The cost of doing business: cost structure of electronic immunization registries

Health Services Research, Oct, 2002 by John M. Fontanesi, Don S. Flesher, Jr., Michelle De Guire, Allan Lieberthal, Kathy Holcomb

Success in immunizing the pediatric population has progressed to the point that disease burden is essentially zero for many of the childhood vaccine preventable diseases; however, reaching this level has required substantial resources in the form of time, personnel, and financing, raising concern about our ability to maintain this degree of disease protection (Centers for Disease Control and Prevention 1998, 2000a; Herrera et al. 2000; Wood et al. 1999; Zimmerman and Burns 2000). These concerns have been voiced by the National Vaccine Advisory Committee (NVAC), Institute of Medicine (IOM), and the Centers for Disease Control and Prevention (CDC). All have identified electronic immunization registries (defined as confidential, computerized information systems that contain the immunization history and status of patients) as a critical component in the long-term strategy to maintain these historically high levels of coverage rates (National Vaccine Advisory Committee 1999; Shefer et al. 1999).

Although registries have primarily focused on immunization of children aged 0-2, the potential to serve a larger population has not been lost on developers. Indeed, some are using registries to monitor immunizations for all age groups and for other public health data such as tuberculin skin tests. A registry, in combination with a geographic information system (GIS) (Clarke, McLafferty, and Tempalski 1996), would permit identification of population-based pockets of need, could be used to guide public health policy, and could serve as a core for development of population-based electronic medical records. The use of a registry to perform epidemiological analysis takes on great importance in light of the events of September 11, 2001, and the potential need to monitor vaccination for agents of bioterrorism, such as anthrax and smallpox.

The belief is that registries should be able to generate an individual's unified immunization record from multiple providers, identify when a child is eligible for immunization and when they may be post-due, create population-level coverage rates, as well as provide reports to individual providers about their clientele's coverage rates in a far less costly and more timely manner than any present system (Linkins and Feikema 1998). To be successful, however, registries must be widely available and easy to use, yet capable of protecting individual privacy. While some of these factors have been investigated, it is unclear what the cost of meeting these goals will be, who incurs the costs and who may benefit. As important is the need to investigate whether the defined policy objectives and the proposed methods for meeting these objectives (i.e., registries) will correspond.

The importance of understanding the capitalization requirements was clearly stated by NVAC: "The barriers to creating a national system of state-based registries are mainly political and financial rather than technical" (National Vaccine Advisory Committee 1999, p. 24). The NVAC further stated that the "Centers for Disease Control and Prevention (CDC) should pursue immediately further study to completely characterize start-up and maintenance costs of registries and compare these to costs of alternative systems" (National Vaccine Advisory Committee 1999, p. 8). This study attempts to meet this information need.

We are not the first to try. As of September 1997, more than three hundred registries were in development, supported by at least $142 million in 317d Federal categorical immunization grant funds and more than $200 million in other public, private, and foundation funds (Centers for Disease Control and Prevention 2000b; Wood et al. 1999). Early research using data from some of these registries found costs ranging from as low as $0.65 to as high as $217 per child per year (Rask et al. 2000a; Slifkin, Freeman, and Biddle 1999). Other authors' findings fell erratically within this range (Home, Saarlas, and Hinman 2000; Urquhart 1999). In examining each of these studies it was evident that different components, different time periods, and different processes were evaluated. Consequently, comparison of results between studies and generalization of findings to a national level was unrealistic. What was needed was a method for converting the collective experiences in such a way as to permit "comparing apples with appl es."

Yet medical informatics is not the first industry to be confronted with the need to anticipate the cost of development and deployment of an application. Such diverse industries as banking, manufacturing, shipping, and retailing have been confronted with the need to predict costs, anticipate benefits, and develop a realistic capitalization plan for large distributed computer applications (Cost Estimating Group 1999; International Society of Parametric Analysts 1998).

In the case of immunization registries, the numerous applications being developed provide enough data to deconstruct expenditure data (such as personnel and equipment costs) and examine how performance issues (such as expected database size, number of end users, communication and record retrieval speed, and reliability) and application objectives (such as reminder/recall notifications, identification of duplicate records, and assuring patient confidentiality) are related (see Table 1). In doing so, a cost model emerges that permits comparison of alternative methods for accomplishing the same policy objectives. In other words, it becomes possible to answer the questions: