This is an announcement that this blog will host a Soft Systems Methodology (SSM) model for Parenting. Using the blog as a tool to gather comments from parents with different parenting Weltanschauung (worldview), a master model will be created. SSM for CDSS will continue to be offered free and I invite your suggestions and comments to build a group to discuss about SSM.
Lee Philip and Loo Grace (2001),"A Soft Systems Methodology Model for Clinical Decision Support Systems", DEXA 2001 Proceedings, IEEE Computer Society Press.
Abstract
A systematic approach for the development of clinical decision support systems is essential for an effective design and implementation of medical information system/ (health informatics). This research paper proposes the use of the Soft Systems Methodology (SSM) as a tool for analysing the Clinical Decision Support System of medical information system using various models.
From the SSM model so developed, a quantitative survey is designed and used for analysing the attitudes of doctors towards the use of clinical decision support systems. Useful findings have been obtained from a research project conducted in Malaysia. These results provide useful pointers for the development of national health information system in Malaysia and some other countries of similar background, and they serve the needs to minimise wastage of resources and optimise benefits generated from the use of Clinical Decision Support systems.
Keywords: Clinical decision support systems, health informatics, medical information systems, Soft Systems Methodology.
1. Introduction
Many countries are focusing on developing its medical information system (MICIS), in order to keep pace with the progress made in other aspects of national and international development. America, Australia and Europe have been actively developing their MICIS. The Good European Health Model (GEHR), later known as the Good Electronic Health Model, started in the early nineties [Schitcliff 1999] is one such example. In Australia, Malaysia and New Zealand medical specialists and information systems researchers are actively working on health informatics to advance the quality of life of their people.
New Zealand though a small nation, follows closely with other developed countries in health informatics. Since the early eighties, its Ministry of Health has
embarked on certain national health strategies in a structured manner. Information management and technology is vital for providing the ability to exchange high-quality information between partners in health care processes, focusing on achieving better health outcomes. Clinical Decision Support Systems (CDSS) is a natural choice of information system tool.
Malaysia develops its health care system in its own way. In various parts of the country, there is an integration of western medical science and the traditional health values (alternative medicine) during the last decade. However CDSS has not gained grounds among the doctors, as in some western countries by the late nineties. This research project initiated in Malaysia in 1999, highlights the associated issues and proposes some solutions for consideration by clinicians and researchers. It investigates the attitudes of doctors, socio-technical issues and past implementation efforts and offers suggestions to improve adoption of CDSS for MICIS.
In this paper, section two discusses clinical decision support systems (CDSS), section three introduces the background of Soft Systems Methodology and its usage in SSMM-CDSS. A description of a proposed model and implementation of the research done is given in section four, with findings presented in section five. This is followed by the conclusion, future work and acknowledgment.
2. Clinical Decision Support Systems
Decision Support Systems (DSS) are interactive computer based systems that assist decision makers to utilise data, models, solvers, and user interfaces to solve semi-structured and /or unstructured problems [Sprague 1980]. In this research project, clinical decision support system (CDSS) is defined to be any software (DSS) designed to directly aid clinical decision making, whereby through the use of specific CDSS, useful characteristics of the patient are made available to clinicians for considerations [Hunt et al. 1998].
CDSS does not make decisions but supports diagnostic decisions of doctors. CDSS is viewed as information technology, defined as mechanisms to implement desired information handling in the organisation [Avgerou & Comford 1993]. Thus CDSS also supports work procedures and organisational goals. There are studies done on CDSS in the areas of drug use and preventive medicine. The quality of such studies has been increasingly improved over the years since the late 1950s [Shortliffe 1987]. Some CDSS incorporates fuzzy logic to overcome restrictions of diagnostic programs. Pathophysiologic reasoning has also been used to represent temporal evaluation of single and multiple disease process [Mazoue 1990].
