U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on Using Performance Monitoring to Improve Community Health; Durch JS, Bailey LA, Stoto MA, editors. Improving Health in the Community: A Role for Performance Monitoring. Washington (DC): National Academies Press (US); 1997.

Cover of Improving Health in the Community

Improving Health in the Community: A Role for Performance Monitoring.

Show details

5Measurement Tools for a Community Health Improvement Process

Chapter 4 has outlined a community health improvement process (CHIP) through which communities can assess health needs and priorities, formulate a health improvement strategy, and use performance indicators as part of a continuing and accountable process. This chapter reviews in more detail the two kinds of indicators and indicator sets proposed for use in a CHIP. Discussed first is the community health profile , with component indicators proposed by the committee, which can provide a broad overview of a community's characteristics and its health status and resources. The second part of the chapter focuses on the development of indicator sets for performance monitoring, which are intended for use with health improvement strategies for specific health issues. The committee presents some examples that illustrate how communities might approach selecting such performance indicators.

Role for a Community Health Profile

A community health profile is an integral component of the problem identification and prioritization cycle of the community health improvement process described in Chapter 4. The health profile is intended to be a set of indicators of basic demographic and socioeconomic characteristics, health status, health risk factors, and health resource use, which are relevant to most communities.

The committee's proposal is consistent with the efforts of others over the past several years to identify small sets of indicators for key issues. One source of interest has been health promotion and Healthy Cities/Healthy Communities activities by the World Health Organization (WHO, 1986) and others (e.g., Canadian Healthy Communities Project, 1988; National Civic League, 1993). In the United States in particular, the inclusion of 300 indicators in Healthy People 2000 (USDHHS, 1991) led to interest in also selecting a smaller set of indicators that could be used to monitor health status (e.g., CDC, 1991; Stoto, 1992). In other work, a small set of indicators was proposed for monitoring access to health care (IOM, 1993).

The health profile can help a community maintain a broad strategic view of its population's health status and factors that influence health in the community. It is not expected to be a comprehensive survey of all aspects of community health and well-being, but it should be able to help a community identify and focus attention on specific high-priority health issues. The background information provided by a health profile can help a community interpret data on those issues.

A community health profile is made up of indicators of sociodemographic characteristics, health status and quality of life, health risk factors, and health resources that are relevant for most communities; these indicators provide basic descriptive information that can inform priority setting and interpretation of data on specific health issues.

Health profile data can help motivate communities to address health issues. For example, evidence of underimmunization among children or the elderly might encourage various sectors of the community to respond, through ''official" actions (e.g., more systematic provider assessments of patients' immunization status) and through community action (e.g., volunteer groups offering transportation to immunization clinics). Even as raw numbers, these data may be an important signal to a community, especially when small numbers of cases make it difficult to construct meaningful rates. For example, any work-related deaths, births to teenagers, or cases of measles might be a source of concern. Working with small numbers of cases raises potential problems of privacy and confidentiality, which communities must consider. Further discussion of privacy and confidentiality considerations appears later in this chapter. Care should also be taken that evidence of health problems not be used as a basis for negative labels for particular population groups or neighborhoods in a community.

Comparisons based on health profile data may be another source of motivation and may help communities in assessing health priorities as well. These comparisons can be based on measurements over time within an individual community, comparisons with other communities or with state or national measurements, or comparisons with a benchmark or target value such as an objective from Healthy People 2000 (USDHHS, 1991). A variety of specialized compilations of data may provide additional reference points (e.g., Andrulis et al., 1995; Annie E. Casey Foundation, 1996; Wennberg, 1996). The opportunity for such comparisons will be increased if there is widespread agreement across communities on a basic set of standard health profile indicators and their operational definitions.

In making comparisons, however, communities must consider underlying factors that might contribute to observed differences. Some factors, if recognized, can be captured in quantitative form. For example, there might be a greater number of hospitalizations in an older population than in a younger population even though the age-specific rates are the same in both groups. Less easily addressed is the effect on the validity of comparisons among communities of different physical, social, political, and cultural contexts and different local needs and priorities, all of which may influence community profile indicators and, for some, argue against standard indicator sets (Hayes and Willms, 1990). (See Appendix B for further discussion of methodological issues in selecting and using health profile and performance indicators.)

The committee emphasizes that communities should update their health profile data on a regular basis to maintain an accurate picture of community circumstances, including identifying positive or negative changes that might influence health improvement priorities. The health profile is not, however, intended to be a tool specifically to monitor changes in stakeholder performance or to establish responsibility and accountability for health outcomes. Some of the indicators that are included in a profile might, however, serve as performance indicators if they are applied to other CHIP activities. Immunization rates, for example, are a useful community health descriptor but could also be monitored as an outcome measure for targeted efforts to reduce the risk of vaccine-preventable disease.

Proposed Indicators for a Community Health Profile

To promote community use of health profiles, the committee is proposing a basic set of 25 indicators (see Table 5-1). They provide descriptive information on a community's demographic and socioeconomic characteristics and highlight important aspects of health status and various health determinants, including behavior, factors in the social and physical environments, and health care. Some the indicators include multiple measures within a broader category (e.g., causes of death and incidence of infectious diseases). Appendix 5A reviews each indicator individually.

TABLE 5-1. Proposed Indicators for a Community Health Profile.


Proposed Indicators for a Community Health Profile.

Selection of Community Health Profile Indicators

The committee's selection of indicators reflects consideration of several factors. Measures were sought that would be relevant across a broad range of communities. Recognizing the diversity among communities in health needs, priorities, and resources, the committee selected a limited number of indicators that could be expected to be widely applicable. The list draws extensively from the "consensus set" of indicators for assessing community health status (CDC, 1991) that was developed in response to Healthy People 2000 Objective 22.1. This objective calls for developing a set of health status indicators appropriate for use by federal, state, and local health agencies and implementing them in at least 40 states by the year 2000 (USDHHS, 1991). The committee gave these indicators a high priority because they and Healthy People 2000 have had an important influence on community health assessment activities since 1991. The committee agreed, however, that the consensus indicators per se were not sufficient to constitute an adequate community health profile.

The committee considered four other factors in selecting indicators: consistency with the field model framework for the determinants of health; attention to the health needs of specific populations; existence of a measure with an operational definition; and availability of data. The mix of indicators was also examined to ensure relevance across the age spectrum (Stoto, 1992). Table 5-2 summarizes the filed model domains and current or potential sources of data for each proposed health profile indicator.

TABLE 5-2. Features of Proposed Community Profile Indicators.


Features of Proposed Community Profile Indicators.

The broad perspective on health embodied in the field model (Evans and Stoddart, 1994) is a fundamental component of the committee's approach to health improvement and performance monitoring. For the community health profile, proposed indicators were mapped to the domains of the field model (social and physical environment, genetic endowment, behavior, disease, health care, health and function, prosperity, and well-being) to identify potential gaps and to assess the distribution of indicators across domains. Only the domain of genetic endowment is not represented directly; its contribution can be seen, however, in indicators such as infant mortality, cardiovascular disease mortality, and obesity.

