PACE: A Case Study of Information-Driven Innovation and Care

PACE, the Program of All-inclusive Care for the Elderly, is an innovative and rapidly growing model of services for older adults needing long term, chronic care. Today there are 159 PACE programs operating in 32 states and the District of Columbia. These programs operate more than 300 PACE centers, serving more than 72,000 participants across the country (National PACE Association, n.d.).

PACE is predicated on a simple idea—providing services based on what the individual needs, not on what is reimbursable. Like an HMO, PACE receives capitation payments instead of multiple service fees for the delivery of all needed services. Unlike an HMO, which focuses on risk management—spreading the cost of services for the few sick across the mostly healthy many—PACE focuses on service management. PACE optimizes the management and delivery of services needed by a group of people, all of whom need continuing care and support. Information played a key role in the development of PACE and is key to its success.

Program Genesis—Three Key Ingredients

PACE had its genesis in the Chinatown/North Beach section of San Francisco in the early 1970s. Three forces came together to make it happen. A caring public health department dentist brought together community leaders with the goal of building a nursing home in their community to care for their frail elders. The group received a small grant to hire a Swiss social worker with a commonsense, can-do attitude to help them. That social worker found that starting a nursing home not only was not possible but also was not a good option. She proposed instead the development of an adult day care center modeled after British day hospitals. The group applied for and received one of four federal Administration on Aging (AoA) day center demonstration grants.

The social worker was hired as the executive director of the program now known as On Lok. She found a location, hired a core team of community health professionals and began to create an adult day center. Halfway through that first demonstration, a data-oriented researcher joined the effort. That researcher interviewed the management and clinical team to identify the questions and measures of success and started collecting and organizing data on the people served, their levels of need, the services they were given, and their outcomes. As that first demonstration was finishing, the executive director asked the researcher to use that data and the lessons learned to write a proposal for funding the next step. The data collected showed: 1) that many of the people served were eligible for skilled nursing care, and 2) the day health center needed to be part of a larger community care system including social day care, in-home services, and housing. The proposal expanded the service model to address these concerns.

Adult Day Health Services, from Demonstration to Permanence

That second demonstration was funded. A social day center and in-home services were added and a housing funding drive launched. The primary focus, however, was the refinement and more permanent funding of Adult Day Health Services (ADHS). Assessment data was used to identify those participants who were nursing home eligible. The less impaired went to the new social day center.

Data systems were refined to collect data on the nursing and therapy services received in the Adult Day Health Care (ADHC), the changes in level of care, and the cost of the ADHC services. Using that data, discussions were held with the California Department of Health Services (DHS) to make ADHS reimbursable under Medicaid.

‘Daily rate funding was significant because it eliminated the need and extra cost of traditional fee-for-service billing.’

The researcher worked with DHS to study On Lok’s ADHS program. State staff certified the Skilled Nursing Facility (SNF) eligibility of the participants served. Data was collected on the services participants received, participant satisfaction, and costs. The study found that ADHS participants received more professional and supportive services, at a lower cost, and were happier with the services they received.

Based on the data, On Lok worked with DHS to fund On Lok’s ADHS as a new service under Medicaid. On Lok’s cost data was used to calculate a single daily rate for all services provided by the ADHS. Daily rate funding was significant because it eliminated the need and extra cost of traditional fee-for-service billing and allowed clinical staff to do what was needed.

The DHS–On Lok Medicaid ADHS demonstration was approved. A few years later the demonstration was expanded to include other sites across the state. The data collection systems developed at On Lok were adapted and used by the state for the operation and management of this growing program. A few years later ADHC became a permanent California Medicaid long-term care (LTC) service and has served as a prototype for Medicaid Adult Day Services programs in other states (Zawadski & Ansak, 1983).

ADHS to PACE—Expanding the Integrated Service Model

While ADHS was evolving into a permanent Medicaid service, On Lok evaluated its second demonstration to see how it could be improved. The evaluation found that other community services were helpful but needed to be coordinated with ADHS. Moreover, it found that Medical services were inexorably intertwined with long-term care services and they both needed to be integrated. Guided by these findings, On Lok’s research team proposed a third demonstration, a Community Care Organization for Disabled Adults (CCODA) that would integrate all services—medical, social, and supportive—needed by a long-term care–eligible person (Zawadski & Ansak, 1983).

That third demonstration was funded by the federal Office of Human Development Services. That demonstration allowed On Lok to hire primary care physicians, contract with medical specialists, and manage and control acute and long-term care admissions. In talking with policy makers, federal and state health administrators, and legislators, it became clear that while keeping people out of nursing homes and providing higher quality care in the community was admirable, containing or reducing health and long-term care costs was paramount.

Data collection was expanded to include Medicare measures of acuity, compatible tracking of medical services, inpatient days (acute and SNF), and their costs. The data showed that integrated community services reduced acute admissions and length of stay and almost eliminated SNF days (Zawadski et al., 1985). These medical care savings more than covered the cost of the expanded community care services.

Based on these findings, it was proposed to the Centers for Medicare & Medicaid Services (CMS) that Medicare and Medicaid fund On Lok’s CCODA for its full package of services on a capitation basis like an HMO. The research team expanded the data-collection system to include medical information used by Medicare and contracted with a team from Brandeis University to analyze Medicare data to identify costs for beneficiaries with the disability level of the CCODA population. Average Medicare costs for that population was 2.39 times higher.

