INTRODUCTION

National immunization surveys in 1997 demonstrated that only 76% (±0.7%) of children 19–35 months of age had received the full complement of immunizations appropriate to their age (4 doses of diphtheria-tetanus-pertussis vaccine, 3 doses of oral poliovirus vaccine, 3 doses of Haemophilus influenzae type b vaccine, and 1 dose of measles-mumps-rubella vaccine), with a range among states of 67% (Idaho) to 88% (Connecticut) and a range among urban areas of 63% (Newark, NJ) to 86% [King County (Seattle), Washington]. However, these data do not tell the full story. Rates vary greatly within metropolitan areas, with immunization rates in inner-city areas lagging behind suburban rates. In inner-city New York, Chicago, and Los Angeles, immunization rates are 50%–60%, ≤30% lower than corresponding suburban or state immunization levels (41, 70, 77).

The failure to raise immunization levels in poor, underserved populations is primarily caused by inequitable and insufficient distribution of essential child health services to meet the demand (19, 27, 29, 76, 77). Exacerbating these structural problems is the lack of timely and accurate child-specific information on immunizations for providers and parents. Studies have shown that providers frequently miss opportunities to vaccinate children in their offices because they do not have access to accurate immunization information (67). Both provider records and parent-held records for immunizations may be inaccurate (32, 50). Parents are often confused by the complex immunization schedule and commonly believe their children's immunizations are up-do-date when they are not (25, 78). Immunization registries are a new tool in health care that can be used to address these and other barriers to effective immunization delivery. Moreover, immunization registries have the potential to track and improve delivery of a broad range of important health and social services for children and families.

An immunization registry is a computerized database that gathers immunization information on all children (with preschool children commonly a high priority) in a population defined by a specific geographic area, health maintenance organization (HMO) enrollment, etc. It should bring together in one database immunizations and demographic information from all sources including private and public medical providers, the parent-held immunization card, schools, hospitals, and other agencies that give immunizations. A fully functioning immunization registry tracks all children in the population from birth through at least 2 years of age and preferably through childhood. It determines immunizations needed by individual children and can respond to provider inquiries for information in a time frame that is useful for service delivery. Moreover, a fully functional registry produces population immunization reports by a number of important parameters such as assigned provider, health plan, geographic area, child age, and other factors (17, 44). The provision of immunization reports to clinics has proven to be a powerful incentive to improve the quality of immunization practices and has resulted in higher immunization rates over time.

The development of community-based registries first began in Delaware in the early 1970s (71). Somewhat later several large HMOs started to develop immunization registries before the current community-oriented approach expanded. Group Health of Puget Sound, WA, had an operational registry covering its 350,000 enrollees in the late 1980s (53). In the early 1990s, through a collaboration with the Centers for Disease Control and Prevention (CDC), several other large HMOs on the West Coast established immunization registries and vaccine adverse-events reporting systems (11). During the same period, the Robert Wood Johnson Foundation (RWJF), along with several other foundations operating under the program name All Kids Count (AKC), funded the development of the initial community-based immunization registries in 24 communities (71).

In 1998, immunization registries were in various stages of planning or development in all states as well as many communities and managed care organizations. Thirty-four states and 255 localities have registries that are functional at some level, meaning that health clinics are transmitting immunization information to a central registry (12). Population-based registries that are fully functional (i.e. track ≥95% of children under 2 years of age in the population) have been used to raise immunization levels through the use of child-specific reminders to providers at the point of service and to parents of children behind in immunizations (42, 68, 81). Immunization registries implemented by large HMOs have used similar approaches to raise immunization levels (5, 43, 53, 69). Registries can provide the means to build predictive models from historical data that prospectively identify children at high risk for underimmunization (77). Registries also offer great promise for supporting public health-oriented outreach efforts (i.e. identifying children at risk for anemia, high lead levels, etc) through collaborations with Women, Infants, and Children (WIC) programs and other community-based agencies that serve poor, inner-city areas (28, 36, 76).