In its selections, the committee favored measures that are in use and have a recognized operational definition or lend themselves to the construction of such a definition. Being able to specify clearly how an indicator is measured will help communities determine what data they need and will help them identify points of comparison with other communities and at state and national levels. For some of the selected indicators, generally recognized measures have not been established. This applies in particular to the indicators on satisfaction with the quality of life in the community and with the health care system in the community. The committee felt, however, that these indicators were of sufficient importance for understanding health in the broadest sense that they should be proposed for inclusion in a community health profile to encourage the development of suitable measures. The Centers for Disease Control and Prevention (CDC) has developed survey questions on the influence of personal health on quality of life that are now in use in the Behavioral Risk Factor Surveillance System (BRFSS) and is attempting to identify community-level indicators of health-related quality of life (Hennessy et al., 1994; Moriarty, 1996). Once valid and reliable measures are available, issues of data collection can be addressed.

Availability of and Access to Data

The availability of data is a special concern at the community level. For most of the health profile indicators proposed by the committee, data are already being collected at the state or national level, but not necessarily by communities themselves or in a form that can produce community-level information or as frequently as might be desired. Few communities have the financial resources or expertise to collect such data on a routine basis or to perform the additional analysis that may be needed to make available data meaningful at the community level. In some cases, however, opportunities may exist to develop sources of data for communities. In selecting indicators for the community profile, the committee frequently chose to suggest such potential sources of data rather than limit its list of indicators to only those for which community-level data are typically available now.

As noted in Chapter 4, the committee believes that states have an obligation to ensure that communities have access to the data needed to construct health profiles. Some states have already assumed this responsibility, and an Assessment Initiative managed by the National Center for Health Statistics (NCHS, 1995a) is assisting other states in developing the capacity to provide such data. Information is often produced in printed reports, but some states such as Illinois and Massachusetts are also developing data systems that give local health departments online access to data. In Massachusetts, the MassCHIP (Massachusetts Community Health Information Profile) data system makes community-level health profile data available to the public as well as to the state's community health network areas (see Box 5-1 for additional information on MassCHIP). Minnesota is providing electronic access to county data from its Substance Abuse Monitoring System (Minnesota Department of Human Services, 1995). Evolving computer and communications technologies can be expected to facilitate access to information not only within states but across the country. Some states, federal agencies, and private companies are already making data available through the Internet.

Box Icon

BOX 5-1

Massachusetts Community Health Information Profile. The Massachusetts Department of Public Health (MDPH) has established as a priority improving the availability of health status data for community-based health promotion and prevention. In 1996, MDPH (more...)

One promising source of community-level data on adults may be the BRFSS, through which the states and CDC collaborate to produce state estimates for a variety of health status, health behavior, and health risk topics (CDC, 1993). Modifications to the sampling methods and inclusion of additional questions could make it possible to generate county or other substate estimates. Illinois, for example, is adopting a program to produce periodic county-level estimates by oversampling different groups of counties for each BRFSS round. In Massachusetts, similar arrangements are being made for cities and regions of the state. The school-based Youth Risk Behavior Surveillance System (YRBSS)—a collaborative effort involving states, cities, and the CDC (1995)—may lend itself to similar approaches to generating community data for adolescents. Neither the BRFSS nor the YRBSS as they are currently designed will provide information on children. To obtain such data, modifications of those surveys or separate data collection methods will be needed. If local data remain unavailable or are not feasible to obtain, communities that are similar to the state as a whole may find some state-level data useful.

Adding location identifiers (e.g., zip codes, census tracts) to survey and other types of data could improve their usefulness at the community level. This approach may be especially valuable for some forms of environmental risk monitoring. The additional information may also make it possible to link data from state sources with local data systems such as an immunization registry. The committee strongly supports more extensive use of such ''geocoding," particularly for data collected by states.

Community-level data collection may also be possible—perhaps essential—for obtaining some types of information. NCHS (1995b) is testing the feasibility of a telephone survey to obtain data related to the consensus indicators, particularly the supplemental indicators for which data sources were not available at the time the consensus indicators were issued. The committee has included some of these supplemental indicators in its health profile.

Because most of the proposed health profile indicators rely on population-based measures, health departments and other public agencies with responsibilities for an entire community will tend to be the principal sources of needed data. Nevertheless, health plans, insurers, employers, and others in the private sector could contribute to community data resources, particularly for numerator data needed to calculate rates. Rate calculations pose other challenges as well. For many indicators, small numbers of cases at the community level will mean that calculation of stable rates will require aggregating data over multiple years. If, however, data are collected only infrequently (e.g., every five years), aggregation may not be practical, either because of the delay created in producing a usable measurement or because circumstances in the community change sufficiently that combining data could be misleading. Communities may also need assistance in developing intercensal population estimates accurate enough to be used as rate denominators. These estimates are especially important if the population is changing rapidly in size or composition.

Further Development of the Community Health Profile

The community health profile proposed by the committee should be viewed as a starting point for further development, not a final product. The indicators chosen reflect the committee's judgment in balancing three considerations: (1) importance in shaping or contributing to understanding community health, (2) usefulness across a broad range of communities, and (3) feasibility of measurement. Communities may, through their health improvement activities, identify topics of local importance that should supplement the basic profile. Indicators that address issues beyond the traditional realm of "health" (e.g., education, literacy, employment, crime, housing, community participation) may be relevant. For example, the categories of measures for the National Civic League's (1993) Healthy Community Indicators include health, family income, housing and homelessness, food assistance, child care, education, youth employment, transportation, public safety, and environmental issues. The Sustainability Indicators developed by the Regional Municipality of Hamilton-Wentworth (1996) in Ontario, Canada, include measures such as air quality, water and electricity consumption, voter turnout, and applications for affordable housing.

Access to a wide array of data, perhaps through state sources, can also support an expanded health profile. In expanding the profile, however, communities should not be aiming to produce a comprehensive health assessment tool. Such assessments are valuable, but if resources are limited, comprehensive assessments should probably be prepared less frequently than updates to a health profile. For a profile, communities should focus on indicators that can contribute most to an understanding of the population's health status and the factors that affect it in a positive or negative way. As was the case with the basic profile, the field model will be a useful guide for examining a broader array factors that may be determinants of a community's health and for selecting indicators to add to a profile.

Part of the committee's intention in proposing a basic set of indicators for a community profile is to encourage the development of common indicator definitions and common practices in data collection, analysis, and reporting that will facilitate comparisons over time and among communities. State programs that provide data to communities can promote this kind of comparability. Activities at the national level related to Healthy People 2000 and the consensus indicators, including reporting requirements for some block grants (e.g., CDC, 1994; MCHB, 1995), should also contribute to standardization of measures suitable for community health profiles. In addition, the work being done to develop indicators for state reporting for the proposed federal public health Performance Partnership Grants (PPGs) can also be expected to promote standardization (USDHHS, no date; NRC, 1996).

The committee encourages reexamination and revision of its basic community profile. Individual indicators in the current set might be modified as new or better data and measures become available. The profile might also evolve toward a greater focus on positive measures of health and health promoting features of individual behavior and the community environment. For example, measures on diet and exercise, topics for which questions have been developed for the BRFSS, might be considered. In general, however, such measures are less well developed than those for health "problems." Work is also needed to further the development of community-level measures to supplement those for individuals (Patrick and Wickizer, 1995).

A formal process, which might be organized by federal agencies, national professional organizations, or foundations, could promote the development and improvement of measures suitable for community-level data and the adoption of standard measures. Participation by a broad array of public and private stakeholders representing national, state, and local perspectives should be encouraged.

