Design Participation as Insurance: Risk- Management and End-users Participation in the Development of Information Systems in Healthcare Organizations Vivian Vimarlund Ph.D., Toomas Timpka M.D., Ph.D. MDA, Departments of Computer Science and Social Medicine, Linköping University, Linköping, Sweden. Address for correspondence Vivian Vimarlund, Ph.D. Department of Computer and Information Science Linköping University S-581 83 Linköping Sweden Phone: +46 13 28 44 73 Telefax: +46 13 14 22 31 E-mail: vivvi@ida.liu.se 1
Abstract By their indirect contribution to the attainment of the organizational goals, activities that prevent negative secondary effects from information system development in healthcare organizations can be regarded as sunk costs. In this paper economic theories and a case study are used to build a theoretical framework for analysis of when decision-makers should use end-user participation as a form of insurance for unforeseen consequences of the implementation of IT systems in healthcare organizations. The result suggests that to invest in broad design participation can be a productive activity that transforms potential current income into future benefits. Keywords: Health economics, health informatics, end-users participation, information system development, risk aversion. 2
1. Introduction Achieving a significant return from invested resources in Information Technology (IT) has been a continuing problem for today s organization [1]. In this respect, an alternative for a health care organization is to be poor today but have good future prospects. When purchasing a large, expensive item such as a Health Information System (HIS) which has not been produced in large quantities previously, there exist risks of cost-plus variety [2]. A HIS consumer agrees beforehand to reimburse the costs related to the acquired goods. However, nothing is usually decided about the unintended consequences that the implementation of a new system can have for the organization. A non-optimally designed HIS can lead to increased workloads and patient waiting time, and thereby reductions in the total service production [3]. In all, a healthcare organization may find the total effect of the changes that occur after the implementation of the new but noneffective system unacceptable. However, if the issues were clearly presented beforehand, decision-makers in healthcare would normally be willing to adopt measures to protect the organization from evident risks incurred by receiving non-optimal new systems. The ability to make effective decisions and buy the appropriate information system is closely related to the costs of acquiring information about the system s functionality and learning about the technical characteristics of the system [4]. Consequently, because the goals of healthcare organizations are to promote and restore health, activities that prevent negative secondary effects from information system development may be included in the total cost structure as sunk costs (a 3
cost incurred "in the past" that will not be affected by any present or future decision), in situations when they significantly contribute to the attainment of these goals. In the Scandinavian countries it has long been a tradition to incorporate end-users as key members of design teams and not see them merely as experimental subjects [5-8]. End-users participation in information system development is instead regarded as an important source of benefits, ranging from increased workplace democracy to that human-computer interaction issues are brought forward in the design process [8-9]. However, despite the increasing interest in end-users participation, few studies have discussed its economic consequences for healthcare. The aim of this paper is to explore the value of end-users participation in information system development projects. In specific, a formal theory is proposed for the planning of the use of such approaches in healthcare organizations. 2. Method Data were collected in a case study of the development of an information system in a small clinical setting. At the time of the study, the clinical unit provided case management for patients with multiple problems of medical, social and psychological nature. It only received its patients by referrals from other institutions, e.g., Primary Healthcare Centers (PCH), Social Insurance Offices, Occupational Healthcare Clinics, the Employment Exchange Office and Social Welfare Units. A participatory design method was used to develop the system. During 4