Using that frailty cost adjuster for a Medicare capitation rate, the research team calculated Medicaid’s cost for the LTC portion. On Lok approached CMS offering to cover all of the services required by this LTC population for a risk-based fixed monthly capitation rate that was 5% lower than their current cost of care. CMS was interested and, because they felt it was high risk for a small organization, even agreed to a risk corridor around the rate. (Risk corridors limit risk to a small amount above or below the capitation rate with the funder [CMS] paying any costs above 105% and taking back any savings below 95%.)

‘Useful service research needs to start in the trenches and be program based.’

An interesting aside, the state balked at the risk corridor and agreed to move forward only after On Lok agreed to take “full risk” and the community leaders and On Lok executive director agreed to use their homes to backstop any losses to the State of California. That was a demonstration of extraordinary leadership commitment to their community and confidence in the data—and provided a big implicit incentive to manage costs well.

With that, PACE was born. On Lok’s CCODA operated successfully with this capitation reimbursement for a few years and was able to use the savings it achieved from better service management to build a reserve. On Lok’s R&D team, realizing that one site’s success was not sufficient to create the policy change needed to ensure the program’s funding permanence, approached and received funding from a group of national foundations to fund the development of replication sites across the country. CCODA was renamed PACE. A number of these replication sites were successful and spawned other sites. With successful replications in multiple states, federal legislation was eventually passed to make PACE a permanent program and allowing states to include it in their state plan.

Lessons Learned—Value of Program-based, Policy-oriented Research

Information Can Improve Services

Management theorist Peter Drucker (1967) said, “What is measured, improves.” This is especially true in the delivery of health services. Meaningful service research builds upon data collection that addresses critical questions and shares the data effectively to guide change. Successful businesses have used data for Continuous Quality Improvement (CQI). The CQI concept, learning from each demonstration what worked and what didn’t, was key to the development of PACE and can be used to continually improve program quality and effectiveness.

Bottoms Up—Value of Program-Based Research

Too often data collection is mandated by external sources like funder reporting requirements and research project needs. Staff who collect that data often do not know why or see its benefits. Useful service research needs to start in the trenches and be program based. Program staff, both management and clinical, need to be involved. Those delivering the services evolving the model need to identify the important issues and questions and provide input into the measures tracked. Equally important, the collected data needs to be shared with those staff if it is to help guide program change.

In the early days of On Lok, nurses were asked to record their services. After a month or two the data, showing four to five services per day, were shared back. The nurses were shocked that their data was actually being used. Service counts tripled in the following month.

Politically Aware–Benefits of Being Policy-Oriented

To effect change, health services research needs to be plugged in and aware of the factors that influence policy decision-making. While On Lok wanted to develop a service system that helped its community elders receive better services while staying in their homes, they recognized that the policymakers were inevitably concerned about costs. Embedding the concerns of policymakers in the research plan and collecting the necessary data across sites made policy change possible.

Conclusion—From a Small Acorn a Large Oak Grows

In 2024 Medicare and Medicaid will spend more than $6 billion dollars on PACE capitation payments. Venture groups have raised billions more to invest in expanding this program to new sites and new states. This large and rapidly growing policy initiative known as PACE all started from the commitment of a small group of local community health professionals, led by a hard driving, problem-solving social worker turned director, all of whom used research to guide and help actualize the model’s development.

Increasingly powerful computer technology has enabled more powerful models of using data, e.g., AI, or Artificial Intelligence. These models provide more powerful tools for using information to improve health services. That benefit will be enhanced if these new tools are used as part of a program-based and policy-oriented ongoing research effort. Research data was critical to the development of PACE, but it can offer even greater ongoing benefits as part of a Continuous Quality Improvement program supporting improved clinical decision-making and guiding future program and policy change.

Rick T. Zawadski, PhD, joined On Lok in 1974 as research and development director, and over 15 years designed and helped develop On Lok's service model and financing systems from its first Adult Day Center thru the PACE Replication and built the information system to help manage and study these innovations. He went on to found RTZ Associates to integrate these services in County systems and create enterprise information systems to manage and improve these programs. Jennie Chin Hansen RN, MSN, FAAN, came to On Lok in 1980 to work in research for Rick and continued for nearly 25 years, serving more than 10 years as its CEO, working in operations, replication of the model, and advancing the PACE federal legislation. Chin Hansen also served as CEO of the American Geriatrics Society, president of AARP, consulted on a San Francisco initiative to develop Geriatric Emergency Departments, and was a stakeholder to California’s Department of Health and Human Services involved in developing its first ever Master Plan for Aging.

Photo caption: Elders at On Lok engaged in a bowling game.

Photo credit: Courtesy of On Lok, 2019.



Drucker, P. F. (1967.) The effective executive. Harper & Row Publishers.

National Pace Association. (n.d.)

Zawadski, R. T., & Ansak, M. L. (1983.) Consolidating community-based long-term care: Early returns from the On Lok demonstration. The Gerontologist, 23(4), 364–369.

Zawadski. (1984.) Community Based Systems of Long Term Care. The Haworth Press.

Zawadski, R. T., Shen, J., Yordi, C., & Hansen, J. C. (1985.) On Lok’s CCODA: A research and development project: Final Report, 1978–1983. On Lok Senior Health Services.