The committee also sees a need for a variety of forms of technical assistance that can help communities understand how to use health profile indicators and obtain appropriate data. States may be able to provide some of the assistance that communities need, but states themselves may benefit from technical assistance in these areas. National efforts such as those suggested for the development of community-level measures would also be useful for improving analytic techniques and developing resources for technical assistance.

As presented here, the community health profile is based on a "community" defined by geographic or civic boundaries, frequently a county or city. This reflects the current form in which data are generally available and not a necessary or preferable basis on which to define a community. Discussions at the committee's workshops emphasized that data for much smaller units (e.g., neighborhoods) are often needed to generate support for health improvement activities. The committee encourages the development of data for a variety of "community" units. It believes that states should work toward developing interactive electronic data systems that will permit users to define the specific population, including demographic or socioeconomic groupings, for which they want data. The MassCHIP system (see Box 5-1), for example, is designed to provide data on cities and towns, neighborhoods in three large cities, predefined regions such as the state's Community Health Network Areas, and user-defined combinations of cities, towns, and regions (Massachusetts Department of Public Health, 1995).

Indicator Sets for Performance Monitoring for Specific Health Issues

Communities, through mechanisms such as the problem identification and prioritization cycle of a CHIP (see Chapter 4), need to identify the health issues that key stakeholders consider important. Data from a community's health profile can point to issues, but health priorities may also be identified by other means, such as community meetings or surveys. In working from the broad perspective of the field model, critical "health" issues may also be found not only among conditions that create a substantial burden in terms of illness or costs of care but also in areas such as education and housing. Some issues may be of great concern but will not yet be suitable choices for more targeted health improvement activities because effective interventions have not been developed.

Once a health issue has been selected, a CHIP moves on to the analysis and implementation cycle, and a community's information needs expand from the descriptive measures in a community profile to the more "actionable" indicators that are crucial to performance monitoring and health improvement activities. As noted in Chapter 4, communities will have to assemble a set of performance indicators to address the multiple dimensions of a health issue. Discussed here are factors that should be considered in selecting sets of indicators for issue-specific performance monitoring. Prototype indicator sets in Appendix A to this report illustrate how communities might use the committee's approach.

For the community health improvement process, a performance indicator provides a concrete measure of a specific capacity, process, or outcome related to an accountable entity that is part of a defined health improvement strategy for a specific health issue. Such indicators can be used to measure performance at varying levels of specificity: a community as a whole; particular categories of accountable entities (e.g., health departments, schools, or insurers); or specific entities in a community (e.g., a specific school or health plan). A set of performance indicators is used to assess the multiple dimensions of a health issue and monitor the contributions of various accountable entities to the health improvement strategy.

Assessing the Scope of an Issue

Almost any health issue will have many dimensions and present many possible opportunities to respond. As an initial step in the analysis and implementation cycle of a CHIP, a community will need to think broadly about the nature of the problem, what can be done, who can take action, and what indicators can track progress most effectively. The field model provides a helpful framework for accomplishing the kind of systematic review that is needed. To gain a clear understanding of the features of a particular health issue so that an effective intervention strategy can be developed, it may be useful to gather additional information from key stakeholder groups. A community that wants to reduce adolescent tobacco use, for example, will need information on topics such as the age at which use begins, how adolescents obtain tobacco products, and the kinds of school-based prevention programs available.

As a community moves on to the process of identifying potential performance indicators, it should specifically include consideration of (1) the domains of the field model that could be addressed by those indicators and (2) the potential to engage the interest and action of a variety of community stakeholders. A narrow focus on any one stakeholder group or health factor may limit opportunities for effective action. For example, efforts to reduce the adverse impact of depression that look only at the quality of care provided by mental health specialists will neglect the contributions that might be made by primary care providers or by activities based in settings such as schools and workplaces.

Considering the Health Field Model

A narrow view of health interventions might be limited to the diagnosis and treatment of disease. Examining an issue in the framework of the field model, which presents health and well-being as the product of a more complex mix of forces, can point to a broader array of possible interventions and related performance indicators. Communities may find it beneficial to use the field model in conjunction with other assessment tools, such as the analysis of risk factors and direct and indirect contributing factors suggested by APEXPH: Assessment Protocol for Excellence in Public Health (NACHO, 1991). An assessment of a health issue may, however, point to important concerns for which satisfactory indicators and data have not yet been developed, often because of gaps in our understanding of the complex processes that produce "health."

Engaging Stakeholders

Successful health improvement efforts in a community will require the interest and support of a variety of stakeholder groups and, for some stakeholders, may require changing their responsibilities and activities. Therefore, health issues identified as community priorities and the performance indicators selected to assess progress should engage key stakeholders who must act or who can encourage action. The mix of stakeholders and their degree of involvement can be expected to vary depending on the health issue being addressed. Nursing homes, for instance, could be expected to be key participants in efforts targeting the health of the elderly but would probably have little role in improving prenatal care.

Selecting Performance Indicators

Many potential performance indicators, generally "process" and "outcome" measures,1 will emerge in discussions about an issue, but some will be more appropriate than others. Selecting those that will be used is a critical stage in a CHIP. Sofaer (1995) points out that indicator selection is a normative process, reflecting community expectations as to which aspects of a health issue are important and what stakeholder actions should achieve. A CHIP and the community coalition that is at its core should provide a framework within which a community can reach agreement on the values and expectations to be represented in performance indicators.

Indicator selection should also reflect a strategic consideration of the value of individual indicators and of the collection of indicators to be used in connection with a specific health issue (Sofaer, 1995). Individual indicators become more valuable to the extent that they are effective proxies for multiple dimensions of performance. A set of indicators will usually be needed to cover a range of relevant performance areas and must be assembled carefully to assure that, together, the indicators effectively serve the process of improving the community's health. The set should appeal to many stakeholders and reflect broad consideration of the domains of the field model, but it should be limited to a comprehensible number of indicators. Too many indicators become distracting and, in practical terms, could make collecting and analyzing data prohibitively burdensome (Sofaer, 1995).

Operational implications and costs of data collection and analysis also must be considered in selecting indicators and indicator sets. Even though indicators may be formulated with the intention of promoting actions that will have positive effects on community health, they must be based on an accurate understanding of their effect in the setting in which they will be applied. For example, reducing the number of cigarette vending machines as a way to limit youth access to tobacco will not have the anticipated impact if teenagers buy most of their cigarettes in convenience stores. It also is possible to frame indicators in a way that creates "perverse incentives" for action, which produce a "better" measured result but do not achieve intended health goals. For example, lower rates of sexually transmitted diseases might be "achieved'' through less complete reporting rather than through true reductions in disease rates.

A reasonable balance must be struck between the information value of an indicator and the cost of collecting the necessary data. A conceptually appropriate indicator will not be helpful if a community cannot afford to obtain the data it requires. Costs of data generation may include designing data collection instruments, collecting or locating data, analyzing and summarizing the results, and reporting information to the community. In some communities and for some indicators, these activities may require new resources. In other cases, it may be possible to apply existing resources (e.g., funds, expertise, data systems) to producing CHIP data.

Another concern is how time factors are addressed in performance monitoring. Communities must approach performance monitoring with an understanding of when to expect measurable effects from health improvement. It is generally easier to examine factors that affect health in the short term (e.g., vaccination or care for acute illness), but some important influences on health operate over much longer time frames. For example, changes in lung cancer rates can lag changes in smoking patterns by 20–30 years, and the health benefits of interventions targeted at critical developmental periods in early childhood may not be seen until adolescence or adulthood. Interventions with important long-term benefits should not be neglected in favor of those that operate more quickly. Indicators based on intermediate goals such as changes in risk factors (e.g., decreased prevalence of smoking) can help bridge the period until changes in health outcomes can be measured.