the initial phase of the development, the future end- users of the system: one physician, one psychologist, one social worker and one secretary, were allowed to actively influence the system design and test every new tool that was considered for implementation. The final HIS supported internal messages to individuals and broadcast file sharing, group and individual scheduling, and Internet mail [10]. Both qualitative and quantitative methods for data collection were used. For a period of two weeks, in face-to-face interviews, the healthcare team and the designers were asked about the system in use and about the main outcomes obtained from the participation process. In order to validate our results, one week after the first interview the members of the staff were individually given the opportunity to comment on the conclusions, which could be drawn from their interviews. Economic theories of industrial organizations, uncertainty, and consumer behavior were then used to interpret the collected data. 3. Case Study: Use of Working-Time in Design Participation 3.1. Findings: In the initial period, the clinical team acquired first-hand experience and information about the physical attributes of the information system. It also learned about the social significance of consuming the system and how the system might serve the needs of their clients. The endusers and designers discussed basic goals for the organization as well as the potential changes for teamwork that the system under design could support. At this stage, the outcome of the design was not seen to be the information system itself, but the value of its role in the production of healthcare 5
services. The new information system was seen as a tool for improving the quality of the service and, as such, a factor that would directly affect the service production possibilities at the clinic. Also, the requirements that were relevant for an improvement of the collaboration between the organizations the patient was involved with were established. Some of the effects pointed out by the work-team were: i) The possibility to improve communication with non-healthcare organizations, ii) The possibility to design the teamwork activities from a patient perspective, rather from that of a single service, iii) The possibility to create network structures that would allow creativity in production and processes, e.g., new forms of interaction and communication with patients and healthcare providers. 3.2. Comments: The information acquired during the design process allowed the clinical team to distinguish the functional value of the HIS, and to understand how it would help to easily integrate the healthcare professionals work tasks with each other. As previously found by Dobbs and Hills [11], experience with the application in the practice setting changed the clinical team s perception of the HIS. This in-depth experience may be the key factor for a high acceptance level of the changes that a new system inevitably is going to lead to for an organization. Moreover, the participation allowed the end-users to work closely with designers in their role as producers of modules, by supplying feedback to practical problems that arose during the design stage. This situation allowed the module producers to present broad specifications for the required module, while engineers developed the detailed specification. A system with these 6
characteristics was consequently considered to be of more value than any other potential systems without such attributes. This is due to the fact that the proposed solutions were seen to be more in accordance with the endusers own preferences [12-13]. The design process was also seen as an investment in human capital that should increase the effectiveness in the use of the systems and thus as a tool to manage the changes that technological innovations give to any organization 1. 4. A theory for Investments in Design Participation When a system development project is initiated, the final quality of the goods (i.e. the information systems) that will be available to the healthcare unit and the subsequent effects on the work-routines cannot be known with certainty. If the information system is not sufficiently appropriate, the economic effects on the organization are significant for two reasons. The first is the irreversibility involved in any investment [14] and the other is that clinical organizations can not raise unlimited resources to compensate for organizational failures. Consequently, when healthcare organization s managers fear negative secondary consequences of a HIS with regard to the clinical work flow, and if they can find a possibility of shifting these risks, they will engage in activities that allow them to be sure that the system will 1 The theory of economic growth signals four determinants of total output: growth in the labour force, investment in human capital, investment in physical capital and technological changes. Lipsey G, Courant PN. Economics.(11 th Edition). New York: Harper Collins College Publishers 1996. 7
be adapted to all aspects of their organization. The underlying logic for investing resources in design participation processes is the same as in choosing between ordinary commodities x and y within a single period of time. Just as individuals, organizations choose the level of investment that maximizes their well being, and borrow or lend resources to achieve a preferred time-pattern of consumption. Once the management has committed funds to a specific project, such as the development of a new HIS, it must forego other investment opportunities that might have been undertaken by using the same funds. Risk averse decision-makers calculate the present discounted value of the effects accrued over time to the unit, and predict the amount of resources they are willing to pay to acquire an insurance (such as design participation) that will protect the organization from future loses [15] e.g., decrease