Documenting Change: The Community Tracking Study

Why And How HSC Tracks Changes

n virtually every community in the country, the health care system is changing. Some of these changes are dramatic---a for-profit hospital chain buys a number of previously not-for-profit hospitals in a given community. Others are less obvious but nonetheless important---providers expand outreach for diabetics to get preventive care. Yet, until recently, there has been little work done to systematically note, analyze or fully understand these changes.

There are several reasons for this. Many key data are not collected on a longitudinal basis, and what are collected are not made available for analysis in a timely enough manner. Also, there is little systematic collection of data at the community level, where health care is organized and delivered. For example, to track changes in access by the uninsured, we can learn much more by studying people in a sample of communities with different proportions of uninsured and different safety nets than through a national sample.

To fill in these gaps, Health System Change is collecting data from multiple sources to investigate what is happening over time in health care financing and delivery at the community level. Equally important to tracking organizational change, HSC is assessing how these changes affect people.

Community-Focused And Longitudinal Research

The major effort of HSC is the Community Tracking Study, a longitudinal study that focuses on changes in the health system in 60 sites that are representative of the nation. Every two years, HSC conducts surveys in all 60 communities and site visits in 12 of them. The first round of surveys and site visits, conducted in 1996 and 1997, provided baseline data-the starting point against which changes documented in subsequent surveys and site visits will be tracked. The second round got underway in June 1998. HSC staff supplement survey and site visit findings with secondary data from government and private-sector sources. The secondary data enable HSC to do national tracking before the second round of surveys is completed and also provide insight to help understand local variations.

Based on their multiple sources of data, HSC's researchers are answering two broad questions that are critical to all health decision makers:

• How is the organization of the health system changing? How are hospitals, health plans, physicians, safety net providers and other provider groups restructuring, and what key forces are driving organizational change?
  
  
• How do these changes affect people, and how are insurance coverage, access to care, use of services, health care costs, quality and satisfaction with health care changing over time?

Targeting Communities That Represent The Nation

HSC randomly selected 60 sites that are representative of large and small metropolitan areas and rural areas in the United States. From the 60 sites that have a population of more than 200,000 people, HSC staff randomly selected 12 communities: Boston, Mass.; Cleveland, Ohio; Greenville, S.C.; Indianapolis, Ind.; Lansing, Mich.; Little Rock, Ark.; Miami, Fla.; Newark, N.J.; Orange County, Calif.; Phoenix, Ariz.; Seattle, Wash.; and Syracuse, N.Y. These 12 communities are studied in more depth than the other 48 sites, with extensive site visits and larger sample sizes for the surveys.

Getting Answers From Households, Physicians And Employers

Much of the information for the Community Tracking Study comes from three surveys that were designed by or in collaboration with HSC staff to determine how health system change is affecting people.

What Americans Have to Say: The Household Survey

Many Americans are concerned about the changes that are taking place in health care. Is access to services getting better or worse over time? Are more people uninsured today in spite of legislative action in Congress and the states to expand insurance coverage?

From July 1996 to July 1997, nearly 60,000 persons in 33,000 families participated in a household telephone survey designed to address these issues. Questions covered health insurance, access, service use, satisfaction with and perceived quality of the care received and general characteristics, such as health status, family income, employment status, age, race and gender. The survey was conducted by Mathematica Policy Research, Inc.

Sample Household Survey Questions

• Does your plan require that you sign up with a certain primary care doctor whom you must go to for all of your routine care?
• Was there any time during the past 12 months when you did not get medical care you needed?
• How would you rate how well your doctor listened to you and explained things in a way you could understand?
• If you smoke, did your physician counsel you to stop smoking in the past year?
• Do you trust your doctor to put your medical needs above all other considerations when treating your medical problems?

Responses from Practitioners: The Physician Survey

Practicing physicians can provide important insights into how the health care system is changing. For example, has the complexity or severity of the medical conditions that primary care physicians treat without referring patients to specialists increased, decreased or stayed the same during the past two years? What types of organizations are physicians moving into? What kinds of financial incentives do physicians face?