In selecting indicators, communities will also have to consider factors such as how issues manifest themselves (e.g., social isolation among the elderly could be a function of lack of transportation but might also result from fear of street crime); what information resources are available; and what actions are organizationally, socially, politically, and economically feasible within the community (e.g., gun safety programs might be acceptable when gun control is not). These concerns should be addressed through the community processes described in Chapter 4.

Selection Criteria

The committee identified several specific criteria to consider in selecting individual performance indicators.2 Ideally, every indicator should satisfy them, but compromises may be necessary until improvements such as better measures and data systems or stronger scientific evidence are available. Communities may need to act cautiously in the face of such limitations but should not neglect important health issues that cannot yet be addressed through quantitative approaches to performance monitoring.

The committee proposes the following criteria for selecting indicators:

  • Established validity and reliability. To be of value, a performance indicator must be valid for its intended use; that is, it must measure what it purports to measure. It is also essential that performance indicators be reliable, that is, producing consistent responses when measured by different people or at different times. Indicators must also demonstrate validity and reliability in varying cultural contexts. These basic qualities of a good measure are particularly important in a monitoring system where progress, or lack thereof, is being followed closely and the results will affect important decisions.
  • Evidence-based link between performance and health improvement. Performance indicators measure how well specific actions are being carried out by those who accept responsibility for them. There should be (under the best of circumstances) clear scientific evidence that the action being monitored will, indeed, lead to improvement in health. In some cases, available evidence may not be conclusive, but expert judgment represented in sources such as clinical practice guidelines (e.g., see IOM, 1992; U.S. Preventive Services Task Force, 1996) may suggest actions that could be expected to produce desired effects. Without such evidence or consensus in expert judgment, it is not reasonable to expect accountability for health improvement when, even under ideal circumstances, it may not be possible for the action taken to have the desired impact.
  • Responsibility and accountability for performance. A critical element of performance monitoring is identifying where responsibility and accountability lie for actions that can improve health. It should be possible to link performance indicators to specific community stakeholders who have accepted or been assigned responsibility for some aspect of health improvement. In some cases, a stakeholder may have responsibility for a defined portion of the total population (e.g., health plans and their enrolled members, schools and enrolled students). When similar health needs exist in the remainder of the population, communities will have to determine where responsibility for serving that portion of the population lies.

Under some circumstances, a stakeholder may have to assume responsibility for producing or assuring the existence of an enabling precondition for achieving health improvement, rather than assuming more direct responsibility for the health outcome itself. Determining whether an "intermediate" activity such as this will be monitored at the community level or by an individual stakeholder organization will depend on a community's approach to the health issue and the nature of the precondition to be achieved.

  • Robustness and responsiveness to change in health system performance, particularly in targeted populations. A performance indicator must be able to detect the effect of reasonably small changes in the performance system so that progress can be measured, even in small increments. If the performance indicator is unable to detect small initial changes, failure may be declared prematurely. Consideration must also be given to whether the indicator can reflect the impact of system changes on small subgroups in the population. In addition, indicators should be sufficiently stable and well defined that they are not subject to substantial random variation.
  • Availability of data in a timely manner at a reasonable cost. The need to collect performance indicator data on a recurring basis makes ease and cost of collection important considerations. Because financial constraints are a concern for most communities, it will be imperative that performance indicators be measurable at reasonable cost and in a timely manner. Communities should consider the roles that the public and private sectors should each have in supporting data collection, analysis, and reporting.
  • Inclusion in other indicator sets (monitoring sets). Some health-related indicator systems are already being used to assess performance, although rarely community-wide. Using existing indicators makes it possible to benefit from the indicator development experience of the parent group and to avoid duplication of effort or variation in specification that may hinder comparisons. A few of these indicator sets include HEDIS, the Health Plan Employer Data and Information Set (NCQA, 1993, 1996), which is focused on managed care organizations and includes some measures framed specifically for Medicaid and Medicare enrollees; the accountability measurement sets being developed by the Foundation for Accountability (FAcct, 1995, 1996) for specific health issues; Healthy People 2000 (USDHHS, 1991), which sets out roughly 300 health promotion and disease prevention objectives and is the starting point for objectives outlined in the Healthy People consensus indicators (CDC, 1991); Healthy Communities 2000 (APHA et al., 1991); and the measures required for reporting on some federal grants (e.g., CDC, 1994; MCHB, 1995). The indicators established for the proposed PPGs may also prove helpful. Some of the additional indicator sets that communities might consult are noted in Chapter 4.

Using Indicator Sets

Once communities establish performance indicator sets for specific health issues, they are able to move further through the health improvement process outlined by the committee. The indicator data should reflect whether appropriate actions are being taken by accountable entities and whether those actions are having the intended health effect. To interpret performance monitoring results, communities will have to take their specific circumstances into consideration. The resources available to a community, the mix of risk factors, and the interventions chosen will all influence the results achieved through a given health improvement strategy. Information provided by the performance indicators should guide subsequent steps: moving on to a new health issue, continuing or modifying the current effort, or perhaps returning to an earlier stage in the process to reassess the intervention strategy and the appropriate indicators to use.

Prototype Performance Indicator Sets

To illustrate the application of its proposed approach to performance monitoring and indicator selection, the committee has assembled, with advice from outside experts, examples of indicator sets for several health issues: breast and cervical cancer, depression, elder health, environmental and occupational lead poisoning, health care resource allocation, infant health, tobacco and health, vaccine-preventable diseases, and violence. Appendix A presents for each topic a discussion of the health issue, the application of the field model and stakeholder interests, and the selection of a limited number of specific performance indicators. Comments are offered on likely sources of data and special considerations in using specific indicators. Table 5-3 shows the relationship to the field model domains of the indicators suggested for health improvement activities for vaccine-preventable diseases.

TABLE 5-3. Field Model Mapping for Sample Indicator Set for Vaccine-Preventable Diseases.


Field Model Mapping for Sample Indicator Set for Vaccine-Preventable Diseases.

These health issues were selected to be generally representative of the spectrum of health concerns facing many communities. Most are associated with significant morbidity or health care costs. The committee's selections were also made to illustrate varying perspectives from which health issues might be viewed, including factors affecting population groups (infant and elder health); acute and chronic illness (breast and cervical cancer, depression); prevention and health promotion (tobacco and vaccine-preventable diseases); environmental and occupational health risk (lead exposure); operation of the health care system; and broad societal issues that have health implications (violence). Similarly, the committee selected health issues that present an opportunity for a variety of stakeholders to respond, including public health and other government agencies, health care providers, schools employers, community groups, and individuals.

The committee's aim has been to demonstrate how performance indicators can be selected and to present credible indicator sets as models for work on a variety of other health issues as well as the ones discussed here. The committee is not attempting to prescribe intervention strategies or specific indicator sets for these health issues because it cannot adequately address the unique combination of circumstances that each community will have to consider. Instead, the examples use community-level indicators to illustrate issues discussed by the committee. Individual communities will have to formulate performance indicators that are based on performance expectations for their particular accountable entities and that reflect specific needs and resources. Some of factors to be considered include who provides specific services, what data are available from what sources, and whether important stakeholders are willing to accept responsibility for particular tasks.

