Integrating Hospital Information Systems. The challenges and ...

0 downloads 0 Views 621KB Size Report
ABSTRACT. With the new technologies availabletoday, more ... hospital databases, managed by new generation ... systems, it was decided to provide end-users.
Integrating Hospital Information Systems. The Challenges and Advantages of (Re-)Starting Now. Umberto Tachinardi, MD, MSc* Marco A. Gutierrez, EE* Lincoln Moura, EE, MSc, DIC, PhD* Candido P. Melo, EE, MSc* *

Divislo de Informitica InCor HCFMUSP Sao Paulo - SP - Brazil -

and Communication Systems") and RIS ("Radiology Information Systems") are examples of "sub-systems" that have slowly been incorporated into fully integrated systems [1,21. On the other hand, most hospitals have their systems running, meaning a great deal of investment already done. Multiple dedicated and/or stand-alone systems are usually found. However, owing to technical problems, is not always possible to exchange data among them. This poses a major problem for system administrators that cannot simply turn-off and throw away these systems and their data. The move must be carefully planned before any action is taken. Obviously the ideal solution for each case is unique, but there are some issues that are common to many cases. This paper briefly describes the history of InCor in this field, and its experience in planning and designing its own HIS.

ABSTRACT With the new technologies available today, more complex and useful Hospital Information Systems (HIS) can be designed and implemented. These new technologies have allowed that information from different sources and nature such as documents, images and signals be integrated within a single environment. Open standards, reliable networks, powerful hardware and software and lower prices are among the issues that make all this possible. One of the main issues is what to do with old systems that do not adhere to this new HIS concept. At the Heart Institute (InCor), a decision was made towards starting developing a new system called I3S. This paper gives a brief description of that system.

INTRODUCTION Since their introduction a few decades ago, Hospital Information Systems (HIS) have evolved into useful and indispensable tools in health care management. Although early systems were developed mainly as an aid for administrative tasks, newer systems have gradually integrated more medical and scientific information. With the new technologies available today, more complex systems can be designed. Powerful software tools and hardware, at lower prices, reliable networking and standards add new prospects in this field. Important sources of information, such as the patient record and images can now be incorporated into large hospital databases, managed by new generation computers systems. PACS ("Picture Archiving

0195-4210/92/$5.00 C 1994 AMIA, Inc.

-

THE FIRST STEPS

InCor is a university hospital aimed at medical research, training and patient care in the field of cardiovascular diseases. InCor was founded in 1975 as one of the Institutes of the Sao Paulo University Medical School. By that time 5 Hewlett-Packard mini-computer systems were purchased for tasks such as ECG and blood pressure analysis. In order to be used by InCor, these systems had to be customized, but since there was no local expertise for that (at that time), some engineers were hired and given the task of understanding and adapting the software. It was not an easy task since many of the systems had

84

Working with a newly formed multiprofessional team was a challenge that required some extra "energy" to equalize knowledge and methods.

to be entirely rewritten in order to suit local needs. It is important to note that those systems, were developed for hospitals in the U.S. and not in South-American countries. The engineers thus hired formed the embryo of the Computing Division. In 1987, with the popularization of personal computers (PC), and with the increasing pressure for small departmental systems, it was decided to provide end-users with microcomputers. This decision has proven to have both positive and negative aspects. Users became more skilled in managing computers and a lot more interested in this subject. This was the most important benefit observed. The two main problems arising from that experience, were: 1) the use of multiple software titles made maintenance, training and development more difficult; 2) a high level of data-redundancy and inconsistency became evident, as departmental stand-alone systems kept information fragmented and unlinked. At that time some important advances in networking and database management technologies occurred, making it possible to devise an alternative solution for the problem of having users satisfied with their PC's yet keeping consistent information. The first step was to get an institutional commitment with the concept of an integrated network of hospital information. A strategy to put things together was then engineered. Called I3S ("InCor's Integrated Information System"), the project is a conceptual plan with some general guidelines regarding standardization and methodologies. The backbone of this project is a large corporation network using both "de facto" and "de jure" standards, in such a way that a heterogeneous environment can be fully connected using commercially available products. The creation of the Computing Division was one of the results of InCor's choice for a fully integrated HIS. The Computing Division was formed by two formerly distinct groups, one oriented to administrative applications and formed by data processing professionals, and the other, comprised by biomedical engineers, physicists, physicians and statisticians, oriented towards research and development in biomedical signal and image processing.