Nascent immunization registries represent innovative technologic solutions to the challenge of monitoring health problems and health care access on a population basis. This is a fundamental activity of public health agencies and one increasingly shared by large HMOs (24). By developing at an early stage of the information systems explosion in health care, immunization registry advocates and developers have had to solve many vexing technological and sociopolitical problems related to sharing patient information in our largely private, disjointed health care system. The study of the development of immunization registries across the United States also provides a very timely and important case study in how public health agencies will use the rapidly developing health information infrastructure to fulfill their primary responsibilities of health assessment, health assurance, and health policy development.

KEY ATTRIBUTES OF A REGISTRY

The CDC and the National Vaccine Advisory Committee (NVAC) (44) have published policy statements describing 12 key attributes of immunization registries. These 12 key attributes are summarized in Table 1. For selected attributes, we describe how they are being implemented by developing registries.

TABLE 1

Twelve key attributes of an immunization registry

Consolidate All Immunization Records from Multiple Providers

Data quality is central to the effectiveness and credibility of a registry. However, the quality of data in a registry can be undermined by errors in data recording (missing data or mistakes in data entry), data transcription, and transmission; changes in child or parent name, address, or other identifying information; or delays between administering immunizations and submission of the data to the registry (40). A study of data quality in a clinic-based electronic registry demonstrated a 10% rate of data transcription errors, and 7% of data were missing from the registry. The reports produced by the registry overestimated the number of underimmunized children by 100% (73). However, because any one source of immunization information (i.e. private provider records, public health clinic records, school records, or parent-held cards) is seldom completely accurate, a registry that brings together and reconciles information from multiple sources may be the best alternative. A study in Austin, TX, comparing the accuracy of immunization data in the community immunization registry to those in physician office charts found that the registry was consistently more accurate (55).

A number of approaches have been developed to validate registry immunization data and identify incomplete or duplicate child records (14, 46). Many immunization registries use sophisticated probabilistic matching algorithms coupled with a powerful name and address parsing engine (breaking name and address into various components for comparison) to match records with extremely high confidence. These programs are adapted from those used in the banking industry that can rapidly and with extreme accuracy locate a single person among 400 million customer records. Another method of assuring data quality is to compare registry information with parent-held cards or clinic chart immunization records for a sample of children in the registry (46). This quality assurance methodology is being performed by officials at over half of the state registries. Mississippi and Arizona State immunization registries are well along in development of validation procedures that cover the records of most of the children in each state. They have invested significant resources to conduct regular audits of all public and many private offices or clinics. They report the data errors and missing data back to the clinics and then work with the clinic staff to clean up the data. Like feeding back immunization rates to clinics, these reports are powerful incentives for the clinics to improve their performance in data collection and data entry. Thus, although registries have powerful tools to check for redundant records and reconcile data from multiple sources, the most important step in data quality assurance is for providers to ensure the accuracy of data at the point of service (46).

Electronically Store and Retrieve All Core Data Elements Approved by the Centers for Disease Control and Prevention and National Vaccine Advisory Committee

National recommendations on core data elements in an immunization registry are established by CDC and the NVAC. The purpose of the recommendations is to assure that registries can uniquely identify children and to promote consistency and appropriate levels of detail in the data so as to allow construction of a useful immunization record (60). These elements are patient name (first, middle, last); patient birth date; patient gender; patient birth state/country; mother's name (first, middle, last, maiden); vaccine type; vaccine manufacturer; vaccination date; and vaccine lot number. Through ongoing surveys, the CDC has found that, despite the wide variation in state laws and registry design, the vast majority of registry projects use all nine core data elements (33).

Link Immunization Registries to Birth Certificate Registries

In the CDC registry survey, 47% of state and other geographically based immunization registries currently link electronically to birth registries. Another 41% plan to do so within 12 months (33). Electronic linkage to the birth record enables tracking of the receipt of immunizations in the first few months of life when many immunizations are due. Not all hospitals, communities, or states have automated birth certificate data, requiring registries to obtain and input the data manually. This delays tracking of newborns several months, greatly compromising the value of linkage to birth certificate data. A significant problem in the use of birth certificate information is that newborns are frequently assigned the mother's last name on the initial copy of the birth certificate, only to be changed to the father's last name when the mother finalizes the birth certificate months later. These name changes impair the ability of registries to uniquely identify children. The assignment of permanent social security numbers at birth would greatly reduce this problem.

