Ninth Annual Maternal and Child Health Epidemiology Conference / December 10-12, 2003
ALAN HINMAN: Thank you very much, Kay. I'm going to shift gears a little bit here because you've been hearing primarily about the use of linking data sets for research purposes. Much of what I'm going to be talking about is work that we've carried out in conjunction with the Genetic Services Branch of the Division of Services for Children with Special Healthcare Needs and the Maternal and Child Health Bureau of the Health Resources and Services Administration, otherwise known as GSB HRSA. And since 1998 the Genetic Services Branch has been providing grants to facilitate the development of integrated child health information systems to include newborn screening systems. They also are trying to facilitate the opportunity to improve service delivery to children and their families that's community based, culturally competent, comprehensive. This sounds like a definition of the medical home and it is intended to do so. Then to facilitate the enhancement of the ability to coordinate care across multiple programs and providers.
Now before getting started I want to talk just briefly about issues of linkage and integration of information systems. Some people think of linkage as taking existing information systems and enabling them to share information, and integration as building a giant database that has everything in it. From our perspective integration involves providing a range of information to the end user in a useful, simple, comprehensive format so that he or she can readily take all appropriate actions as needed. So integration, from our perspective, relates to the end user, not to the background machinery, not the hardware or software. It is to serve the purposes of the provider, the patient or family, or the program. In conjunction with the Genetic Services Branch we have over the past couple of years developed what is called a source book for planning and development. It's titled "Integration of Newborn Screening and Genetic Service Systems With Other Maternal And Child Health Systems."
Part of the handout is a two-page summary of this document. It is available on our website or at the Genes-R-Us website that's operated through funding from the Genetic Services Branch. And a part of this activity was to try to, through a series of site visits of integrated Child Health Information Systems to try to develop a roster of what the key elements for success in this kind of an endeavor are. And these key elements are really pretty straightforward. They're on the handout. They start with leadership. Leadership to define and perpetuate a vision and sustain support. Governance. Governance includes defining functional outcomes, creating accountability, setting priorities, making major policy decisions and overseeing finances. And that's to be differentiated from project management, which is the actual implementation of the Program. So these are separable functions and each one of them is critical to the success of an integrated information system. Involvement of stakeholders defined as broadly as possible, including from the outset in the case of Child Health Information Systems families and parents, organizational and technical strategies, technical support and coordination, financial support and management, policy support and evaluation. And as these are described in a little bit more detail in the handout and in considerably more detail in the document that's on the website.
From these key elements and the site visits we derived five primary lessons. The first one is that data are for sharing. Pat *Nolan, the State Health Officer in Rhode Island has a sign in her office that says, "Information is not a possession. Information gains value as it is shared and used." We need to remember that. The second lesson is listen up. Listen to what people are saying and what they want. The third is that change is hard. This is not an easy process. There's a lot that has to go into it. Fourth is let program needs drive technology rather than the other way around. The technology should serve the program needs. A crude way of saying this is, "Don't put the geeks in charge." And the notion is that the Program, as Milt pointed out, the technology is not the limiting factor. One of the limiting factors is knowing what it is you want and being able to enunciate that and being consistent about that. And one of the real problems that program people provide to their technical counterparts is mission creep. They say, "Here's what we need. This is all we need." And the technical folks go out and start working on it. And then they come back and the program folks say, "Well, yeah, but you know, we really could use a little bit more of this and that." And you need to be clear and consistent and explicit about what it is you really need and what you cannot do without. And the final lesson is to stay the course. This is a lengthy proposition. It's not going to get done today or tomorrow. It takes time. It takes years. And it takes money.
Now one of the functions of the grants that are being given out by the Genetic Services Branch, and there are now been a total of 15 of these, the most recent set have interestingly been joint projects funded by the office of data information and management ODEM, that Mike Kogan is in charge of at HRSA, as well as the Genetic Services Branch. And these are for, among other things, the purpose of developing child health profiles to provide up-to-date information about children's health status to families, healthcare providers, public health programs, thereby facilitating appropriate care such that authorized users, and authorized users can include families or private practitioners or health programs, can determine at a glance the child's status with respect to all components of the Program. And ideally, this is done in a push fashion. It's in-your-face information. For example, in Missouri what's being done is to develop essentially a dashboard approach in which there is a series of red-green-yellow signals on the screen when you open it up. And if a child is in good shape for immunization, that's green. If the child is due for immunization, that'd be yellow. If they were overdue, that'd be red. Similarly for other programs that are components of the integration process. You want to make as easy as possible for people to do the right thing. And programs also can view information about individual children or summarize information.