THE I3S PROJECT

Designing a fully integrated system for hospital information is a complex task. Every piece of data has multiple connections and its meaning and usefulness are specific to the application in which it is being used. For example, patient name and registration number are used as indexes for many databases, being thus vital to the integrity of the whole system

[3]. Modeling data and processes is a prerequisite to any integrated system. In the case of complex and ever-changing structures such as hospitals, there should be a commitment between the ideal and the possible. In our case only critical macro-structures were mapped at first. With the material thus created it was possible to have an overview of the needs and to plan the system's core. The main topics raised at the I3S design were: data and processes modeling; software development technology; quality and security issues and hardware and software requirements Typical HIS are implemented from one out of two main approaches: the administrative and the medical. The I3S was planned to congregate both of the two worlds. The I3S target is to merge HIS with PACS and RIS. This idea is not original in its essence [4], but our project introduces a singular approach regarding its view of Open Systems usefulness in HIS implementation. OPEN-HIS

The I3S was idealized at the time when open systems became fashionable. New hardware and software products were starting to appear. Object-orientation, graphical interfaces and new networking options were also being released, and helped to shape the I3S concept. The features presented by these innovations suited well our very own needs. Developing countries in general, with many financial and political problems present a very unstable environment to develop long-term solutions that require much time and money. Governmental rules changing on an almost daily

85

thus Database Management Systems (DBMS) must handle all these objects properly. Information is widespread distributed, but there are several forms of links among its pieces, requiring further control. The use of Relational DBMS's and even Object Oriented-DBMS, with SQL calls and the appropriate Application Program Interfaces (API), provides the means for co-existing with multiple platforms in a truly distributed environment. Most communication issues are supplied by network services. Medical images, however, deserve special attention. Sequences of tomographic slices can, for example, use a substantial amount of network resources. This problem can be reduced by compression techniques and faster pathways. Routing information within the network can be mostly done by network and DBMS services. in the Transactions hospital environment are usually done in real-time. There are even some descriptions of a specialized data bus for medical data [7]. From billing data to lab tests, reports must be available for the end-users as fast as possible. Although some processing can be done in the system's spare time, it is better to provide the system with a priority controller in order to manage processing and data-flow. Patient data are described by means of a combination of coded and free-text, since there is still no well established method for handling free-text automatically. In the I3S information is retrieved by means of coded-text which has been properly indexed. In general, data files are described by a series of different formats. For the I3S, some of the most popular and standardized ones were chosen. The choice for standard file formats allows the users to access data more easily. In order to make file formats widely available, format conversion software and proper drivers should be either bought or developed. Network services such as mailing, printing and file managing proved to be very important in many respects. Cost reduction and redundancy decrease are among the benefits brought by these services. Interfacing devices within health care services is a critical issue. Some vendors are still committed to proprietary systems and blackboxes, and seem not to understand that data is a property of the users. At InCor we are facing

basis, high inflation rates that force billing and payment systems to be more flexible, and the inefficiency of the national health policy, are only a few examples of the problems. In addition, the complex scenario posed by modem hospital characteristics such as the advances in health care technologies, higher costs and the fast changing and expanding medical knowledge, InCor's own features do not help to make things easier. InCor performs some 15 open-heart surgeries a day, 1,000 Nuclear Medicine exams a month and some 150,000 consultations a year. The capability to react fast and at a low cost is therefore essential to guarantee success of any enterprise in this area. All the above mentioned factors, along with InCor's group past experience, made the I3S idea more a bundle of norms, standards and definitions than a single computer-system. USER-ORIENTED HIS The aim of any information system is to satisfy its users by providing what they need at the right time and in the most suitable way. This statement seems to be obvious but most users seem to believe that systems are engineered to satisfy developers egos and create extra work for them. One of the challenges presented today is to design and implement systems that really fit user's needs, enhance productivity and are easy to use. User-friendly interfaces are a good start for that, graphic-interfaces with their windows are appealing and well accepted by most users. At InCor, the choice for preserving the users familiarity with PCs is helping to reduce the rejection for new systems. Enabling users to get information through their best-loved software (text editors, spreadsheets and DBMS), proved also to be a wise choice [5] although somewhat tricky. User participation in systems modeling has been essential. Hearing from their needs and their wishes, has helped to guarantee the acceptance of new systems [6] and cooperation with the information team.