Assess Immunization Status for Individual Children and Groups of Children in the Target Population

Almost all state, county, and city registries produce reports on children who are up-to-date or behind in immunizations (33). The Arizona registry produces immunization status reports for children by provider, medical group, and health plan, which assists them to comply with the mandatory reporting law (1, 13). The state registries in Washington, Michigan, and South Carolina produce Health Employer Data Information Set (HEDIS) immunization reports for managed care organizations that provide the registry with enrollee name and identifying information (7, 23, 64). Registries in Arizona, New York State, and San Antonio, TX, have geographic information systems (GIS) that can report immunization rates by small geographic areas, such as blocks or census tracks, which helps target outreach to underimmunized populations.

Another form of outreach to underimmunized populations is sending reminder or recall notices. The specific approach taken to recall or remind families varies widely among registries. Some registries are organized to send out recall notices centrally by letter or by automated phone message. Registries in Rhode Island, Washington, and Michigan are organized to proactively send out standard immunization reminder notices at birth, 2, 4, 6, and 12 months of age to all children born in the geographic area. However, in many communities providers prefer to notify the patients directly, and they either use their own immunization reports or request reports from the registry. The Savannah, GA, registry, one of the few public registries that has gained the participation of nearly all private and public providers, adopted a novel approach. Each provider records a reminder and recall message to be accessed by the registry. Officials of the registry use the provider's voice to send phone reminders and recall notices to the provider's patients. This approach has motivated the providers to promptly submit immunization data to the registry to ensure that their patients receive accurate messages.

Child and Family Confidentiality

The immunization record includes sensitive identifying information, such as address, parental employment status, parental marital status, and eligibility for public programs, an indirect measure of family income. It may also include sensitive medical information, including the immunocompromised status of the child or other household member. Because of this, developers of immunization registries have, from the inception of these registries, had to demonstrate that the privacy and confidentiality of children and families can be maintained (26). Immunization registries have electronic and personnel safeguards to prevent inappropriate disclosure of information or inappropriate access by unauthorized persons or organizations. These safeguards include assigning levels of sensitivity to each data element and defining protection efforts for each level; defining all purposes for which access to the data may be granted; establishing data access policies for all potential users (e.g. providers, schools, government agencies, etc); delineating all verification protocols for each type of user; and creating audit trails for all transactions. Registry officials regularly review and update confidentiality and security procedures and offer ongoing education in security procedures to staff and users. By state law, officials of all immunization registries must disclose to parents all potential uses of the data and the safeguards in place to protect client confidentiality (40, 45).

In 1998, there was as yet no comprehensive federal statute protecting the confidentiality of health information. As a result, immunization registries have developed under a diverse set of state laws (26). In a recent survey, 18 states were found to have legislation specifically authorizing rules for data sharing by immunization registries, and 9 had existing regulations governing general health information sharing, which were applied to immunization information (37). In 33 states, consent to collect, record, and share immunization-related information was implied, not requiring the parent to give formal written consent for each release of immunization information (which is the standard used for most health information). In 24 of those 33 states, however, parents were given the right to refuse or limit their child's participation in the registry. In 12 states, either by state law or by health department regulations, written parental consent is required before the child can be included in the registry (37).

In the near future, federal standards for health care information security will be issued as required by the 1996 Health Insurance Portability and Accountability Act. Under this legislation, the Department of Health and Human Services is to develop security standards for the storage and sharing of patient records and other sensitive healthcare information (including child immunizations, patient address, and other identifying information). The practice of sharing health care information electronically has grown dramatically in recent years as health care providers communicate with partners in integrated delivery systems and as the World Wide Web has become ubiquitously available (59). Standards for encryption of public and private data and other security technologies, adapted to health care from the banking industry, have been developed by the leading information technology professional associations. The Department of Health and Human Services is likely to adopt these standards. Once the regulations are issued, immunization registries and all health organizations will have two years to bring their information systems and data communications security into compliance (58). Health departments and other health care organizations, by having already invested the effort to design comprehensive immunization registry data security programs, will be in a good position to meet the requirements of the new Department of Health and Human Services/Health Insurance Portability and Accountability Act security regulations.