Privacy and Confidentiality

The performance monitoring component of the CHIP outlined by the committee will require increased access to potentially sensitive data such as an individual's income level, employment status, medical diagnoses (e.g., HIV status, other sexually transmitted diseases, genetic conditions, mental illness), and lifestyle information (e.g., sexual practices, drug and alcohol use). Ensuring that this information is not misused must be a priority.

Matters of both privacy and confidentiality must be considered. Privacy can be defined as the capacity of the individual to determine which personal information is communicated to whom (Westin, 1967). In the health care setting, privacy refers to the implicit right of an individual to have control over personal medical information. confidentiality, however, refers to the duty of those who hold information about others to protect that information from inappropriate disclosure to third parties. Underlying this duty is the knowledge that uncontrolled access to some types of personal information can result in harm to individuals (IOM, 1994).

Data in the form of person-identified (or identifiable) records are the most vulnerable, but access to such data can be vital for the success of some activities, particularly linking information from separate sources. Immunization registries, for example, must be able to update a child's record each time a vaccine dose is administered, regardless of who the provider is. Techniques that create unique but anonymous identifiers can make it possible to omit personal information such as name, address, and social security number from stored records. Risks of misuse are lower for aggregated data and for individual records that do not include personal identifiers. Even in this form, however, distinctive combinations of characteristics such as age, race, occupation, and diagnosis could suggest the probable identity of an individual in a community. Thus, policies are needed regarding the level of detail provided even in supposedly anonymous data.

Developing appropriate procedures to safeguard data from misuse is important for two reasons—it will prevent harm to individuals and it will help maintain the integrity of the data system (Gostin, 1995). All states have privacy protection laws regarding health data held by government agencies (e.g., communicable disease reports); the specific protections and penalties for violations vary from state to state (Gostin et al., 1996). Various state provisions also protect privately held health data, but federal legislation such as the Employee Retirement Income Security Act (ERISA) may take precedence without offering protection comparable to state laws (IOM, 1994). Federal legislation that would provide more comprehensive protection for health data has been proposed (e.g., U.S. Congress, 1995, 1996).

Recommendations from the Institute of Medicine (IOM, 1994) report Health Data in the Information Age, which outlines a role for community-based ''health data organizations" (HDOs), aim for a balance between ensuring confidentiality of information and the security of automated databases and providing access to information for activities that will improve the health of communities. In particular, the report recommends that HDOs have explicit mechanisms for developing and implementing policies and procedures governing the acquisition and dissemination of information that will provide for protection of privacy and confidentiality. The report also recommends passage of preemptive federal legislation that is designed to ensure that data systems protect privacy and confidentiality and would impose penalties for inappropriate use or release of data (IOM, 1994).

Communities should ensure that a CHIP incorporates adequate protection for all data that are used. Access to technical assistance from state agencies and experts in academia and the private sector may help communities establish policies and implement technologies that provide needed protections.


The health improvement process outlined by the committee depends heavily on access to information provided by indicators such as those discussed in this chapter. Both the broad perspective of a community health profile and the narrower focus of issue-specific indicator sets are needed. To aid communities in assembling and using indicators and indicator sets, the committee has proposed specific indicators for a health profile and has illustrated how communities might develop indicator sets for specific health issues.

Communities will have to translate these proposals into the realities of their particular circumstances. An immediate aim should be to identify a manageable number of indicators to be included in a community profile and to begin collecting and publishing data for those indicators on a routine basis. Over time, a community will have an information resource that allows it to see whether strengths are being preserved, progress is being made, or problems are emerging. More challenging will be the development of appropriate measurement tools to support an issue-specific performance monitoring process.

The guidance offered in this chapter and, by example, in the prototype indicator sets in Appendix A should help communities begin. Further work should be undertaken at the national and state levels to develop ways to make expertise in measurement and analysis available to communities that desire it.