To capture such information, some 12,350 physicians who spend at least 20 hours a week in direct patient care were interviewed by telephone between August 1996 and August 1997. Survey questions covered physician supply and specialty distribution, practice arrangements, physician time allocation, sources of practice revenue and level and determinants of compensation, physicians' perception of their ability to deliver care, career satisfaction, effects of care management strategies and provision of charity care. The survey was conducted by The Gallup Organization.

Purchasers' Perspective: The Employer Survey

Most Americans have access to private health insurance through the workplace, so what employers do plays a particularly critical role in shaping the health care system. To analyze the types of plans offered by employers and the cost of premiums to employers and employees, HSC collaborated with RAND to develop community-level findings from the 1997 Robert Wood Johnson Foundation Employer Health Insurance Survey.

From September 1996 to October 1997, more than 22,000 private and public employer establishments were interviewed, primarily by telephone. Survey questions covered a range of issues relating to employer-offered insurance, including the number, types and cost of premiums of health plans offered by employers; the number of employees enrolled in the plans and their share of premiums; and techniques used by employers to control costs, such as participating in purchasing pools and modifying financial incentives to encourage employees to shop for low-cost plans. The survey was conducted for RAND by the Research Triangle Institute.

Sample Physician Survey Questions

• In the last two years, were you part of a practice that was purchased by another practice or organization?
• For roughly what percent of your patients do you serve as a gatekeeper?
• During the last two years, has the number of patients whom you refer to specialists increased a little or a lot, stayed the same or decreased a little or a lot?
• Is it possible for you to provide high-quality care to all of your patients?
• How large an effect does feedback from patient satisfaction surveys have on your practice of medicine?

Voices In 12 Communities

An important component of the Community Tracking Studying is the series of interviews with key players in the health system in each of the 12 sites. By asking probing questions of leaders from the major organizations in the community, HSC was able to obtain an understanding of the health care system in each site that complements the survey data. Based on this information, HSC was able to document similarities and differences in the health systems across communities, explain trends and identify issues to track nationally in the future.

From May 1996 to April 1997, researchers from HSC and The Lewin Group conducted between 36 and 60 interviews in each site. Key among the people interviewed were leaders of insurance companies and health plans, hospitals, physician groups, employer and other purchaser groups, health departments, community health centers and agencies and policy and consumer groups. After analyzing the data, individual case studies were written describing each local market, including how care is organized and delivered, the role of employers and other community forces and the impact of public policy. Findings from the case studies are referred to throughout this report.

Research Collaboration

HSC has established numerous collaborative arrangements with individual researchers and research institutions. David Blumenthal of Massachusetts General Hospital helped develop the physician survey and will contribute to the analysis of its findings; Jon Christianson of the University of Minnesota helped develop the design of the site visits and has participated as a team leader and report author. Research institutions working with HSC include: The Gallup Organization (Physician Survey, under the direction of Linda Keil); The Lewin Group (site visits, under the direction of Raymond Baxter); Mathematica Policy Research (Household Survey, under the direction of Richard Strouse); and RAND (Employer Survey, under the direction of Stephen Long and Susan Marquis). Jon Gabel of KPMG Peat Marwick, working in partnership with HSC, analyzed trends in employment-based insurance.

HSC is also working with researchers on a number of collateral studies that build upon the Community Tracking Study. Kenneth Wells of RAND is tracking issues related to substance abuse and mental health care in the 60 sites. RAND's Robert Brook, Beth McGlynn and Steve Asch are conducting a pilot study assessing quality in the 12 sites by reviewing medical records of respondents to the Household Survey. With funding from the U.S. Department of Health and Human Services, Rose Martinez of Mathematica Policy Research is undertaking a study of public health departments in the 12 sites.

HSC also consults with its two advisory committees to provide guidance on research design and dissemination. Members of the Researchers Group and the Users Group are listed on page 29.

Sample Employer Survey Questions

• How many different plans do you offer at this location?
• Do you pay part of the premium for employees and, if so, how much?
• Are plan enrollees required to select a gatekeeper to make referrals to most specialists?
• If enrollees use a provider who is out of the plan network, must they pay the full cost of these visits?
• Do you use the results of patient satisfaction surveys or scores on a report card of selected outcomes, such as HEDIS measures, to select health plans?