KEY ISSUES IN REGISTRY DEVELOPMENT AND IMPLEMENTATION

Registry Planning Design

Although immunization registries were initially implemented in the public health departments, public-private partnerships have been a prerequisite for successful implementation (39). Registry projects across the nation have collaborated with community and health care stakeholders in registry planning and implementation (56). Most registries have established advisory groups that can

Arizona formed a statewide coalition, entitled The Arizona Partnership for Infant Immunization, composed of private provider groups, managed care plans, state agencies, professional organizations, pharmaceutical companies and other businesses, the faith community, the media, and many other groups (35). The coalition formed five subcommittees in the areas of advocacy and policy, community awareness, provider awareness, survey and assessment, and strategic planning. In 1996 The Arizona Partnership for Infant Immunization joined with managed care organizations and the medical professional societies to successfully support legislation that required all communities in the state to have immunization registries. The state legislation also mandated reporting of childhood immunization to local registries by all health care professionals. This legislation resulted in a dramatic growth in provider participation in the state and local immunization registries.

Community Needs Assessment

The structure of immunization delivery systems varies widely from state to state and community to community. In some states, such as Mississippi and Nevada, the public sector delivers 75%–85% of childhood immunizations, whereas in other states, such as Rhode Island and California, the public sector delivers ≤15% of child immunizations. To better understand community issues, registry projects in the development phase almost always include community needs assessments (39). The needs assessment process commonly starts with a public and private provider survey to assess current immunization delivery practices, computer and telecommunications capacities, record-keeping practices, and provider readiness for electronic data exchange (52, 60). The survey also assesses providers' interest in a range of potential services that could be offered by the registry. Planners and officials of many registry projects have made site visits to provider offices to better understand provider reporting needs, how to increase the registry's user friendliness and reduce data entry burden on office staff, and how to minimize the negative impact on patient flow.

Provider Participation in Registries

Because most publicly sponsored immunization registries were initiated by public health departments, public health department clinics were the first delivery sites to be connected (20, 56). In the early 1990s, changes in immunization financing shifted much of immunization delivery from public clinics to private providers. For example, although the overall level of vaccinations delivered in Los Angeles County has remained level, the number of child vaccinations delivered in Los Angeles Department of Public Health clinics dropped from >1 million doses per year in the late 1980s to just over 300,000 doses in 1997 (G Feldman, personal communication). Given the growing role of private providers in vaccination delivery, officials of public registries have increased their efforts to gain the participation of private providers. However, despite the increased effort, engaging the private sector has been a complex and slow process. A 1997 CDC survey demonstrated that, in the entire United States, only two registries had 100% of community private providers participating and only four registries had ≥50% of private providers participating (33).

Surveys demonstrate that providers are willing to participate in a local public registry if the following conditions are met: (a) data in the registry must be sufficiently accurate and reliable to be used in immunization delivery decisions (i.e. the data help providers provide immunizations to patients at the point of service); (b) the costs of participation in the registry must be offset by cost savings, such as reduced vaccine wastage or by producing practice-specific immunization coverage reports (10, 64); (c) patient- and practice-specific information in the registry must be kept confidential. States have used incentives and mandates for private providers to promote their participation in immunization registries. South Carolina and Rhode Island give private providers new computers, modems, and other free tools to help them link to the state's immunization registry. North Carolina and other states make participation in the immunization registry a prerequisite for providers to receive free vaccines through the public vaccine purchase program (4, 18). Ten states have passed laws mandating that private providers report immunization information to state and local registries (37).