I3S CHARACTERISTICS Information in health care systems are multi-modal (texts, images, graphics) by nature,

86

re-design of the system with a well studied substitution schedule. Theory in this case may prove to be very different from reality. Sometimes it is necessary that developers be ready to give provisional solutions and wait till a final solution can be achieved.

two distinct problems: the first refers to old analog equipment and the second to proprietary digital systems. For the former we have developed systems for converting data to the digital form. That group is formed mainly by old signal and image systems (i.e. ECG-recorders, spirometers, echocardiographs and x-ray). In the case of proprietary systems, the solution varies but demands heavy negotiation to be solved. From the introduction of I3S every new piece of equipment purchased by InCor has to meet our own interfacing criteria. Routine work with repetitive tasks sometimes require that software be developed in a more traditional way. Software development should be in accordance to the established rules and tuned to integration matters. The use of standardized software libraries, a common programming language, and new generation programming tools has helped to minimize integration problems.

REFERENCES [1] Mosser, H; Urban, M; Durr, M; Ruger, W and Hruby, W. Integration of Radiology and Hospital Information Systems (RIS. HIS) with PACS: Requirements of the Radiologist. Eur J Radiol, 16: 69-73, 1992. 2) Kimura, M. PACS and Patient Data Management Systems. Computer Methods and Programs in Biomedicine, 36:107-112, 1991. 3) Sideli, RV and Friedman, C. Validating Patient Names in an Integrated Clinical Information System. Proc Annu Symp Comput Appl Med Care, p. 588-92, 1991. 4) Boheme, JM 2nd and Choplin, RH. Systems Integration: Requirements for a FUllY Functioning Electronic Radiology Department. Radiographics,12 (4): 789-94, 1992 5) Burnakis, TG. Translating Mainframe Computer Data to Spreadsheet Format. Am J Hosp Pharm., 48(12):2619-22, 1991. 6)Abendroth, TW. End-User Participation in the Needs Assessment for a Clinical Information Svstem. Proc Annu Symp Comput Appl Med Care, p. 233-7, 1991. 7) Gardner, RM; Hawley, WL; East, TD; Oniki, TA and Young, BF. Real-Time Data Acquisition: Experience with the Medical Information Bus (MIB). Proc Annu Symp Comput Appl Med Care, p. 813-7, 1991. 8) Hammond, JE; Berger, RG; Carey, TS; Rutledge, R.; Cleveland, TJ; Kichak, JP and Ayscue, CF. Making the Transition from Information Systems of the 1970's to Medical Information Systems of the 1990's: The Role of the Physician's Workstation. J Med Syst, 15(3):257-67, 1991.

CONCLUSIONS As reported in the literature [8], the idea of a fully digital medical record emerges as a natural consequence of integrated medical information system. New technological achievements, such as more processing power, networks, high-resolution displays, powerful DBMS, have made it possible for InCor to decide for a re-start, including replacing old systems. One of the highlights of the I3S is the idea of connecting the users to the HIS via their workstations. PCs and other workstations are increasingly substituting office tools and becoming most valuable for daily work. People are getting used to PCs and feeling more comfortable with them. Data management in a distributed environment tends to be very complex and demands special attention. This is probably the main issue in a modem environment. Since users are more on control of information than the information group, the new approach requires robust mechanisms to ensure information quality and correctness. Security is a special problem as there are overlapping areas between access restrictions and system usefulness. Co-existence with older non-integrated non-open systems deserves a case-by-case approach. The decision at InCor was for a total

87