INFORMATION MANAGEMENT IN THE

CLINICAL LABORATORY

D. CALVA MENDEZ, A. LANDA QUEZADA, M. LEHMAN

Software Integral para Laboratorio (SOFILAB) SACV

Lisboa 14-A, Col. Juárez, México DF, 06600 México

Abstract: - Developing a Laboratory Information Management System (LIMS) able to offer enough flexibility and functionality to most clinical laboratories, especially when a certification process is involved, becomes a process in which decisions must be taken. The main achievement of our system is the possibility to adapt to the necessities of each clinical laboratory. This gives to the system certain interesting properties we analyze here.

Key-Words: - LIMS, clinical laboratory, data patient record, image processing

1 Introduction

Laboratory Information Management Systems are important and well-known tools in most clinical laboratories and have been developed for years [1]. However, today's challenge has to do with designing and developing a system able to provide customized solutions without prohibitive capital or maintenance costs in money and personnel. Most laboratories have to accurately and quickly process and manage overwhelming amount of data for the tests performed, the way these tasks take place differ from one laboratory to another and make the sample testing life cycle a very particular process [2]. The challenge of designing a particular solution for each different place complicates when regulatory compliance and certification are involved, since most of the times these implies changes in the way the sample testing life cycle is handled. Although this cycle is mostly the same everywhere as shown in Fig. 1., the paper work involved and the formats each laboratory uses differ from one place to another.

Having a LIMS designed for the laboratory personnel to customized it according to its emerging needs, means that apart from having to do with all that is implicated in a certification process, the laboratory personnel will have to deal with programming whichever new report or catalogue is needed. The proposed LIMS is one the end user will not have to program or customize. In order to achieve this, the database design will have to be predictive in order to consider and contain information the laboratory is not yet taking into account in their daily processes and tasks, but that in the future will be requested either by certification demands or any new need or request form the laboratory or hospital administration.

2 Information System Structure

In developing a LIMS, several aspects have to be considered based primarily on the sample-testing life cycle. This assures all the information to be contained in the database will comply with the laboratory actual needs. Apart from these, certification demands have to be considered since constant changes resulting from certification and audits are common especially when these processes are in their initial stages.

The information managed in the clinical laboratory, begins with the patient identification (usually a set of characters) and from there, grows exponentially according to data associated with the sample-testing life cycle (type of sample, test names, results, reference values, etc.) [3]. Because of the amount of information managed daily in the clinical laboratory, most LIMS including the one presented here choose a relational database model, which also allows the management of information based on the application of mathematical set theory which makes it flexible and practical.

The database analysis and design is the most important part in developing a LIMS. In the current design we tried to consider every possible data, information or even image, currently used and needed by the laboratory, together with information that according to certification procedures and experience will be requested in the future in order to construct a relation-entity diagram. This kind of diagrams are specially useful in a structured design since they show the relation and hierarchy each data has in respect with the others, and therefore every function in the final application depends directly and entirely on this first approach [4]. Information was grouped according to the specific tasks and activities in the clinical laboratory in order to construct separated modules with their specific relations, and then the relations between each of these modules were considered. The way information was grouped and related is shown in Fig. 1.

Figure 1- Work flow in a clinical laboratory.

3 Implementation and Results Obtained

The development of the LIMS first version, from analysis to implementation took a year, however, laboratories emerging needs took us to detect several weaknesses as well as automation needs in several clinical processes which take place in the laboratory as a routine. This kind of clinical processes have to do mainly with the microscopic analysis of samples such as urine and fecal grounds. The aspects that traditionally have to do with tasks the LIMS should perform were pointed out during and after implementation.

The implementation process involved the whole laboratory personnel and was divided into three stages: a) Familiarization with the interface and the use of the computer, since none of the personnel had any previous experience, b) Automation and knowledge of the different steps and shortcuts the users needed in order to perform their tasks faster than when there was not a LIMS involved in the sample-testing life cycle, c) Gathering, classification and discrimination of information from the users feedback, for the modification and optimization of the LIMS. This led to the planning of the LIMS second version, where the database had to be predictive and take into consideration every aspect of laboratory certification, specially when dealing with printed reports and document control, the aspects of internal and external quality control were also included although most of the laboratory instruments can document and perform the needed statistics and manage the associated results, since managing such information from the LIMS homogenizes the criteria as well as reports, and consolidates all the information in a single database [5].