The dramatic growth of Medicaid managed care as well as private managed care has helped registries gain access to immunization data from private providers. Medicaid managed care plans must report immunization status of their enrolled children <2 years of age to satisfy HEDIS reporting requirements. HEDIS reporting is voluntary for private HMOs; however, there is a high level of participation, with 292 private managed care plans submitting and publicly reporting HEDIS data in 1997 (48). Cooperative agreements have been developed between registries and managed care plans in which reports are exchanged on their enrolled populations for collaboration and submission of child immunization data to the registry. In Oregon, Maine, and Connecticut, the Medicaid managed care plans send immunization data of enrolled children, many of whom see private providers, to state registries, and they also contract with the state registry for immunization reports by provider and by health plan.

Technical Design Considerations

There are two basic hardware configurations for immunization registries. Older projects in large health care organizations have generally used a host/mainframe registry design, linking dumb terminals or gateways via direct lines or a network to a mainframe computer. This registry design is highly controlled, stable and fault tolerant, relatively simple to design and bring on-line, and potentially very responsive. However, organizations with centralized computing structures have difficulty keeping pace with the changing immunization recommendations, as well as developing interfaces to local immunization registries.

The most popular design for newer immunization registries is the client-server architecture. The client-server design uses a high-capacity server to house the central database, which is connected to personal computers via an existing network infrastructure and client software or through the World Wide Web with a Web browser. The personal computers connecting to the main database have registry software resident which allows the provider or group to maintain a local registry, produce reports, do reminder recall, etc. A significant advantage of the client server is that it gives providers the ability to manage the immunizations of their own patients even while the community registry is in development. Furthermore, the client server design can easily be scaled up to include additional providers, an important characteristic as registries grow.

Health providers use a number of communication methods to send or receive community-based registry information. The most common method is a simple fax or paper transfer. A number of registries have developed private interactive voice response and fax-back technologies, which automatically process a voice phone request from a provider and fax the response to the provider within 5 minutes. Officials of the Michigan public registry have found that this technology is sufficiently rapid to support immunization decisions at the point of service (J Weihl, personal communication). The second most common method is for providers' offices to directly dial into the registry via modem. This method has the advantage of using widely available commercial communication software; however, it has the disadvantage of being limited to the relatively low speed of the modem. Larger physician groups or HMOs require higher-speed transfers and more secure communication, and they typically install direct telecommunication linkages to the registry. The next most common communication method used to send or receive community-based registry information has developed from the dramatic growth of the Internet and the availability of new encryption and other Internet data security methods, permitting both large and small providers to increasingly use Web browsers to exchange information with immunization registries.

Initial Financial Support and Long-Term Sustainability

Before registries become fully functional and can be used to raise revenues through services to providers, the substantial startup costs associated with software/hardware design or purchase and outreach to and enrollment of providers must be covered. Cost analyses conducted on 16 of the original AKC projects found that annual registry costs during the development phase averaged $3.65 per preschool child in the target area (38). Although the first grants by the RWJF were relatively small ($675,000 over a 5-year period), most projects greatly augmented these funds with in-kind local health department personnel and federal immunization funds. In 1998 the RWJF allocated additional funds to 16 successfully developing registry projects with the goal that they become fully operational by Jan 1, 2000 (enrolling and tracking 95% of all children <2 years of age in the catchment area). In 1996, the CDC's National Immunization Program allocated $42.5 million to the 65 state, city, and territory granting agencies. In 1997 the National Immunization Program/CDC allocated $35 million for immunization registry development (44).

States and communities have creatively patched together funding for the initial development and early implementation of immunization registries. The Philadelphia Kids Immunization Database System received funding from CDC, the RWJF, and from Mercy Health Plan, the largest Medicaid managed care organization in the city. The Oregon State Health Department immunization registry partnered with the state Medicaid program and a private, nonprofit corporation representing Oregon's largest health plans to help underwrite the costs of registry development (15). Michigan's registry project augmented state and federal dollars with some of the revenues from the state's tobacco tax.