in service production. In such circumstances, the acquired insurance converts an uncertain outcome into a certain cost (the amount of resources the organization pays for allowing its personnel to actively participate in the development process). It follows that the organization is better off with an insurance if the possible future negative effect is severe or likely. The benefit for the buyer of the insurance who exchanges a risk-filled situation for a less risky one, is then substantially higher in terms of gain in private benefits or consumer surplus and wellbeing [16]. Consequently, healthcare decision-makers that want to prevent any negative secondary effects that affect the social or economic outcome of their organizations should be willing to actively contribute out-of pocket investments in 8
participation in order to screen out a high risk of obtaining a less appropriate system. Insert here text box 1 5. Discussion Organizations with utilitarian goals, such as healthcare organizations, always aim to maximize the utility of the production of services over time. They will therefore undertake policies that ensure optimal growth by adjusting a variable input factors, such as workforce, and means of production in the short run, and by choosing the method that will produce the best social and economic output in the long run. The results of our case study suggest that when time and effort are invested in allowing healthcare staff to participate in information system development processes, the benefits can well exceed the costs over the life cycle of the project. Because understanding the current process provides the basis for discussions regarding new alternatives (such as innovation of services, system capabilities, system costs, technical architecture), end-users participation in design can be the key positive influence on the quality of the service and thereby organizational effectiveness. In reverse, improved effectiveness stimulates productivity due to the fact that understanding the current process also provides the basis for decisions regarding new alternatives that can generate growth [17]. It is this association that makes design participation attractive for the healthcare organizations and for society as a whole. The central mechanism behind this is that learning-bydoing increases the accumulation rate of human capital for the organization 9
[18-19]. We further found that effects of design participation not only can induce a better understanding of an organizations needs in the production of the system, as argued by Greenbaum [20], but it can also increase the interpersonal or social interdependence among individuals. The main yield of these effects is a broader scope for productivity improvement [17]. Further, in every economic system, there is a responsible agent upon whom the burden of any given risk falls in the first instance. Healthcare organizations, due to their responsibility as agents for patients, and as representatives for their stakeholders (who are themselves largely risk averse), are always likely to behave in a risk averse manner [21]. In making choices, these organizations always wish to be efficient by using no more input than necessary. Because the development processes of HIS necessarily involve changes that take place over time and involve numerous actors [22] it seems rational that healthcare organizations should adopt approaches that diminish or at least divide risks of obtaining an inappropriate HIS. Consequently, if it is possible to use design methods that will decrease the risk for negative unforeseen secondary consequences, these methods will be preferred, even if this implies increased short-run costs during the requirement specification process. Decision-making, whether or not of an economic nature, always involves assumptions about the future. The purpose of adopting a new plan of action, such as introducing an information system, is to make the future different and presumably better. To make an optimal choice in economic terms, it is essential to predict, with the least possible uncertainty, the effects that this 10
choice will have on the future evolution of the economic unit in question. If any organization worries about risk it will pursue safer lines of activity even though these suggest fewer short-term benefits. The organization has to decide what to consume, how to employ the resources it has, and how much to save so as to acquires more resources [23]. On occasions, organizations save in order to produce new resources for themselves in-house. More often, however, organizations save by physically acquiring new resources (such as new technology), making capital investments in order to create and exploit profit opportunities in the future [14]. As saving increases future income, the decision to consume today or to save, such as to decide between direct investments in new ready-made products or design participation, is a productive activity that transforms potential current income into future income. Hence, if a healthcare organization makes wrong choices in this decision-making, its survival may be threatened, whereas if it makes a shrewd estimate of the future effects of an investment, it may be rewarded with large or at least improved benefits and with new sources of competitive advantage. Acknowledgements The Swedish Board for Technical and Industrial Development NUTEK/VINNOVA supported this work through the MTO program. References [1] Landauer TK. The Trouble with Computers, Usability and Productivity. Cambridge, MA: MIT Press 1995. 11
