|Abstract. The evaluation of client’s/patient’s health condition is the basis for determination of nursing problems/diagnosis and the nursing planing. Changed health condition to better is also the indication of successful performance of nurse’s work. The evaluation of the appropriateness of nursing interventions is based on positive alternations of individual daily activities. That is why a constant, in-fixed-intervals monitoring of client’s changed health condition on the basis of attributes or indicators, determined within the individual daily activities, is suggested. At that stage a problem, how to aggregate the individual estimates of indicators into an overall health condition evaluation, emerges. A multicriteria decision-making theory is used for that purpose. Practically that means the introduction of the measuring scale of 3 or 5 descriptive values and the aggregate function into the hierarchically structured indicators, which are in accordance with the International Classification for Nursing Practice (ICNP). In the following article the evaluation case using the shell of expert system DEX (Decision Expert) will be presented. The logical operations and the transparent interpretation of the evaluation within the expert system DEX enables a full presentation of the aggregated knowledge.|
Community nursing (CN) is defined as a special form of health care that assures an active health and social care of individuals, families and communities that are, due to their biological features or a particular disease, more exposed to harmful effects from the environment. In Slovenia it includes: health and social treatment of individuals, families and communities, nursing of women in childbed, new-born child at their homes and nursing of patients at their homes. Community nurse works on primary, secondary and tertiary prevention and for promotion of health , , [3i].
CN is an integral part of the primary health care and it plays an important
role in implementing WHO goals. In the process of CN is the individual,
his family and community the subject of treatment. V. Henderson´s
and D. E. Orem´s nursing theories are theoretical basis for philosophy,
scope and the nature of the CN , .
They are orientated into responsibility for individual own health and into
support of risk groups. That is why their implementation value is in accordance
with the WHO strategy’s goals. Both theories define fourteen daily activities
but none of the theories has its objective evaluation system for evaluating
the achieved or altered health condition of individual, patient or community
installed , .
The evaluation of the client’s health condition is the basis for the determination
of nursing problems or nursing diagnosis and nursing planing. This provides
CN with feedback which is not only essential for CN activity but it also
enhances it’s quality and effectiveness. The evaluation of
the appropriateness nursing activities is based on positive alternations
of basic daily activities. When monitoring the client’s health conditions
alternations in the fixed–intervals in daily activities we are met with
the problem of aggregating the evaluated indicators into overall evaluation
of the client’s or patient’s health condition and his independence.
2. DEX - An Expert System Shell for Multi Attribute Evaluation
DEX closely follows the concept of multi-attribute decision making , which is based on the decomposition of a decision problem into smaller less complex problems. Options are decomposed on to different dimensions, usually called attributes, performance variables, criteria, etc. These are evaluated independently. The total utilities of options are then obtained by some aggregation procedure for example a weighted sum. The procedure is designed by the decision maker as to best represent his or her preference knowledge about the options. The obtained utilities are finally used for option evaluation.
In DEX, this approach is combined with some elements of expert systems and machine learning , . Attributes and aggregation procedures are treated as an explicit knowledge base that consists of: (1) tree of criteria, (2) aggregation procedures expressed by decision rules, and (3) description of options. The components will be presented in next section.
DEX basically consists of two operational parts: (1) knowledge based acquisition and (2) evaluation and analysis of options. The first part supports the user in designing the criteria tree and decision rules for particular problem. This is actually the process of decision problem structuring and preference knowledge elicitation, which is continuously supervised by computer based tools for checking the consistency of rules. The second part of DEX, which is actually shown in Figure 1, utilises the so acquired knowledge base to evaluate and analyse options. At the beginning, each option is described by a set of values that correspond to the leaves of the criteria tree. DEX then evaluates each option according to the knowledge base, i.e., according to the structure of the criteria tree and defined decision rules. For each option, an adequateness estimate is obtained. The analysis of the results can follow, which consists of one or more of the following activities:
In summary, the decision support offered by DEX is based on preference
knowledge modelling. Such a support makes decision analysis transparent
by providing the decision maker with the explanations regarding the evaluation
results and the background of the evaluation process itself. The knowledge
representation is based on an integration of multi-attribute decision-making
approach with expert systems. This offers a user-friendly decision support,
where decision knowledge is expressed simply and naturally by words, rules
and hierarchies of criteria.
3. Evaluation of Client’s/Patient’s Daily Activity
The evaluation will be explained on daily activity Physical Activity. The activity according to ICNP , ,  is presented in Figure 2 as hierarchical tree structure. The knowledge base that has been developed for the evaluation consists of the attributes (indicators, parameters) in that tree and the decision rules determine in each node of the tree the interdependent effect of lower parameters on the node - aggregated parameter.
Table 1 presents a DEX print-out of linearised tree from Figure 2 together with description of attributes. Each attribute is measured along the 5 grade Likert-type scale (Table 2). So the problem of each attribute concerning physical activity can be expressed on the scale from no problem to a problem to a very high degree.