The costs and challenges to achieve full registry functionality and ongoing sustainability are significant. To progress beyond the development phase and become fully functional (tracking 95% of children <2 years old in the catchment area), registries must enroll all children from birth in the catchment area, perform extensive data quality monitoring, and produce individual patient queries and population coverage reports. These functions are costly to perform and have quickly outpaced the resources available from foundation grants and public appropriations. One solution to the problem of limited public and foundation funding has been to form strategic partnerships with information management businesses that can supply additional resources and additional technical expertise. For example, the Washington State registry contracted with a for-profit information management business that invested capital to upgrade the registry's technology. The partner acts as a broker or intermediary between the registry and providers, increasing outreach to providers (i.e. marketing) and offering a broad range of services that providers can purchase on a fee-for-service basis (2). A number of registries have increased revenues by directly charging providers and/or managed care organizations for services such as immunization coverage reports, patient-level requests for information, or reports of children overdue for immunizations. The Seattle registry contracted with managed care organizations at a capitated rate of $4.00 per child per year, which covers the average per-child costs of basic registry functions. Maine and New Hampshire, which operate a joint registry, developed a contract with their state Medicaid programs to pay for immunization reports and queries on enrolled children (54). The Connecticut Medicaid managed care organizations contract with the state immunization registry on a capitated per-member, per-month basis for predetermined registry services (80). At the national level, there is consideration for imposing a surcharge on new vaccines and using the funds to underwrite core registry functions not covered by provider revenues (22, 72).

CURRENT STATUS OF IMMUNIZATION REGISTRIES

Given their recent development, there are a limited number of published studies on immunization registry function and outcomes. In 1997 the CDC's National Immunization Program conducted a phone survey of states, major urban areas, and territories. The study was the first attempt to collect detailed and comparable information on the status of registry projects in the United States (33). The survey assessed registry projects across the country for (a) the size of the target and enrolled populations, (b) the number of providers linked to the registry, (c) the technical design (e.g. hardware and software systems used), (d) the functions supported by the registry, including provider reminder and patient reminder/recall, (e) the legal environment and local or state legislation that supports or inhibits registry function, and (f) the current funding sources and long-term business plans. Differences in community and state public and private health systems, as well as political and legal environments, have produced significant variations in overall registry design and function. However, there are also many common issues faced in the design and implementation of registries that warrant documentation. The results from the CDC survey are presented throughout this paper and are available on CDC's web page (http://www.cdc.gov/nip/registry).

The survey identified 255 community registries based in city and county health departments and large managed care organizations, in various stages of planning or implementation (Figure 1). Seven public health departments were in the preplanning, needs assessment phase, and 100 health departments were at some level in the planning stage. Fourteen local public health departments were pilot testing their registries, 105 were at some level in the implementation phase, and only 29 registries had all public clinics on line (44). At the CDC grantee level (50 states, 6 major cities, and 8 territories), 24 grantees had immunization registries that received and sent data from between 75% and 100% of public clinics, and an additional 24 grantees had registries that communicated with at least one public clinic (Figure 2).

However, in all but a few areas, little progress had been made in recruiting private providers or in establishing registry links between immunization partners such as managed care organizations and WIC (Figure 3) (33). From 1992 to 1997, AKC registry projects were evaluated on their progress in the implementation of key registry attributes described above. Although no project was fully operational as of 1996 (∼4 years after their start), 75% of projects had completed their system design and had established community support. Only 25% of projects were able to link the central immunization database with private providers, and only two projects were able to produce up-to-date immunization reports for the entire population (8). The slow and difficult process experienced by these pioneer registry projects demonstrates the complex, politically difficult, and costly nature of developing a community-based immunization registry.

The AKC program has developed qualitative and quantitative indicators of the developmental progress of registries. These indicators include (a) time for entry of birth records into the registry, (b) time for entry of immunizations administered at provider offices into the registry, (c) the proportion of public and private providers participating in the registry, (d) the proportion of the target population tracked in the registry, and (e) the completeness and accuracy of immunizations recorded in the registry. The AKC program will survey their 16 registry projects on these indicators and describe their progress. In addition, a number of other CDC-funded registry studies are underway to estimate the costs of operating a registry and the cost effectiveness of key registry functions (e.g. patient and provider reminder/recall), to identify best methods to promote private provider participation and to evaluate methods to uniquely identify individuals within a registry (57).

