Global Burden of Disease project, the world’s most comprehensive epidemiological study commissioned by the World Bank in 1990, which tracks mortality and morbidity from major diseases, injuries and risks factors.

“The study has made a major contribution to global public health because before this study we just didn’t have a comprehensive, systematic understanding of the things that cause death and disability in low- and middle-income countries. But now we do,” he said. “It’s hugely ambitious and very sweeping in scope — and a lot of my work is around providing the evidence we need to inform policy.”

Much of Salomon’s work is global in nature. He’s most recently focused on older adults in one rural South African community, which has a high prevalence of HIV and one of the world’s highest levels of hypertension. His research there aims to inform urgent prevention initiatives tailored to older adults where HIV and cardiovascular risks are moderate or high, as in similar communities in sub-Saharan Africa.

“People don’t expect a high level of ongoing HIV transmission in older adults,” he said. “The double burden that we find, with a very high level of HIV, as well as the high prevalence of diabetes and heart disease, creates enormous strains on the health-care system.”

Emily Tuong-Vi Nguyen, a Stanford student studying human biology, writes about the Asia Health Policy Program’s international conference on diabetes

The Asia Health Policy Program at the Shorenstein Asia-Pacific Research Center hosted the Net Value in Diabetes Management Workshop in March to discuss progress on an international research collaboration. Research teams from Hong Kong, Singapore, China, Taiwan, South Korea and the United States convened at the Stanford Center at Peking University (SCPKU) in Beijing to work on research that compares utilization and spending patterns on diabetes across different countries and to develop a method for measuring the net value of diabetes internationally, based on previous methods discussed in a Eggleston and Newhouse et al. 2009 study with Mayo Clinic Data for Type 2 diabetes.

The research teams from various Asian countries are attempting to calculate the net value of diabetes in those countries by observing the changes in diabetes value and spending. These calculations include monetizing the value of health benefits of new treatments and improvements in health, as well as avoided spending on treatments when prevention was effective, and associated mortality and probability of survival. Previous models used to measure diabetic values and risks, such as the United Kingdom Prospective Diabetes Study (UKPDS) risk engine that was created from U.K. data and populations, are not very relevant for Asian populations. The goal is to create separate risk models specifically suited for populations from Hong Kong, Singapore, China, Taiwan and South Korea.

During the workshop that spanned two days, the research teams had an opportunity to share updates on their individual projects and to discuss methods and ideas for future collaboration.

On the first day, each research team presented its work, describing data sets and explaining the risk models that were used or developed. Karen Eggleston, director of the Asia Health Policy Program, delivered introductory remarks and shared current progress by the Japan and Netherlands research teams on calculating value and risk for diabetes with data from the Netherlands and Japan. The data sets from those two countries were best estimated by the JJ Risk Engine for the Japan data and the UKPDS model for the Netherlands data.

Chao Quan of the University of Hong Kong presented the risk model used for Hong Kong populations. His work primarily looked at how the UKPDS risk engine predicted risk in Hong Kong populations as compared to a local Hong Kong risk engine and how to best calibrate the Hong Kong risk engine. His next step will be to monetize the value for improved survival in diabetes in Hong Kong. He offered to re-estimate the model using the risk factors available on others’ datasets so that the Hong Kong risk model could potentially be used by other teams as well.

Stefan Ma and Zheng Li Yau of the Ministry of Health of Singapore discussed the 5-year prediction model and statistical methods they used for all-cause mortality of Singaporean individuals with diabetes. Their work is based on Singapore’s extensive administrative and claims data as well as data provided by the national health surveys conducted every six years by the National Health Service of Singapore. The researchers plan to look into how their overall risk model compares with models for specific subpopulations, such as Chinese, Malay and Indian populations in Singapore.

Katherine Hastings from the Stanford University team, led by principal investigator Latha Palaniappan, presented preliminary ideas about measuring cardiovascular risk with the Atherosclerotic Cardiovascular Disease Risk Score in analyses of Stanford health system diabetic patients. The researchers are collaborating with a clinical bioinformatics team at Stanford to use machine learning to expedite the analysis.

Min Yu and Haibin Wu of the Zhejiang Center for Disease Control and Prevention shared results from their analysis of health data collected from community health centers for diabetes management, diabetes surveillance data, cause of death data and insurance claims data that showed relationships between different patient characteristics and insurance types. The researchers then estimated the annual cost of Type 2 diabetes and its complications in Tongxiang province, China.

Hai Fang and Huyang Zhang of Peking University worked with claims data of diabetic patients insured by the New Cooperative Medical Scheme in Beijing, and at the workshop, shared regression analyses on the relationship between outpatient visits and inpatient admissions.

Jianqun Dong of the People’s Republic of China Center for Disease Control and Prevention presented ongoing research about diabetes management in China, including preliminary results of a randomized control trial of diabetes self-management strategies.

Wankyo Chung of Seoul National University shared preliminary estimates of a risk model for mortality among diabetic patients in South Korea and discussed next steps for estimating net value of diabetes management using the detailed clinical and claims data available in South Korea.

On the second day, the workshop concluded with a videoconference between workshop participants in Beijing and collaborators at Stanford Graduate Business School, including Stanford professor Latha Palaniappan and Harvard visiting professor Joseph P. Newhouse, using the Highly Immersive Classroom.

The workshop was a good opportunity for the research teams to discuss preliminary models, to offer each other suggestions regarding research methods, and to discuss the future direction of the international collaboration on the net value of diabetes. All research teams are preparing comparative research papers that will be included in the working paper series of the Asia Health Policy Program. A follow-up event will be held at Stanford in November 2017 in recognition of World Diabetes Day.

Medicare Post-election by Stanford Health Policy on Exposure