The conclusion drawn by a review done by Hunt et al. [1998] over a twenty-five years period showed that the CDSS improve health practitioners’ performances, but however, less frequently, improve the patients outcome. Through a case study done for this project, it had been shown that the adoption of CDSS in Malaysia has met with unfavourable comments. The review from Hunt et al. could have asserted some influence.
3. Soft Systems Methodology
In the early seventies, when software developers were faced with ill-defined problems or situations that required information systems development, there were very few established methods of systems engineering available. At that time, the systems engineering defining assumption was that the system of concern exists, could be named, and could be manipulated in the interests of efficiency [Checkland 1981]. This basic assumption held well if the optimum solution could be found for a particular situation. However this was not always the case with real-world situation.
Soft System Methodology (SSM) evolved then to meet that need when no optimum solution was available. SSM placed emphases on people’s perception of reality and worked with the notion of a problem situation in which various players might perceive various aspects to be problematical. SSM had a system of enquiry that was formally expressed to allow learning and make sense of complex situations to enable purposeful actions.
4. SSMM: Design and Implementation
Triangulation was used in this project, incorporating SSM, survey, official statistics, observation, interviews: According to McNeil [1985], it was sensible to use a mixture of methods (triangulation) in the overall research design. The strength of one method was used to compensate for the weakness of another method. Each method had its advantages and limitations. Both a quantitative approach (close-ended survey and official statistics) and a qualitative approach (SSM) were used to take account of values as well as behaviour. SSM was adopted not merely as a seven-stage model, but as a mouldable methodology for investigating unstructured situations. The emancipatory research strategy was utilised to resolve conflicting views in the human activity system. The qualitative approach strived for quality data (SSM), while the quantitative approach represented the Malaysian doctors community.
4.1 Model / Infrastructure
The model/infrastructure (SSMM) proposed in this project, was defined as below:
(i). Formulate some models of human activity systems that were relevant to the problem situation.
(ii). Use the formulated models in (i) to question the perceptions of the real world in a process of comparison.
(iii). Use the debate so initiated by the comparison to define purposeful action which would improve the original problem situation.
In this infrastructure four perspectives were considered: (i) People perspective studied using attitudinal survey instrument; (ii) Organisation, technology and function perspectives were considered in a feasibility study of the viability of CDSS implementation; (iii) Data collected from interviews with medical practitioners and observations in Malaysian hospitals; and (iv) Organisational theory was used to analyse reasons for the slow adoption and resistance to change.
Holons and surveys used in the project are elaborated on below.
Holons
Holons defined by SSM to represent human activity systems relevant to real purposeful actions, built from some declared perspectives or worldviews [Checkland & Scholes 1990] were used to identify different perspectives of the purposeful activities of CDSS. In SSMM, holons were defined for the adoption situation of CDSS in Malaysia.
Surveys
The study was conducted on three occupational groups in Malaysia: medical practitioners with clinical responsibilities, medical students and hospital/health departmental managers. These groups were considered to be most affected by the CDSS. A survey instrument was developed on the basis of a series of interviews conducted using SSM, adaptation of surveys from literature and pilot
testing. Feelings and attitudes were summarized from the series of interviews. A sample size of 120 was decided on through examining the results from the pilot survey. Some questions were also restructured to improve on clarity after receiving comments. The scales to measure attitudes followed a Likert format. Both face-to-face interviews and e-mail data collection were used, in different parts of Malaysia.
Data were analysed using the statistical package SPSS (Release 7.5, 1996). With the help of a SPSS manual, the reliability of the scale was calculated as Cronbach’s alpha. Data obtained from open-ended questions were categorized into relevant themes for ease of analysis.
4.2 Feasibility study on adoption of clinical decision support system (CDSS) in Malaysia
This section is intended to guide the reader in considering whether it is worthwhile to construct a medical information system with CDSS in Malaysia. The aim is not to provide objective criteria by which the viability of CDSS could be measured from the organisation standpoint, but the emancipatory research strategy adopted entails seeking information requirements of all users in the information system. It seeks to answer the question, “ How feasible is CDSS for use by different people in the information system environment? ”. This discussion considers the technical, legal, organizational, social and economic perspectives as outlined by Avgerou and Cornford (1993). Some details are given in the following subsections.