Reaching Diverse Audiences With HSC Findings

As a result of the surveys, site visits and data analyses, key decision makers-from members of Congress to community leaders-are learning more about the health care system and how it is changing at the national and local levels.

At the national level, for example, as Congress debated allowing near-elderly individuals to buy in to Medicare, HSC published a timely analysis of the vulnerability of different age groups with respect to uninsurance. After analyzing data from the Household Survey, HSC concluded that most uninsured persons ages 55 to 64 would have serious problems affording coverage under the proposed buy-in.

At the local level, communities are learning how they measure up relative to other markets, fueling discussion about how to improve. Release of an HSC analysis about uninsured children inspired a front-page story in the Miami Herald and a round of questioning by local leaders about why the uninsurance rate for children in their community was 50 percent higher than any other site studied by HSC. HSC's research, which was based on Household Survey data, identified a number of factors that appear to contribute to Miami's high uninsurance rate, including the prevalence of small firms, which are less likely to offer health insurance, and the constant influx of immigrants into the area.

HSC is eager to get more researchers involved in tracking change and comparing health systems across communities, and so will be releasing data from the Household and Physician Surveys in public use files this year. Also helpful to research and policy activity are HSC's recent primers on how to make sense of seemingly disparate cost data and how to discern trends on access to care.

"This study by HSC about the uninsured is an important contribution to the Medicare policy debate. It clearly demonstrates that a significant gap exists between the cost of coverage and the ability of the uninsured near elderly to pay."

–Robert Reischauer,
The Brookings Institution

Round One: What HSC Is Learning

HSC's 1997 annual report focuses on early findings from the Community Tracking Study in three important areas: the extent of managed care; access to care, particularly for the uninsured; and health care costs. Since the first round of HSC surveys and site visits provides mostly baseline data, tracking changes over time is not yet possible. (There are a few exceptions to this, where survey questions and site visit interviews specifically asked about changes.) Ongoing analyses of round-one data in 1998 will permit staff to conduct more in-depth analyses, including how being in a managed care plan affects access and the perceived quality of care delivered, why insurance coverage varies across communities and other issues.

Managed Care

In a relatively short period of time, managed care has become well diffused across the country and has changed the way most Americans receive their health care. By 1996, according to KPMG Peat Marwick data, more than 70 percent of both small and large firms offered some form of managed care plan. As an indication of how far the managed care pendulum has swung, only 57 percent of large firms (those with more than 200 employees) even offered a conventional, fee-for-service plan.

Managed care today means far more than just getting a discounted rate from providers for services. HSC staff were able to document the pervasiveness of a vast array of care management activities that are changing the doctor-patient relationship, physicians' relationships with plans and hospitals and other key interactions in the health system-all of which have the potential to affect care delivery from the consumer perspective. These care management activities include gatekeepers, capitation, disease management programs to address chronic conditions and outreach efforts to ensure that all children receive necessary preventive care services, among many other techniques.

But managed care is not evolving in recognizable stages across communities. There is no single pathway being followed. This provides its own set of challenges for analysts and underscores the importance of HSC's community-based research approach. The work described below reflects HSC's initial focus on the extent of managed care techniques. Staff are currently analyzing the impact of managed care on consumers.

Gatekeepers

HSC's Household and Physician Surveys document just how pervasive gatekeeping has become. Nationwide, 40 percent of Americans with health insurance report that they are in some form of a gatekeeping arrangement. While nearly two out of three insured persons in a community with high managed care penetration such as Boston are in gatekeeping arrangements, as many as one out of three in low managed care penetration communities such as Greenville and Syracuse has a primary care gatekeeper. Looking at this from a different perspective, more than 90 percent of primary care physicians report that they serve as gatekeepers for at least some of their patients.

"If your market intelligence comes from exposure to the speaking circuit, your blueprint of reality will be woefully inadequate. The systematic collection and analysis of data will help us separate fads from trends, and provide a far more accurate picture of how the health system is evolving.