  • Andrulis, D.P., Ginsberg, C., Shaw-Taylor, Y., and Martin, V. 1995. Urban Social Health: A Chart Book Profiling the Nation's One Hundred Largest Cities. Washington, D.C.: National Public Health and Hospital Institute.
  • Annie E. Casey Foundation. 1996. KIDS COUNT Data Book: State Profiles of Child Well-Being. Baltimore: Annie E. Casey Foundation.
  • APHA (American Public Health Association), Association of Schools of Public Health, Association of State and Territorial Health Officials, National Association of County Health Officials, United States Conference of Local Health Officers, Department of Health and Human Services, Public Health Service, Centers for Disease Control. 1991. Healthy Communities 2000: Model Standards. 3rd ed. Washington, D.C.: APHA.
  • Canadian Healthy Communities Project . 1988. Canadian Healthy Communities Project Start-up Kit. Ottawa: Canadian Healthy Communities Project.
  • CDC (Centers for Disease Control and Prevention) . 1991. Consensus Set of Health Status Indicators for the General Assessment of Community Health Status—United States. Morbidity and Mortality Weekly Report 40:449–451. [PubMed: 2056991]
  • CDC. 1993. Special Focus: Behavioral Risk Factor Surveillance—United States, 1991. Morbidity and Mortality Weekly Report 42(SS-4).
  • CDC. 1994. Organization and Implementation of the Uniform Data Sets (Preventive Health and Health Services Block Grant). Atlanta, Ga.: U.S. Department of Health and Human Services. November 1. (mimeo)
  • CDC. 1995. Youth Risk Behavior Surveillance—United States, 1993. Morbidity and Mortality Weekly Report 44(SS-1).
  • Donabedian, A. 1980. Explorations in Quality Assessment and Monitoring . Vol. 1. The Definition of Quality and Approaches to Its Assessment. Ann Arbor, Mich.: Health Administration Press.
  • Donabedian, A. 1982. Explorations in Quality Assessment and Monitoring . Vol. 2. The Criteria and Standards of Quality. Ann Arbor, Mich.: Health Administration Press.
  • Donabedian, A. 1985. Explorations in Quality Assessment and Monitoring . Vol. 3. The Methods and Findings of Quality Assessment and Monitoring: An Illustrated Analysis. Ann Arbor, Mich.: Health Administration Press.
  • Evans, R.G., and Stoddart, G.L. 1994. Producing Health, Consuming Health Care. In Why Are Some People Healthy and Others Not? The Determinants of Health of Populations. R.G. Evans, editor; , M.L. Barer, editor; , and T.R. Marmor, editor. , eds. New York: Aldine De Gruyter.
  • FAcct (Foundation for Accountability). 1995. Guidebook for Performance Measurement: Prototype. Portland, Ore.: FAcct. September 22.
  • FAcct. 1996. In Practice—Diabetes. Accountability! (Fall):1–29.
  • Gostin, L.O. 1995. Health Information Privacy. Cornell Law Review 80(3):451–528. [PubMed: 11660159]
  • Gostin, L.O., Lazzarini, Z., Neslund, V.S., and Osterholm, M.T. 1996. The Public Health Information Infrastructure: A National Review of the Law on Health Information Privacy. Journal of the American Medical Association 275:1921–1927. [PubMed: 8648874]
  • Hayes, M.V., and Willms, S.M. 1990. Healthy Community Indicators: The Perils of the Search and the Paucity of the Find. Health Promotion International 5(2):161–166.
  • Hennessy, C.H., Moriarty, D.G., Zack, M.M., Scherr, P.A., and Brackbill, R. 1994. Measuring Health-Related Quality of Life for Public Health Surveillance. Public Health Reports 109:665–672. [PMC free article: PMC1403555] [PubMed: 7938388]
  • IOM (Institute of Medicine). 1992. Guidelines for Clinical Practice: From Development to Use. M.J. Field, editor; and K.N. Lohr, editor. , eds. Washington, D.C.: National Academy Press. [PubMed: 25121254]
  • IOM. 1993. Access to Health Care in America. M. Millman, editor. , ed. Washington, D.C.: National Academy Press. [PubMed: 25144064]
  • IOM. 1994. Health Data in the Information Age: Use, Disclosure, and Privacy. M.S. Donaldson, editor; and K.N. Lohr, editor. , eds. Washington, D.C.: National Academy Press. [PubMed: 25144051]
  • Massachusetts Department of Public Health. 1995. Massachusetts Community Health Information Profile. Boston: Massachusetts Department of Public Health. (brochure)
  • MCHB (Maternal and Child Health Bureau). 1995. Annual Report Guidance for the Maternal and Child Health Services Block Grant Program. OMB No. 0915-0172. Rockville, Md.: U.S. Department of Health and Human Services.
  • Minnesota Department of Human Services. 1995. The Substance Abuse Monitoring System. research NEWS (January). St. Paul: Minnesota Department of Human Services, Chemical Dependency Division.
  • Moriarty, D. 1996. CDC Studies Community Quality of Life. NACCHO News 12(3):10, 13.
  • NACHO (National Association of County Health Officials). 1991. APEXPH: Assessment Protocol for Excellence in Public Health. Washington, D.C.: NACHO.
  • National Civic League. 1993. The Healthy Communities Handbook. Denver: National Civic League.
  • NCHS (National Center for Health Statistics). 1995. a. The CDC Assessment Initiative: A Summary of State Activities. Healthy People 2000: Statistics and Surveillance. No. 7 (October). Hyattsville, Md.: U.S. Department of Health and Human Services.
  • NCHS. 1995. b. Pilot Test of Community Survey. Healthy People 2000: Activity Update. (Brock, B.M. A Telephone Survey Methodology for Local Health Departments' Community Health Status and Risk Factor Assessments Related to Healthy People 2000. DHHS/PHS/CDC Contract #200-94-7064.) Hyattsville, Md.: U.S. Department of Health and Human Services.
  • NCQA (National Committee for Quality Assurance). 1993. Health Plan Employer Data and Information Set and User's Manual, Version 2.0 (HEDIS 2.0). Washington, D.C.: NCQA.
  • NCQA. 1996. HEDIS 3.0 Draft for Public Comment. Washington, D.C.: NCQA.
  • NRC (National Research Council). 1996. Assessment of Performance Measures in Public Health. Phase 1 Report. Draft for Comment. Washington, D.C.: National Academy Press.
  • Patrick, D.L., and Wickizer, T.M. 1995. Community and Health. In Society and Health. B.C. Amick, editor; , s. Levine, editor; , A.R. Tarlov, editor; , and D.C. Walsh, editor. , eds. New York: Oxford University Press.
  • Perrin, E.B. 1995. SAC Instrument Review Process. Medical Outcomes Trust Bulletin 3(4):1, I–IV.
  • Regional Municipality of Hamilton-Wentworth. 1996. Signposts on the Trail to Vision 2020: Hamilton-Wentworth Sustainable Indicators, 1996. Hamilton. Ont.: Regional Municipality of Hamilton-Wentworth Environment Department.
  • Sofaer, S. 1995. Performance Indicators: A Commentary from the Perspective of an Expanded View of Health. Washington, D.C.: Center for the Advancement of Health.
  • Stoto, M.A. 1992. Public Health Assessment in the 1990s. Annual Review of Public Health 13:59–78. [PubMed: 1599602]
  • U.S. Congress, House of Representatives. 1996. H.R. 3482: Medical Privacy in the Age of New Technologies Act of 1996. 104th Congress, 2nd Session, May 16.
  • U.S. Congress, Senate. 1995. S. 1360: Medical Records Confidentiality Act of 1995. 104th Congress, 1st Session, October 24.
  • USDHHS (U.S. Department of Health and Human Services). 1991. Healthy People 2000: National Health Promotion and Disease Prevention Objectives. DHHS Pub. No. (PHS) 91-50212. Washington, D.C.: Office of the Assistant Secretary for Health.
  • USDHHS. No date. Performance Measurement in Selected Public Health Programs: 1995–1996 Regional Meetings. Washington, D.C.: Office of the Assistant Secretary for Health.
  • U.S. Preventive Services Task Force. 1996. Guide to Clinical Preventive Services. 2nd ed. Baltimore: Williams and Wilkins.
  • Wennberg, J., editor. , ed. 1996. The Dartmouth Atlas of Health Care. Chicago: American Hospital Press.
  • Westin, A.F. 1967. Privacy and Freedom. New York: Atheneum.
  • WHO (World Health Organization). 1986. A Discussion Document on the Concept and Principles of Health Promotion. Health Promotion 1(1):73–76. [PubMed: 10286854]

Appendix 5A. Proposed Community Health Profile Indicators


Distribution of the population by age and race or ethnicity.

Data on the basic demographic characteristics of a community are important for understanding current or potential health concerns. For example, a community that has a significant percentage of young families may have a special interest in health issues related to children, pregnancy, teenagers, and injuries, whereas an older community may need to address health issues related to health care resources and utilization, and chronic disease associated with aging. The demographic composition of the population should be understood because significant disparities in health status between minority and nonminority populations may be due to factors including economic resources, health care access, discrimination, and genetic susceptibility to disease. Field model domains: individual behavior, genetics, and social environment. Data sources: decennial census; states may also develop intercensal estimates for communities.


Number and proportion of persons in groups such as migrants, the homeless, or the non-English speaking, for whom access to community services and resources may be a concern.

Subpopulations such as migrants, the homeless, or those who do not speak English are at greater risk for more significant health problems than the general population, may have greater difficulty gaining access to community services and resources, and may benefit from a variety of specialized responses. If a community has a large population of this type, then an attempt should be made to collect health indicator data for that group. In most cases, however, special populations are small, which necessitates special care in the analysis of group-specific data. The size and composition of these populations may change more rapidly than the rest of the population, so care should also be exercised in using data that are not current. Field model domains: individual behavior, social environment, physical environment, and prosperity. Data sources: decennial census; local agencies that serve special populations. Caution may also be needed in using census data if there is reason to believe that a group may have been undercounted relative to others in the community.


Number and proportion of persons aged 25 and older with less than a high school education.

Adults with less than a high school education can be at increased risk of health problems because of illiteracy, low-paying jobs that do not provide health insurance, lack of health information, and poor living conditions. There is also evidence that children living with parents whose educational attainment is low have more health problems than other children, even after other socio-economic factors have been taken into account (Zill, 1996). These problems can begin even before birth because low educational attainment is associated with poor maternal health. Field model domains: individual behavior, social environment, physical environment, prosperity, and well-being. Data sources: decennial census; intercensal data may be available from state or community data systems or estimates.


Ratio of the number of students graduating from high school to the number of students who entered 9th grade three years previously.