4.2.1 Measures of performance
It was imperative that user requirements and alternative systems were identified prior to assessing the feasibility of the systems. This project considered two systems, the diagnosis of patients without CDSS currently, and the proposed system with CDSS. In line with the emancipatory research strategy, negotiation of user requirements were undertaken when deriving the primary task model using Soft Systems Methodology (SSM). This was used to define the measures of performance for the feasibility study as advocated by Wilson B. (1984).
The 5 Es
Wilson B. (1984) advocated the use of 3 Es (Efficacy, Efficiency, Effectiveness) to measure the performance of the alternatives for the situation. Checkland and Scholes (1990) added two more criteria for monitoring and controlling the system; ethicality and elegance. Fig. A.1 presents these five criteria used to measure the following five aspects of the feasibility study.
4.2.2 Primary task model
The primary task root definition for CDSS is after a CATWOE analysis shown in Figure A.2. It follows the schema, ‘a system to do X by Y in order to achieve Z’ by Checkland and Scholes (1990) after a synthesis of various Weltanschauung obtained. Two transformation process (X) were adopted, in a relationship where diagnosis accuracy dominates learning. The minimum activities necessary to meet the requirements were assembled and shown in a model in Fig.A.3. Based on the defined infrastructure, models and feasibility study, some findings are presented below.
5. Project Findings
The survey found that generally Malaysian doctors have attitudes ranging from unfavourable to neutral towards CDSS. Many had misconceptions that CDSS would replace them and pose a threat to them and patients (Fig. 5.1). The analysis showed they had low awareness and were willing to be exposed to this technology, providing the possibility of a paradigm shift towards adoption in the future (Fig.5.2/3/4). Doctors intuitively employ complex decision making strategies based on common sense, instead of fixed organisational and medical guidelines. While the doctor would still be able to decide on the course of action to be undertaken, CDSS would present the standard decision approved by the organisation. Among tangible economic benefits of CDSS would be the savings achieved by using less medical resources and time to arrive at a diagnosis, as CDSS produced a tested fixed procedure to arrive at the optimum decision on a range of cases.
Fig. 5.1 Graph comparing attitudes among different occupational groups.
| | |
|  |
|
|
Theme
|
No. (%)
|
|
|
Insufficient medical knowledge of patients
|
12 (55)
|
Absence of clinical evaluation
|
4 (18)
|
Inaccuracy of system
|
2 (9)
|
Others
|
4 (18)
|
Total
|
22 (100)
|
|
|
Fig. 5.4 Reasons CDSS can be detrimental
if networked with patients on the Internet
|
|
The project had shown that there were various interest groups with different needs, and prospects of CDSS adoption were higher when these needs were met.
6. Conclusion
This project by avoiding the study of single variables and instead focus on total situations (the technical, social and organisational context) of information systems yield some useful lessons.
The participative approach is essential for medical information systems development. By involving users be part of the development process of CDSS, they contribute to the functional requirements of the CDSS; such as consultative participation, representative participation and consensus participation.
The danger is that through the use of CDSS, management may pressure to reduce frequency of expensive investigational treatment to force the doctor’s conformity to organisational goals.
This project has highlighted the problems and hurdles facing CDSS. Many new techniques and developments have emerged during the last one and a half years. It is time for researchers to explore the design and usage of flexible web-based decision support system generator utilising software agents to provide additional functionalities.
|
|
Theme
|
No. (%)
|
Clinical trials to improve accuracy
|
13 (28)
|
Information about the programme, pros and cons
|
11 (23)
|
medical workshops and seminars about CDSS
|
8 (17)
|
Prototype to try
|
4 (9)
|
Public awareness (through mass media)
|
4 (9)
|
Clear misconceptions of doctors
(CDSS replaces them)
|
3 (6)
|
User-friendly programme
|
2 (4)
|
Higher computer literacy rate
|
2 (4)
|
Total
|
47 (100)
|
Fig. .5.5 Ways to improve adoption of CDSS
|
|
|
|
7. Acknowledgement
The authors appreciate the cooperation of numerous medical specialists and health personnel who have very kindly been involved in interviews and the survey. The supports of the University of Auckland (NZ) and the HELP Institute (Malaysia), in enabling this project to be conducted are gratefully acknowledged here.