THE FUTURE ROLE OF REGISTRIES: MONITORING THE HEALTH OF CHILDREN, FAMILIES, AND COMMUNITIES

After public efforts to reform the health care system failed in 1993–1994, market pressures produced a myriad of changes in the health care system that we summarize in the following three major trends. First, a large proportion of publicly and privately insured populations shifted from fee-for-service reimbursement plans into capitated managed care plans. In 1996 the number of Medicaid beneficiaries enrolled in capitated managed care was 13.3 million, a rise of 600% in 6 years (24). In the private sector the growth of managed care plans has been equally rapid, with 77 million Americans enrolled in HMOs in 1997. Second, a wave of mergers, acquisitions, and other partnerships between provider groups, hospitals, HMOs, and other health organizations created many large, integrated-delivery networks (IDNs). In 1998 the consolidation of the health systems in Portland, Minneapolis, San Diego, and other areas resulted in individual markets with only three to four major IDNs (24). Third, many hospitals and health plans converted from not-for-profit status to for-profit companies that were responsible to stockholders and investors for returns on their investment (30).

The role of public health in the newly consolidated, market-driven, capitated health care system is not well defined and is being hotly debated (3, 21, 24). Furthermore, the role of large, capitated delivery networks in performing the traditional public health activities of disease prevention and health promotion activities is unclear (21). Although HMOs were founded on principles of health promotion and disease prevention, the increasingly competitive market has created significant pressure on HMOs/IDNs to focus on short-term cost reduction activities such as reducing utilization of expensive services, reducing hospital per diem costs, and shortening hospital lengths of stay. Many preventive services accrue benefits on a longer timeline, often resulting in social benefits outside the health care system rather than cost savings to the health plan. For example, the costs of child lead screening or early childhood developmental screening may be born by the child's HMO, but the benefits accrue several years in the future to the educational system in the form of reduced learning problems and school failure (21).

However, these same trends have also prompted private HMOs and IDNs to invest heavily in new information systems, which may lay the foundation for new and innovative collaborations with public health departments and immunization and health registries. As hospitals and physician groups and other groups merged to form IDNs, many have found that the lack of integrated clinical information systems is a significant impediment to improving the quality and cost effectiveness of medical care (61). Moreover, the public backlash against cost-cutting measures instituted by some HMOs has increased the demand for objective clinical information on the quality of care provided in HMOs. In response to these pressures, 292 health plans voluntarily reported publicly their HEDIS data in 1997, which includes immunization rates for 2-year-old children (48). Therefore, the dual pressures of cost control and public demand for quality-of-care information have pushed HMOs and IDNs to invest heavily in information systems, adding new clinical functionality to the older administrative and financial health information systems.

New health information systems can collect and track important preventive health and health promotion data, such as immunizations, for large populations within HMOs and IDNs. The data collected by IDNs in a major metropolitan area, if merged by the local health department, would be a major step toward the development of a population-based health registry. Immunization registries in mature managed-care markets, such as Portland and Minneapolis, are aggregating data from HMOs and IDNs and providing reports back to the reporting organizations. The data can also be used by health departments to fulfill their mission to monitor the health of populations and hold HMOs and IDNs accountable for delivering health prevention and promotion services. This effort will be greatly aided by the 1996 federal Health Insurance Portability and Accountability Act, which mandates the standardization of health data and the assignment of universal unique identifiers for all US residents (50).

There are potential benefits for HMOs and IDNs in collaborating with health departments on health data. Publicly based immunization registries can aid HMOs/IDNs by supplying them information on patients that recently joined the network as well as aid them in producing HEDIS reports. Moreover, the registry can be used to identify individuals in need of public health services that are supportive of other aspects of clinical care but are not performed by personal health care systems, such as contact tracing for sexually transmitted diseases or lead abatement for housing of children with high internal lead levels (3).