Now what was done was to start with four programmatic areas; newborn, dry blood spot screening, used to be called newborn metabolic screening, but of course not all of those tests are metabolic tests and they're not all genetic tests, so the one thing they share is that they're dry blood spot. Because they're not all heel sticks either. Because heel stick is a term that's been used in some places. The newer early hearing detection intervention, or EHDI Program, immunizations and vital registration. These four areas were chosen because they share certain characteristics. They're recommended for all infants and children; they're carried out or begin in the newborn period, usually in the hospital at the time of delivery. They are time-sensitive. If they're not done in time, or not followed-up promptly there can be an adverse outcome. They're primarily delivered in the private sector. But they involve very substantial public sector involvement and they're mandated in most or all States. In fact, dried blood spot screening, immunizations, vital registration are mandated in all States. Early Hearing Detection and Intervention, newborn hearing screening is mandated currently, I believe, in about 40. But that number is going up and will reach 50 in the not too distant future.
Now one of the things to try to do was to figure out what it looks like when you try to bring all of these things together and integrate them. And so part of the task was to develop a draft model of practice or framework for integrating newborn screening systems with other related early child health information systems to include a comprehensive set of core functions, activities and services. And the notion here is that all of these program areas have certain things in common. Now you know you've heard that if you see one Health Department, you've seen one Health Department. That's true. But there are things in common in all Health Departments. They all do certain things in common. And the task was to see if we could define what those things were that these different information systems had in common.
We convened a core work group, which included these folks, *Bill *Matkinson from the National Center For Health Statistics, *Tonya *Dean, the State Genetics Coordinator in Iowa, *John *Ikewald, the Head of the Charm Project in Utah, and your handout, by the way says Massachusetts and please if someone is here from Utah, tell *John that I did correct it on the slide for presentation. Jennifer *Heberer, a parent of a child with special healthcare needs from Maine, *Therese *Hoyle, the Manager of the Immunization Registry in Michigan, Pam King the State Genetics Coordinator in Oklahoma, Bob *Kosic a practicing Pediatrician from Massachusetts, Donna Williams from the National Newborn Screening and Genetics Resource Center, and they're the operators of *Genes-R-Us, probably the niftiest URL that you can encounter, and *Amy Zimmerman from the Rhode Island Department of Health. In addition, from the Genetic Services Branch, we had *Michelle Lloyd *Perger, the Branch Chief, Maureen *Mann and the Project Officer, *Deborah *Lindser, and from All Kids Count and Public Health Informatics Institute, several of us.
Now, in describing this framework, we said that this was going to describe a set of activities and functions which would help achieve the desired outcome, improving health and health services of children. And we were focusing, as I mentioned, on integration of these four program information systems building on approved practices and standards and providing a minimum set of core functions. And this is not a technical model. What we did was to take from these four program areas the existing programmatic standards and in the Childhood Immunization Arena in fact there are standards. The National Vaccine Advisory Committee has published standards for pediatric immunization which have been endorsed by the American Academy of Pediatrics and a variety of other organizations. In the Newborn Screening line there have not been standards developed as yet, there are some under development now by the American College of Medical Genetics, these will be published next year. But there are guidelines and recommendations for what should be done in terms of screening both for hearing and for dried blood spot screening.
Then we looked at the functional standards for information systems and in the immunization arena, again, there are definitions. A technical working group convened by the Center for Disease Control and Prevention developed a set of 12 immunization registry core functional standards. It says any immunization registry worthy of its name ought to be able to do these 12 things. So what we did was to say, "Okay, if this is what you need to do to support an immunization program, will these functions also support the information needs for a newborn dried blood spot screening program? Yes or no? And if no, do we have to just tweak one of those functions of an immunization registry or do you have to write something entirely new?" Well, as I say, we reviewed all of these things and from a roster of 12 core functions for immunization registries, we ended up with 30 functions for an integrated system integrating four different program areas, which are quite disparate. And actually I think that's pretty darn good. We thought we might end up with 60 or 70 as we got started, but it turns out, once again, there are a lot of things that these programs have in common.
We also identified 19 over arching principals. I'm not going to go into those in detail, but in your handout you have the summary of the principals and core functions. We described the current system of data exchange between the information user and individual public health programs. We didn't use silos but we did use boxes. It's one to one. It's here if you want to find out if the child has a birth record, you have to go to the vital statistics folks. If you want to find out if metabolic screener dried blood spot screening was done, you have to go to that program. Similarly for hearing. Some of these are hospital based, some are health department based, some are laboratory based and it's difficult to get all this information in one place. And what we hope to achieve through integrating the information is to enable the information user to go to one place and get all this information. And it also means that these different programmatic areas can get information through the integrated system also. So an immunization program, for example, could determine if lead screening was a part of the integrated system. Whether there are children who were under-immunized, were also under-screened for lead. And guess what? They are.