With the LIMS first version, the most basic laboratory needs were covered, the most significant result was the easier and faster access to information, and the reduction of paperwork, as well as an accurate knowledge of the laboratory demand on each test performed, since this was made by only considering and manually counting the medical orders and not taking into account the repetitions, quality control and instruments calibrations, since this information is gathered by the LIMS from the instruments, the task which used to take a couple of days takes now only a couple of minutes. Fig. 2 shows the windows of second version of developed LIMS.

(a)

(b)

Figure 2- LIMS Sofilab: (a) main window, (b) user interface.

3.1 LIMS with image processing

Many of the clinical laboratory tests results, take numerical values which have to fall in between a determined range for the physicians to interpret. However, there are some tests which are already an interpretation of one or several images coming from a microscope, a petri dish and other different sources, some of these tests are the microscopic analysis of samples such as urine, blood, fecal grounds, protein electrophoresis and disks susceptibility tests [6].

Figure 3- Control window for image acquisition and processing.

The results of these kind of tests are always written interpretations and do not include the image of what its being interpreted. Moreover, the interpretation of the samples observed on a microscope, a petri dish or an electrophoresis gel can not be preserved easily due to the natural degradation of such samples and therefore have to be analyzed as they are prepared or received in the laboratory and when a second opinion is needed it is not always easy to transport the sample to another place or there is not enough time for another chemist to arrive to the laboratory and make an interpretation of the sample, also if the physician is not sure of the result interpretation he is not always able to see where the interpretation came from. In order to solve this problem a PACS is added to the LIMS, giving it the possibility of keeping record of the images interpreted so that the interpretations can now be backed up [7,8]. An example of this can be seen in Fig. 3.

This is helpful when dealing with legal issues and certification purposes, since it makes a sample result easy to follow up or justify whenever needed.

4 Conclusions

Eventhough the sample test cycle in clinical laboratories is the same, every laboratory has special needs and ways of dealing with the day to day situations. This makes developing a general LIMS for every laboratory difficult. The LIMS presented here is developed according to the sample test life cycle, but it also considers such special needs. In order to have a robust product and also to meet short developing times, we introduce the use of a database, where most of the issues that might be requiered be included and from there the LIMS development is done according to the laboratories specifications. Also, including an archive of images helps backing up interpretation of results coming from samples observed in the microscope or other sources such as petri dishes.

Acknowledgments

This work was supported by Software Integral para Laboratorio (Sofilab) S. A. de C. V., and partially supported by Consejo Nacional de Ciencia y Tecnología (CONACYT) and Secretaría de Hacienda y Crédito Público (México), through the project NeuroSofilab (Ref. SOF-971021-I75/2002-1).

References:

[1] Paszko Chritine, Pugsley Carol; Consideretions in Selecting a LIMS, American Laboratory, September 2000.

[2] T. Kisiel, Laboratory Information Management System (LIMS): Steps to useful implementation, DAI White Papers, Digital Applications International Ltd. (2002).

[3] R. Megargle, Laboratory Information Management Systems, VCH Publishing (2002).

[4] K. Siau, Helth Care Informatics, IEEE Trans. Inform. Technol. Biomed. 7(1), 1 (2003).

[5] D. L. Nigon, Clinical Laboratory Management, Mc Graw Hill, New York (2000).

[6] R.McDowell, editor. Laboratory Information Management Systems: Concepts, Integration, Implementation. Sigma Press, Wilmslow, U.K., 1985.

[7] Handbook of Medical Imaging, Kim, Horii, ed., Spie Press (USA) 2000.

[8] S. Chu, Information Retrieval and Health/Clinical Management. Synopsis, Yearbook of Medical Informatics, IMIA 2002, 271.