Potential Impact on the Child Health Delivery System

The restructuring of the personal health care system is occurring on top of an already disjointed and fragmented public and private child health care system. The child health care system is made up of many categorical health programs, each with its own eligibility, administrative, and funding criteria (29). Furthermore, child health programs are segregated into personal health services (e.g. primary care services), public preventive health services (e.g. lead screening and child abuse prevention) and other publicly funded health-related services (e.g. WIC nutrition services, early intervention and home-visiting services, school health and education services). Even the public sector child health programs, despite serving similar populations of low-income children and families, rarely share patient information and seldom coordinate their delivery of services.

Immunization registries can form the information infrastructure to support tracking and sharing of patient-based information among the three sectors within the child health system. Many immunization registries have laid the groundwork for these collaborations by including in their development process a broad range of public and private child health providers. Immunization registries have frequently overcome complex interorganizational barriers and developed information exchange protocols and confidentiality safeguards that span these barriers. An example of such an effort is KIDS NET, an information-sharing project of the state health department in Rhode Island. KIDS NET connects child-specific information from vital records, prenatal-care services, universal newborn screening, immunizations, lead screening, infant metabolic and hearing screening, and early intervention services (31). KIDS NET was designed to track and report the above information under strict confidentiality protocols to the appropriate categorical public agency as well as to hospitals, physicians, emergency rooms, schools, and other agencies. The data in KIDS NET are also used to identify high-risk families for outreach and preventive services such as home visitation and nutrition support. It is an example of a registry project with the potential to increase the efficiency and effectiveness of health and preventive services for medically and socially at-risk children and families (31).

Increasing Services to Underserved Communities

Children of low-income families in the inner city or rural areas are at risk for many problems that compromise their health and development. Epidemiologic studies have shown that low-income children have higher rates of underimmunization, lead exposure, anemia, child abuse and neglect, growth and development problems, and chronic disabling conditions such as asthma, learning disabilities, and chronic conditions associated with premature birth (6, 74). Many studies have documented that the access of low-income children to care is not commensurate with their increased need for personal health, preventive, social, and educational services (51, 65, 79). Moreover, studies have documented the substantial and long-lasting benefits to children, families, and society of early preventive intervention for high-risk populations (16, 49). Despite the increased needs of high-risk children and the proven benefit that services bring, there are only a few localities that systematically identify high-risk children and track the delivery of needed services.

Immunization registries are the first step in identifying and tracking young children over time and are an extremely powerful tool to aide in the identification of specific child health needs or health risks. Broader child health registries could track and promote the timely delivery of developmental-screening and early-intervention services, lead screening and treatment, and other services needed by high-risk children and families. WIC and public health agencies in Chicago and Milwaukee have collaborated to create child health registries to track and promote service to underserved populations. The program has resulted in dramatic increases in rates of immunization and screening for lead, tuberculosis, and anemia among inner-city children (31, 54). Moreover, child health registries could measure and track child outcomes such as functional status, child abuse reports, and school performance. Registries that track a range of outcomes and risk factors can be used to develop predictive models to identify children at risk for adverse child health outcomes and proactively target them for specialized preventive services (53). Officials of the Seattle immunization registry are collaborating with other health and social agencies to use birth registry and other data to predict the likelihood of adverse child outcomes such as child abuse or infant mortality (1). The data are provided to health and social agencies so that families identified by this project are provided with preventive support services.

Many of the issues addressed by developing immunization registries lay the groundwork for the future development of more comprehensive, community-based public health information systems. Sharing of health information with multiple and disparate health organizations is essential if health departments are to monitor the population's health in the future managed-care environment. Moreover, by tracking and reporting on patient outcomes such as immunization rates, users of health registries can assess the quality and effectiveness of health preventive and promotion services delivered by managed-care organizations and IDNs. Although there are many technical, political, and financial obstacles to overcome, immunization registries can be viewed as one of the initial steps in building a public health information infrastructure that will enable public health agencies to fulfill their mission to assess the public's health, assure access to needed health services, and develop policies that protect individual and community health.