So we developed these things and then sent them by an external review committee, made some further modifications, they've been submitted now to the Genetic Services Branch and have been presented to the various grantees, to their grantees I'm presenting here now. We'll be talking about these with the Early Hearing Detection Intervention Grantees in the spring and we hope to be able to go on from here. The points to keep in mind are that these principals and functions relate to integrated systems. We realize that individual program areas are going to maintain their own information systems because they may have a lot of information that everybody else doesn't need. They don't speak these principals and functions, they don't speak to system architecture, to the data elements that you use, or to the software. They talk about what the functions are, not how they're to be achieved. And the principals and core functions document that you have I think has a heading that looks like this.
There were 19 principals, one of which was stating the overall purpose, then no surprise, five addressing security and confidentiality. As Milt pointed out this is, if not the, it is mighty close to the principal concern of people involved in information systems, individual information systems, not just integrated or linked information systems. A series about how technology should serve the stakeholder needs, quality assurance and evaluation in financing. Then in the core functions, there were five on confidentiality and security, four on establishing and maintaining records having to do with timeframes in which this should be accomplished. Remember these are serving programmatic needs so that ideally when the child comes for a two-week post-birth visit, you would have this information available. We haven't yet reached that point, but at minimum, by the time the child comes for a two month immunization visit, you should have information about whether the child got a dose of hepatitis B vaccine in the hospital, whether the child had dried blood spot screening and if so, whether the results were normal, or if not normal, was the child followed-up. Similarly for the other activities.
There are six functional statements on services and four on technical issues and three on reports. Then we had a series of desirable functions of which, again, some were on establishing and maintaining client records. And some of these have to do with how rapidly you can deal with children who move or go elsewhere, the *mogies. And this is a real problem for anyone who is operating a programmatic information system. Now the next steps in developing integrated child health information systems, we believe, are to come to agreement on core data sets and information transfer standards, to develop and begin to use performance measures to tell how well these are doing. In the immunization registry arena we have developed performance measures that can give us an indication as to how well the registry is performing. We need to do that for these integrated broader systems. Document their impact on outcomes.
The whole purpose of these things is to improve *help and health services and we need to be able to document that they're accomplishing that. That will take time because there are no integrated systems that are fully functional at the present time. Rhode Island may come as close as anyone else, but they still have some gaps also. We do know that individual information systems, well developed and used, can improve health. We have good evidence in the immunization registry arena that using a immunization registry can improve immunization coverage, can decrease missed opportunities and improve the health services. We also have information that the absence of information is harmful. You can make a calculation from the reports on newborn dried blood spot screening tests that suggest that, at least on the data that are submitted to the national level, you can estimate that there may be 14 cases of PKU every year that are lost to follow-up. That's just by applying the diagnosis rate on those screened not normal, to those who are reported as lost to follow-up. $2.4 million settlement recently in Washington State to a child with PKU which was missed. So we don't have direct evidence on the benefits of the integrated system. I submit it's not a large leap of faith to go from documented benefits for a single information system, clear problem of missing information, to the likelihood that integrated systems will, in fact, improve things. But we need to identify, we need to document this, we need to develop more information on the costs and cost savings associated with integrated systems and identify funding sources to sustain them.
Last week, we convened an invitational conference in Atlanta on developing child health information systems to meet medical care and public health needs. The conference was co-sponsored by this wonderful alphabet soup and if anybody has any question as to what any of these are, I believe I can correctly decipher the alphabet. The objectives of the conference were to review national initiatives and other factors influencing the develop of child health information systems infrastructure and develop based on that concrete recommendations for immediate actions and actions over the next three to five years. And finally to enlist stakeholders and communicating, supporting and implementing the recommendations. Your handout does not include, but the presentation does because the handout came in a week ago before the conference was held, what in essence the overall vision of the conference was, which was improving children's health and health services through the timely provision of accurate and comprehensive information. My personal summary of some of the important things at the conference were that, first of all, these integrated information systems aren't at all useful until the information is used to improve health and health services.
One of the greatest weaknesses of immunization registries, which is the area with which we have most familiarity, one of their greatest weaknesses is that people spend so much time trying to be sure that every last little bit of the data system is perfect that they lose the utility of the system. They fail to realize that by providing even imperfect information, you can provide an important service, you can gain support and you can get better quality information next go around, because people will begin to depend on it and demand it. There are a variety of efforts underway to develop clinical and public health child health information systems. Commonly these are program specific, the silos that people have been talking about, with little attention being paid to integrating the information from those systems and that conscious efforts towards integration are essential and must continue. The conference ended up with a series of action steps in four areas, governance, information infrastructure, economic issues and use of information. These will be posted on our website www.allkidscount.org in the next 10 days with luck, in the next two weeks without luck, and a meeting summary will be posted on our website in mid-January 2004. And if you have any questions, please contact me at my e-mail address. Thank you very much.