1. Avgerou C. and Cornford T., 1993, Developing Information Systems:Concepts, Issues and Practise, Basingstroke, UK: Macmillan.
2. Checkland P, 1981, Systems Thinking, Systems Practice, Chichester, UK: John Wiley& Sons.
3. Checkland P and Scholes J, 1990, Soft Systems Methodology in Action, Chichester, UK: John Wiley& Sons Ltd.
4. Hunt DL, Haynes RB, Hanna SE, Smith K., "Effects of computer based clinical decision support systems on physician performance and patient outcomes: a systematic review", Journal of the American Medical Assoc. (JAMA), 1998; 280: 1339-1346.
5. Mazoue J. G., 1990, The Journal of Medicine and Philosophy, “Diagnosis without doctors”, Vol. 15, pp. 559-579.
6. McNeill P., 1985, Research Methods, UK : Routledge.
7. Shortliffe E. H., 1987, JAMA, “Computer Programs to Support Clinical Decision Making,” 3 July, Vol. 258, No. 1, pp. 61-66.
8. Sprague, R. H. 1980, “A Framework for the Development of DSS”, MIS Quarterly.
9. Wilson B., 1984, Systems: Concepts, Methodologies and Applications, Chichester, UK : John Wiley& Sons.
(The above announcement is to increase visitor count to this blog site. Bookmark it if you are a parent. This parenting blogsite has a goal to be the top blogsite on the Internet with daily tips from an enthusiastic dad).
![Text Box: Scale and item Mean ± S.D.× Reliability
General computer orientation 3.03 ± 0.96 0.80
CDSS will increase the productivity of hospitals. 2.74 ± 0.86
CDSS creates hassles for clinical staff. = 2.91 ± 0.81
CDSS can do jobs better than people. 3.58 ± 1.01
In the long run, CDSS decrease the hospital’s cost. 2.94 ± 0.93
Use of this technology is unavoidable. 2.65 ± 0.89
I am confident that CDSS will succeed. 2.85 ± 0.78
Frequent CDSS breakdown is anticipated. = 3.92 ± 0.73
Job security 2.20 ± 0.99 0.86
Increased use of CDSS mean less work for people. = 2.64 ± 0.94
CDSS will threaten my future where I work. = 2.13 ± 0.92
CDSS will reduce my job security. = 2.13 ± 0.85
CDSS will make me less useful as a worker. = 2.06 ± 1.00
My job skills are rapidly becoming obsolete. = 2.05 ± 1.11
Professional impact 2.70 ± 0.86 0.71
CDSS will enhance opportunities for medical care. 2.41 ± 0.75
CDSS will free up time for more professional activities. 2.88 ± 0.93
CDSS will upgrade job descriptions. 2.82 ± 0.82
Management concern 3.11 ± 1.00 0.65
Management’s interest is to improve productivity
Without concern for employees. = 3.24 ± 1.09
Management will involve people in
Planning for implementation. 2.74 ± 0.89
When management explain CDSS plans, they
will not tell us the whole story. = 3.47 ± 1.02
Fig. A.4 Attitudes of doctors towards CDSS analysed using SPSS programme
x Scores ranged from 1 to 5; the lower the mean, the more positive the response. S.D=standard deviation.
=Negatively worded items were reverse coded before data analysis.
[Note : Lower mean score represent more favourable attitude towards CDSS. Mean score in Fig..A.4.]](file:///C:/Users/philchlee/Documents/Philip/research/CDSS-SSM/PhilipLeeSSM%20Model%20for%20CDSS_files/image009.gif)