– Uwe Reinhardt,
Princeton University

Nature of Plans in Communities

It seems that national plans should have an advantage over regional plans, particularly in terms of economies of scale and access to capital. However, HSC found that regional plans dominate. Their advantage is based on name recognition and established relationships with providers.

With one exception, none of the national plans controls any of the markets in the 12 sites. In most cases, these regional plans have a long local history, for example, the Harvard Pilgrim Health Plan, Tufts Associated Health Plan and Blue Cross and Blue Shield of Massachusetts in Boston. The exception is Orange County, where the dominant competitors--PacificCare/FHP, Kaiser Foundation Health Plan and Foundation Health Systems--were active regionally before going national..

Consumer Choice

What do consumers think about having limits placed on their choice of providers? The general public is divided about which is preferable: limiting the choice of physician and spending less, or having choice and spending more money. Nearly three-fifths of Americans responding to HSC's Household Survey said they were willing to accept managed care's limits on choice of physician to save money (see Figure 1, p. 20). At the same time, a sizable minority-40 percent-were not willing to give up choice to save money. This divergence of opinion was fairly constant nationwide.

These findings may help explain the surging popularity of preferred provider organizations (PPOs) and point-of-service (POS) products, which accommodate these divergent perspectives. Based on visits to the 12 communities, employers are seeking to further increase choice for employees by getting managed care plans to expand the number of providers in a plan's network and providing more options for out-of-network care at affordable prices.

The movement toward more provider choice and unrestricted access is seen by some observers in the 12 communities as a win-win situation. It helps satisfy employees and makes it easier for employers to enroll workers in managed care plans. However, there likely will be a price to pay: providers gain bargaining power, and expanded networks are more difficult to manage from both a cost and quality perspective.

Figure 1: How Willing are Americans to Accept Limited Choice to Contain Costs?

Source: Community Tracking Study

Financial Incentives

While discounts for fee-for-service still dominate, financial incentives for physicians are being used more extensively. The HSC Physician Survey documented that among those physicians in practices with capitated arrangements, the average proportion of practice revenues from capitated contracts accounted for more than one-quarter of total practice revenues, and half of all physicians were eligible to receive some type of financial bonus based on how they or their practice performed.

From the perspective of most physicians, financial incentives have not challenged their ability to practice good medicine: Seven of 10 physicians say they can make clinical decisions in the best interests of their patients without the possibility of reducing their income (see Figure 2). However, three out of 10 doctors do perceive a conflict between the financial incentives they are working under and their clinical decision-making authority. This finding does not address how physicians resolve the conflict.

Access To Care

Access to care takes on different meanings, depending on whether a person has health insurance. Access for the growing number of uninsured-rising from 34.7 million in 1990 to 41.7 million in 1996-is focused on whether these individuals are able to get care for which they cannot pay. For people with insurance, access tends to mean how much choice they have within their health plan to select doctors and to get care from specialists when they perceive it is necessary.

HSC's community-based research approach helps to shed light on the reasons why uninsurance rates are rising in communities, and will eventually help decide whether new legislative approaches are successful in addressing fundamental problems that contribute to uninsurance. With respect to the insured, HSC's surveys will help to better define and track newly emerging issues of access for those who have coverage.

Figure 2: Financial Incentives and Clinical Decision Making

Physicians' responses to the statement: "I can make clinical decisions in the best interests of my patients without the possibility of reducing my income."

Source: Community Tracking Study

Access for the Uninsured

After the failure of President Clinton's Health Security Act in 1995-which promised to address the uninsured problem comprehensively-policy makers began to take an incremental approach to improving access for the uninsured by targeting particular age groups; 1997 was a banner year for this.

• The Balanced Budget Act created the Children's Health Insurance Program (CHIP), which gives grants to states to cover low-income children through Medicaid or private insurance.
• President Clinton proposed that near elderly-people between the ages of 55 and 64-be able to buy in to Medicare for about $300 a month, and that displaced workers ages 55 to 62 be able to buy in for about $400 a month.

Insurance for Children: Uninsured children represent a diverse group, but they are largely children of the working poor. According to HSC's Household Survey, nearly four-fifths have parents in the work force, even though many of them live in poverty or just above the poverty line.