Teenagers who drop out of high school may be at increased risk of unwanted pregnancy, sexually transmitted diseases, substance abuse, low-paying jobs without health insurance, and violence. This indicator is a measure of cumulative dropouts from the beginning of the high school period. Adjustments will be needed to account for students who transfer to or from other schools. Field model domains: disease, individual behavior, social environment, physical environment, and prosperity. Data sources: local school districts; data should be collected by individual districts and for all districts combined.


Median household income.

Median household income in the community provides information on family economic resources and the distribution of income in the community. Household income can affect a family's ability to obtain suitable housing, nutrition, or health insurance and may be related to behaviors that affect health. Comparisons over time within a community, among population groups within a community, or with other communities may be helpful in gauging the possible relationship between income and health status or other factors. Field model domains: individual behavior, social environment, physical environment, prosperity, health care, and health and function. Data sources: decennial census; may be available from state surveys.


Proportion of children less than 15 years of age living in families at or below the poverty level.

This indicator is included in the consensus set recommended by the Centers for Disease Control and Prevention (CDC, 1991) for use by all states and communities. It is similar to median household income but focuses specifically on children in low-income households, whose risk for health problems is high and whose ability to address health risks is limited. Many of these children will be enrolled in Medicaid or qualify for other health-related programs such as WIC (Special Supplemental Food Program for Women, Infants, and Children). Field model domains: individual behavior, social environment, physical environment, prosperity, health care, and health and function. Data sources: decennial census; may be available from state or local surveys.


Unemployment rate.

For individuals, unemployment reduces household income, can limit access to health insurance, and can contribute to psychological stress. For a community, an increase in the unemployment rate can increase demands on social services and might signal broader economic problems. The unemployment rate can fluctuate considerably from month to month; therefore rates should be obtained by month or quarter for one to two years to determine the underlying trend. Field model domains: individual behavior, social environment, physical environment, prosperity, health care, and health and function. Data sources: state employment security office.


Number and proportion of single-parent families.

Single-parent families may experience many economic and social stresses that affect the health status of adults and children. Field model domains: individual behavior, social environment, physical environment, prosperity, health care, and well-being. Data sources: decennial census; data on divorce and births to unmarried mothers can be obtained from the state vital records office to monitor changes in family structure.


Number and proportion of persons without health insurance.

Having health insurance can be key for access to health care services. Without insurance, individuals often do not receive timely treatment or preventive care, which can compound adverse health conditions. Field model domains: disease, social environment, health care, health and function, and well-being. Data sources: no uniform community-level data collection tool is available; state assistance may be necessary to obtain data through community surveys. Oversampling in a state-level survey for the Behavioral Risk Factor Surveillance System (BRFSS) might be a source of information on adults; modifications would be required to obtain information on children.


Infant mortality rate by race or ethnicity.

This indicator is included in the consensus set recommended by the CDC (1991) for use by all states and communities. It is widely used as an indicator of child health. Because there are many reasons why infants die, infant mortality reflects the effectiveness of health departments, personal health care providers, outreach services, and preventive services for the mother before and during pregnancy and for the child during the first year of life. The number of deaths will be small in most communities so caution is required in analyzing these data. Usually, data will have to be aggregated for multiple years to produce a stable rate. Field model domains: disease, genetics, individual behavior, social environment, physical environment, health care, and prosperity. Data sources: state or local vital records.


Numbers of deaths or age-adjusted death rates for motor vehicle crashes (ICD-9 codes: E810–E8251), work-related injuries, suicide (E950–E959), homicide (E970–E978), lung cancer (162), breast cancer (174), cardiovascular diseases (390–448), and all causes, by age, race, and gender as appropriate.

This indicator is included in the consensus set recommended by CDC (1991) for use by all states and communities. These leading causes of death provide a basic understanding of the health status of the community. Data should be analyzed by age, race, and gender if possible to target preventive efforts. Although in some communities the numbers of deaths will always be too small to develop a stable rate, it is important to know the number of events. For example, although there may not be a large number of teenage suicides, any number is unacceptable. At the community level, the number of deaths for any specific cause will be small, and data will need to be aggregated for multiple years to produce stable rates. Field model domains: disease, genetics, individual behavior, social environment, physical environment, health care, and prosperity. Data sources: state or local vital records.


Reported incidence of AIDS, measles, tuberculosis, and primary and secondary syphilis, by age, race, and gender as appropriate.

This indicator is included in the consensus set recommended by CDC (1991) for use by all states and communities. Communicable diseases such as these affect the individuals who are infected and also place the entire community at risk. For some conditions, the numbers of cases may be too small to develop stable rates, but establishing the number of persons with the disease is important since nearly all cases are potentially prevent able. Field model domains: disease, genetics, individual behavior, social environment, health care, health and function, well-being, and prosperity. Data sources: state or local disease surveillance systems.


Births to adolescents (ages 10–17) as a proportion of total live births.

This indicator is included in the consensus set recommended by CDC (1991) for use by all states and communities. Births to young women of school age are usually unplanned and often unwanted. The pregnancy can have a negative impact on the health and well-being of the mother, father, grandparents, and child. Lack of economic and social support can manifest in various diseases and health conditions. Field model domains: individual behavior, social environment, well-being, and prosperity. Data sources: state or local vital records.


Number and rate of confirmed abuse and neglect cases among children.

This indicator is included among the priority data needs to augment the consensus indicators recommended by CDC (1991) for use by all states and communities. Children are the most vulnerable population in a community. Most abuse and neglect cases involve young children who cannot defend or choose for themselves; thus, a community response is required. Child abuse and neglect are thought to be underreported, and inconsistencies in reporting and confirmation practices make it difficult to assess changes in incidence (NRC, 1993). Field model domains: disease, individual behavior, social environment, physical environment, health care, and well-being. Data sources: state or local child protection agency.


Proportion of 2-year-old children who have received all age-appropriate vaccines, as recommended by the Advisory Committee on Immunization Practices.

This indicator is included among the priority data needs to augment the consensus indicators recommended by CDC (1991) for use by all states and communities. The immunization rate reflects the effectiveness of the public health system and personal health care providers in delivering immunization services. It also reflects the impact of family decisions, which can be influenced by personal circumstances, economic factors, and factors affecting access to services. The current series of immunizations recommended for completion by 2 years of age is four doses of diphtheria-tetanus-pertussis (DTP) vaccine; three doses of polio vaccine (oral or inactivated); three doses of Haemophilus influenzae type b (Hib) vaccine; three doses of hepatitis B vaccine; one dose of measles-mumps-rubella (MMR) vaccine; and one dose of varicella vaccine (CDC, 1996). Field model domains: individual behavior, social environment, prosperity, and health care. Data sources: retrospective school records surveys; community immunization register; community surveys; health plan records; reviews of patient records. Except where an immunization registry has been established, there is no routine reporting on immunizations.


Proportion of adults aged 65 and older who have ever been immunized for pneumococcal pneumonia; proportion who have been immunized in the past 12 months for influenza.

This indicator is included among the priority data needs to augment the consensus indicators recommended by CDC (1991) for use by all states and communities. The immunization rate reflects the effectiveness of the public health system and personal health care providers, as well as decisions of the elderly or their caretakers. Field model domains: individual behavior, social environment, prosperity, and health care. Data sources: Medicare claims files; health plan records; community surveys (questions have been developed for the BRFSS).


Proportion of the population who smoke by age, race, and gender as appropriate.