Table 3 presents decision rules in a complex form for aggregation of mouth, eyelid and throat paresis into estimate of overall paresis problem of a patient. For example 1st decision rule can be read as: If there is a problem of a very high degree concerning mouth paresis then regardless of the eyelid and the throat (* means any value) the overall paresis problem is a problem to a very high degree.
Rule 11 can be read as: If mouth paresis is in the interval between 4 (to a high degree) and 3 (to some degree), regardless of the eyelid (including problem to a very high degree), and the throat problem to some degree the overall paresis problem is a problem to a high degree.
These rules are the result of professional consensus and are opened for discussion.
Table 4 presents 3 successive client’s/patient’s estimates of daily
Physical Activity monitored on successive visits. There can be seen a significant
improvement of his or her physical activity from visit V1 to visit V3.
The proposed evaluation of health condition demonstrates ICNP and contemporary information technology as a challenge to deepen the system approach in nursing. DEX as a part of cybernetic loop in nursing process offers a possibility for better (self)control and consequently higher professional work and better quality of nursing care. The approach can be also seen as an active method for testing ICNP in the nursing process.
Knowledge-based multicriteria evaluation and ICNP is a fruitful combination for information support of nursing process which leads towards more client orientated approach using problem solving strategy. Such approach does not only assure a greater functional adequacy of the system but also stimulates the user, i.e. nurses, toward research and development work.
We believe that such an evaluation could be an integral part of an information system for community nursing [3ii], , . Modern information technology (portable computing, networking and multimedia)  offers the possibility of integrated onsite information support. The use of IT does not merely mean that a machine operates or helps us to make what we have been doing manually. The use of IT tools is also a challenge and opportunity for new ideas and solutions.
Further work will be focused on development of appropriate knowledge bases for all 14 daily activities and their testing in practice. Later on the expert system should become an integral part of IS for community nursing, especially in connection with nursing diagnoses and interventions. Some organizational changes regarding process of nurses’ work are also expected.
The work presented in this paper was supported by The Republic of Slovenia,
Ministry of Science and Technology, The Slovene Science Foundation and
INCO Copernicus Project: Co-operative Research in Information Infrastructure.
We would like to thank the subproject Computer-Aided Information System
for Community Nursing team members and specially Dr. Marko Bohanec for
his contribution regarding DEX. We are also grateful to The Danish Institute
for Health and Nursing Research and Mr. Gunnar Nielsen for his personal
support and encouragement.
 M.J.D. Clark, Community Nursing. Health Care for Today and Tomorrow. Reston Pub. Comp., 1984.
 V.K. Saba, A Home Health Care
Classification (HHCC) of Nursing Diagnoses and Interventions. HHCC Project.
Georgetown University, School of nursing , 1994.
[3i] [3ii] O. Sustersic, V. Rajkovic, An Information System for Community Nursing Support. In: Technology and Informatics 34, Medical Informatics Europe 96, Amsterdam, 1996, pp. 665-9.
 V. Henderson, Health Records and Nursing, Conn. Nursing News, 1985.
 D.E. Orem, Nursing: Concepts of Practice. McGraw-Hill, New York, 1980.
 V. Chankong, Y.Y. Haimes, Multiobjective Decision Making: Theory and Methodology. North-Holland, 1983.
 M. Bohanec, V. Rajkovic, DEX: An Expert System Shell for Decision Support, Sistemica 1 (1), 1990, pp. 145-157.
 V. Rajkovic, M. Bohanec, Decision Support by Knowledge Explanation, In: H.G. Sol, J. Vecsenyi (eds.) Environments for Supporting Decision Processes. North-Holland, 1985, pp. 47-57.
 The International Classification for Nursing Practice ICNP with TELENURSE Introduction. Editor: R.A. Mortensen. The Danish Institute for Health and Nursing Research, ICN- International Council of Nurses, Genova, 1996.
 G.H. Nielsen, R.A. Mortensen, The Architecture of ICNP: Time for Outcomes-Part I. Int. Nurs. Rev. 1997, pp. 44,6:182-88 and 176.
 G.H. Nielsen, R.A. Mortensen, The Architecture of ICNP: Time for Outcomes-Part II. Int. Nurs. Rev. 1998, pp. 45,1:27-31.
 G. Rolfe, Closing the Theory-Practice Gap. A New Paradigm for Nursing. Buterworth Heinemann, 1996.
 M.G. Stineman, A. Jette, R. Fiedler, C. Granger, Impairment-Specific Dimensions Within the Functional Independence Measure. Arch Phys Med Rehabil, 1997, pp. 636-43.
 W.T.F. Goossen, P.J.M.M. Epping, I.L. Abraham, T.W.N. Dassen, A. Hasman, Problems with Nursing Information Systems: are there Solutions. In: Technology and Informatics 34. Medical Informatics Europe 96, Amsterdam, 1996, pp. 872-6.
 T. Albreht et al., National
Programme of Health Informatics in Slovenia – Its Development and Perspectives.
In: Technology and Informatics 34. Medical Informatics Europe 96,
Amsterdam, 1996, pp. 892-6.
 I. Testelmans, P. Palmers, Mobile Nurses and Their Computerized Helpsystem in ICARUS.