By expanding eligibility for public insurance, CHIP has the potential to significantly reduce the number of uninsured children in America. But many would argue that other approaches need to be pursued to address the problem, including market reforms to expand affordable, employer-sponsored coverage. HSC research indicates that half of the children who lack health insurance are in this predicament because their parents either work for firms that do not offer insurance or are ineligible for their employer's health benefits. In addition, some parents cannot afford family coverage, even when it is available. This combination of public and private programs may result in equity and crowd-out problems.

“As states roll out CHIP, many of them are grappling with sticky equity issues. For example, how can we provide CHIP benefits to workers making 200 percent of poverty in companies that do not provide insurance, while expecting individuals with similar incomes at companies that do offer insurance to pay high contributions for coverage?”

- Patricia Riiley,
National Academy For State Health Policy

Insurance for the Near Elderly: HSC found that most uninsured near-elderly persons ages 55 to 64 would have problems affording coverage under the President's buy-in plan, but that those in the poorest health would have the most difficulty. Specifically, the average uninsured near-elderly person would have to spend 20 percent to 25 percent of his or her income on Medicare premiums, while that person's counterparts in poor or fair health would have to spend between one-third and one-half of their income to obtain coverage.

The uninsured near elderly are among the poorest and sickest of all uninsured persons. One-fourth to one-third characterize themselves as being in poor or fair health, compared with 16 percent of all uninsured. These sicker near elderly have average annua l incomes of less than $10,000, while overall the near elderly have average incomes of about $46,000 a year. These findings are based on an analysis of HSC's Household Survey data.

The Safety Net: The traditional sources of care for the poor-the so-called safety net-are being challenged to adapt to health system changes in their communities. The most significant factors include:

• hospitals experiencing cutbacks in federal funding and, in those states that abandoned rate setting, loss of subsidies to serve the indigent; and
• community health centers facing increased competition for Medicaid patients, resulting in a loss of some patients and lower payments for those they retain.

HSC's visits to 12 communities provided new insights into how communities are responding to these pressures. In Lansing, the Ingham County Health Department spearheaded a community-wide effort to see that all the local hospitals are providing their proportionate share of indigent care, and is experimenting with managed care for the uninsured. And in Orange County, the Medicaid program reserves a portion of managed care funds for direct contracting with traditional safety net providers.

Targeting Vulnerable Age Groups: Does it Make Sense?

If incremental change is to proceed by age group, policy makers need to define their objectives related to expanding access to health insurance. HSC found that if the goal is to target assistance 0to people with significant health care needs, there is some justification for focusing on the near elderly. Sixteen percent of all uninsured people under age 65 who were interviewed reported being in fair or poor health, but 32 percent of uninsured people ages 55 to 59, and 28 percent of those ages 60 to 64, reported poor health status.

However, if the goal of incremental expansion of health insurcance is to target assistance to those with unmet need, it is less clear that any particular age group should be targeted. Uninsured adults of all ages appear to encounter significant financial barriers to getting the care they need.

Access for the Insured

Historically, most people with health insurance were not very concerned about their access to care, with the exception of people who lived in rural areas where there were few physicians and relatively few high-technology services. There is a new dimension to any discussion of access to care for the insured today, however, largely due to the expansion of managed care.

As noted earlier, issues of access for people with insurance tend to focus on choice. Do they have a choice of plan? How wide a choice of doctors do they have? Can they see a specialist when they think they need one? Findings from HSC's Household Survey s how that most reported no change in their ability to get care over the past three years, but 21 percent insured nonelderly said it had become more difficult over time (see Figure 3).

One group that was dramatically less likely to perceive declining access to care was the elderly. Only 11 percent of people age 65 and older reported that it was more difficult to get care. A likely reason for this relatively good response is that the elderly are almost universally covered by Medicare and experience greater consistency in coverage.

An increasingly important access issue for people with insurance is the ability to get referrals to high-quality specialists when they are medically necessary. While most physicians feel that they can make such referrals, 20 percent of primary care physicians responding to HSC's Physician Survey said that they cannot always or almost always obtain these referrals. This finding closely parallels what consumers think about the issue: 15 percent of people responding to HSC's Household Survey felt that their physicians might not refer them to a specialist when needed (see Figure 4).