This indicator is included among the priority data needs to augment the consensus indicators recommended by CDC (1991) for use by all states and communities. Smoking is the greatest risk factor associated with the leading causes of death. It has been estimated that 19 percent of all deaths are related to smoking (McGinnis and Foege, 1993). It also contributes to morbidity from chronic lung disease and respiratory infections. Smoking adversely affects the health of smokers and also other persons who breathe secondhand smoke. The fetus of a pregnant woman can be adversely affected as well. Estimates of the prevalence of smoking among adolescents (ages 10–14 and 15–19) might serve as a proxy for more direct measures of smoking initiation. Field model domains: disease, individual behavior, social environment, physical environment, prosperity, health care, and health and function. Data sources: community surveys (e.g., oversampling for a state survey for the BRFSS) and school-based surveys (e.g., for the Youth Risk Behavior Surveillance System) for data on adolescents; maternal smoking status is recorded on birth certificates, but the quality of the data needs to be evaluated.


Proportion of the population age 18 and older who are obese.

This indicator is included among the priority data needs to augment the consensus indicators recommended by CDC (1991) for use by all states and communities. Obesity is associated with increased risk for cardiovascular diseases, diabetes, some cancers, and conditions such as arthritis. It also generally reflects a combination of dietary factors and limited physical activity that are themselves associated with increased health risks. It has been estimated that 14 percent of all deaths in the United States are related to diet and activity patterns (McGinnis and Foege, 1993). Obesity can be measured in terms of the body mass index, which can be constructed from weight and height data (kg/m2). Field model domains: individual behavior, genetics, social environment, health care, health and function, and well-being. Data sources: community surveys (e.g., oversampling for a state survey for the BRFSS).


Number and type of U.S. Environmental Protection Agency air quality standards not met.

This indicator is included in the consensus set recommended by CDC (1991) for use by all states and communities. Air quality can have a significant impact on health, particularly for those who have chronic respiratory conditions. Field model domains: disease, social environment, physical environment, and well-being. Data sources: state environmental quality agency; local air quality management agency.


Proportion of assessed rivers, lakes and estuaries that support beneficial uses (e.g., fishing and swimming approved).

This indicator is included among the priority data needs to augment the consensus indicators recommended by CDC (1991) for use by all states and communities. Pollution in a community's rivers, lakes, and estuaries may directly cause disease and also affect the well-being of the community. Field model domains: disease, individual behavior, social environment, physical environment, and well-being. Data sources: state environmental quality agency.


Per capita health care spending for Medicare beneficiaries (the Medicare adjusted average per capita cost [AAPCC]).

Analysis shows considerable differences among communities in health care costs even after controlling for demographic factors (Wennberg, 1996). These analyses also indicate no discernible differences in mortality rates in communities that spend less money on health care. Communities should use this indicator in combination with other information (e.g., AAPCC and morbidity levels over time or across communities) in considering the appropriateness of resource use for health care. Because data do not exist on the total health care costs for most communities, the per capita health care spending for Medicare beneficiaries serves as a proxy for the community's total health care costs. Field model domains: health care and prosperity. Data sources: Health Care Financing Administration.


Proportion of adults reporting that their general health is good to excellent.

This indicator is a good overall indicator of the health status of persons in the community. Field model domains: health and function and well-being. Data sources: community surveys (e.g., oversampling for a state survey for the BRFSS).


During the past 30 days, average number of days for which adults report that their physical or mental health was not good.

This indicator is another approach to measuring the overall health of persons in the community. Field model domains: health and function and well-being. Data sources: community surveys (e.g., oversampling for a state survey for the BRFSS).


Proportion of persons satisfied with the health care system in the community.

Perceptions regarding the health care system can have an influence on perceived health status. This indicator is a broad measure of satisfaction, which could relate to many aspects of the health care system including access, cost, availability, quality, and options in health care. No standard measure of ''satisfaction" has been established, but the committee endorses efforts to do so. Field model domains: social environment, health care, health and function, well-being, and prosperity. Data sources: community survey.


Proportion of persons satisfied with the quality of life in the community.

As proposed by the committee, health is more than just the biological events occurring or not occurring in a person. The ideal of health is a sense of well-being in a person's life. Although quality of life is a difficult concept to measure, this indicator represents an effort to address this state. Standard measures of satisfaction and quality of life would have to be developed to use this indicator. Field model domains: individual behavior, social environment, physical environment, prosperity, health care, health and function, and well-being. Data sources: community survey; questions related to quality of life have been developed for the BRFSS.


  • CDC (Centers for Disease Control and Prevention). 1991. Consensus Set of Health Status Indicators for the General Assessment of Community Health Status—United States. Morbidity and Mortality Weekly Report 40:449–451. [PubMed: 2056991]
  • CDC. 1996. Immunization Schedule—United States, January–June 1996. Morbidity and Mortality Weekly Report 44:940–943. [PubMed: 8531913]
  • McGinnis, J.M., and Foege, W.H. 1993. Actual Causes of Death in the United States. Journal of the American Medical Association 270:2207–2211. [PubMed: 8411605]
  • NRC (National Research Council). 1993. Understanding Child Abuse and Neglect. Washington, D.C.: National Academy Press.
  • USDHHS. 1995. International Classification of Diseases, Ninth Revision, Clinical Modification . 5th ed. DHHS Pub. No. (PHS) 95-1260. Washington, D.C.: National Center for Health Statistics and Health Care Financing Administration.
  • Wennberg, J., editor. , ed. 1996. The Dartmouth Atlas of Health Care . Chicago: American Hospital Press.
  • Zill, N. 1996. Parental Schooling and Children's Health. Public Health Reports 111:34–43. [PMC free article: PMC1381739] [PubMed: 8610189]



The framework of structure, process, and outcome measures was originally developed for quality assurance in health care (Donabedian, 1980, 1982, 1985) but has proved useful in a variety of contexts. Structure applies to capacity to perform (e.g., whether smoking cessation counseling is available to pregnant women). Process applies to activities that are being performed (e.g., numbers of pregnant women receiving smoking cessation counseling). Outcome applies to results of those activities (e.g., proportion of counseled women who stop smoking or, more significantly, the rate of low-weight births among counseled women).


The committee's indicator selection criteria are similar to those specified by other groups for related purposes. The Sustainable Development Indicators project in Hamilton-Wentworth, Ontario, Canada (Regional Municipality of Hamilton-Wentworth, 1996), listed the criteria of measurability, cost and ease of collection, credibility and validity, balance, and potential for effecting change.

The National Committee for Quality Assurance (NCQA, 1996) identified the following desired attributes of measures to be submitted for consideration for version 3.0 of the Health Plan Employer Data and Information Set (HEDIS): relevance (meaningful to users, health importance, financial importance, cost-effectiveness, strategically important, controllability, variance between plans, potential for improvement), scientific validity (reproducible, valid, accurate, risk adjustable, comparability of data sources), and feasibility (precisely specified, reasonable cost, confidential, logistically feasible).

Criteria established by the Scientific Advisory Committee of the Medical Outcomes Trust (Perrin, 1995) for outcomes assessment instruments to be included in the Trust's collection are a conceptual and measurement model, reliability, validity, responsiveness (ability to detect change), interpretability, burden, alternative forms, and cultural and language adaptations.


Diagnostic codes assigned under the International Classification of Diseases, 9th Revision (USDHHS, 1995).

Copyright 1997 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK233011


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.3M)

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...