Although managed care has processes to limit referrals to specialists, the role of managed care in explaining differences in problems with specialty referrals is not easily identified. While Boston has the highest rate of primary care gatekeeping in the Community Tracking Study and a relatively high HMO enrollment rate, it is significantly below the national average for the proportion of physicians and patients concerned about access to specialists.

Other communities also known to have high managed care penetration, such as Miami and Orange County, have higher proportions of physicians and patients concerned about this. Clearly, factors beyond simply the amount of managed care or even gatekeeping-for example, the supply of specialists in the community and the competitiveness of the market-influence this important aspect of access to care.

Figure 3: Access to Care Compared to Three Years Ago for Insured Population

Source: Community Tracking Study

Figure 4: Insured Patients' Trust in their Doctor to Refer them to a Specialist When Needed

Patient's responses to the statement: "I think my doctor may not refer me to a specialist when needed."

Source: Community Tracking Study

Cost Of Care

The cost of health care profoundly affects all aspects of the health system and, from the purchasers' perspective, the news about health care costs has been relatively good: The soaring increases in health care costs that characterized the 1980s have been steadily declining since 1990. The numbers are dramatic: Since 1994, there have been historically low rates of increase. In 1997, a health cost index for the non-Medicare population increased by 3.3 percent and private health insurance premiums rose only 2.1 percent (see Figure 5).

Tracking the Slowdown of Cost Increases

By documenting this far-reaching trend, HSC played an important role in convincing decision makers that low rates of increase are real and not merely an oddity or a short-lived reaction to pressures for health care reform. At the national level, it has been widely predicted that premiums would start rising again, but to date they have not. A group of experts convened by HSC concluded correctly in mid-1997 that the forces that have kept health care costs in check during the past few years will remain in effect for the immediate future, and that rates of increase will continue to remain low. In fact, insurance premiums for 1997 still ran below underlying costs, and while there are anecdotal reports of premium increases for 1998, the average increase turned out to be only 3.3 percent.

While this is the general cost trend, it is important to note that local markets likely will vary. For example, across the 12 Community Tracking Study sites, the Employer Survey found that the rate of premium increase was lowest in Orange County and highest in Little Rock. These community-by-community variations were modest compared with the substantial variations in premium changes reported by individual employers. In 1997, according to the Employer Survey conducted by RAND, when the average annual premium increase was 1.9 percent, 7 percent of employees were in firms that had decreases of 10 percent or more and 19 percent were in firms whose increases exceeded 10 percent. This highlights the importance of drawing data from a representative sample rather than making inferences from the experience of a single employer.

Overall, HSC site visit teams noted aggressive price shopping by public and private purchasers, who increasingly are more resistant to premium increases and are willing to switch plans to save money. A pervasive theme in all 12 communities was purchasers' pressure to reduce health care expenditures and the competitive response of plans and providers.

Figure 5: Private Insurance Premiums have Declined More Rapidly than Costs Sources: Private Health Insurance Premiums, KMPG Peat Marwick; Health Cost Index, Milliman & Robertson

Employees' Share Rising

Data from national employer surveys show that in the early 1990s employees did not benefit from low premium increases to the extent that employers did, although recent data indicate that the trend is leveling off. Many employers were requiring their employees to pay a higher proportion of health insurance costs. Overall, the average monthly contribution for family coverage for workers in large firms grew from $29 in 1988 to $127 in 1996-a 20 percent increase per year on average. For workers in small firms with fewer than 200 employees, contributions increased considerably more-from $34 in 1988 to $175 in 1996, for a staggering 23 percent average yearly increase. At the same time, total premiums for small firms increased only 8 percent annually.

Perhaps as a result of these growing costs for employees, HSC researchers found that while more small firms are offering health insurance, fewer employees are enrolling (see Figure 6). The percentage of employees in small firms enrolled in their firms' plans decreased from 72 percent in 1989 to 66 percent in 1996. The pattern of this decline is reflected among large firms as well. The result is a net decline in the rate of employment-based health coverage.

Source: Health Insurance Association of America, 1989; KMPG Peat Marwick, 1996