[2] Arrow KJ. The Economics of Information. Cambridge, MA: The Belknap Press of Harvard University Press 1984. [3] Vimarlund V, Ljunggren M, Timpka T. Implementation of Computer-Based Patient Records in Primary Care: The Societal Health Economic Effects. In: Cimino JJ (Ed.). Proceedings of AMIA Annual Fall Symposium; 1996 Oct 26-30 Washington D.C. Philadelphia: Hanley and Belfus, Inc 1996. p. 503-8. [4] King N, Anderson N. Innovation and Change in Organizations. London: Routledge 1995. [5] Abowd G, Beale D, Dix A, Finley J. Human Computer Interaction. Prentice Hall International 1993. [6] Mundford E, Henshall D. Participative Approach to Computer System Design. London: Associated Business Press 1991. [7] Naschold F, Cole RE, Gustavsen B, van Beinum H. Constructing the New Industrial Society. Stockholm, Sweden: The Swedish Centrum for Working Life 1993. [8] Timpka T et al. Action Design Version 2.0 CEC/AIM DILEMMA Deliverable DO4. Linköping, The MDA- group. Linköping University, 1992. [9] Sjöberg C, Timpka T. Participatory Design of Information Systems in Health Care. Journal of the American Medical Informatic Association 1998; 5:177-83. [10]Hallberg N, Timpka T. Quality Function Deployment (QFD) for 12
Developing Customized Social Medical Service. Scandinavian Journal Social Welfare 1997; 6:292-300. [11]Dobbs M, Hill MB. Technical Innovation and The Demand for Goods. Scottish Journal of Political Economy 1984; 31(2): 147-56. [12]Becker GS. Accounting for Tastes. Cambridge, MA: Harvard University Press 1996. [13]Frey BS. Not just for the Money. Cheltenham, UK: Edward Elgar Publishing 1997. [14]Dixit AK, Pindyck S. The Options Approach to Capital Investments. In: Brown JS (Ed.). Seeing Differently: Insights on Innovations. Boston, MA: A Harvard Business Review Book 1997. [15]Cooter R, Ulen T. Law and Economics. New York: Harper Colins Publishers 1988. [16]Evans R. Strained Mercy. The Economics of Canadian Healthcare. Toronto and Vancouver: Butterworth & Co. Canada Ltd 1984. [17]Yang X. (Eds.). Increasing Returns and Economic Analysis. New York: McMillan Press Ltd 1998.p.3-63. [18]Lucas RE. Making Miracle. Econometrica 1993; 61:251-72 [19]Arrow KJ. The Economic Implications of Learning by Doing. Review of Economic Studies 1962; 29:155-73. [20]Greenbaum J. A Design of One s own: Towards Participatory Design in the United States. In: Schuler D, Namioka A (Eds.). Participatory Design 13
Principles and Practices. New Jersey: Lawrence Erlbaum Associated, Inc 1993. p. 27-37. [21]Hopkins A, Gabbay J, Neuberger J. Role of users of health care in achieving a quality service. Quality in Health Care 1994; 3:203-09. [22]Perez C. Structural Change and the Assimilation of New Technologies in the Economic and Social System. Future 1993; 15(5): 357-75. [23]Hirshleifer J, Glazer A. Price Theory and Applications. (5 th Ed). New Jersey: Prentice Hall Inc 1992. 14
Text box 1 The reasons why decision-makers should be willing to invest in a "participation insurance" during IS development Risk averse decision-makers are always willing to pay for the reduction of risk just as they are willing to pay for more tangible commodities. Assume that any loss that can occur as a consequence of the implementation of a new system (e.g. changes in work routines, decrease in productivity due to a non-optimal use of the new system, lack of specific knowledge for using central functions, non-optimal system design etc) can theoretically be expressed in terms of a monetary equivalent. In such circumstances, when investing in development of an IS, decision-makers contemplate an uncertain future in which a specific negative effect may or may not occur. Probabilistic methods are often used in decision analysis to determine expected costs and benefit as well as to assess the degree of risk in particular projects. In such situations, when estimating the benefits and costs, it is common that decision-makers attempt to obtain the expected or average values of different events that might or not occur and affect the final outcome of the product. The benefits and costs associated with these different events can therefore be estimated and the expected benefits and costs calculated as the sum over all possible events of the resulting benefits and costs multiplied by the probability of occurrence of a particular event, such as, non-desirable organizational and economic effects. m E [ B ] = ( B ) Pr q t t/q q=1 Eq.1 q = 1, m m E [ C ] = ( C ) Pr q t t/q q=1 Eq.2 q= 1, m 15
Where q= 1, m, represents possible events, (B t\ q ) and (C t\ q ) are benefits and costs respectively in period "t" due to the occurrence of "q". Pr q is the probability that "q" occurs, and E [B t ] and E [C t ] are respectively expected benefit and cost in period "t". Hence the expected net benefit in period "t" is given by E[NBt] = E[Bt] -E[Ct] Eq.3 Let now the probability of occurrence of the unforeseen effects after the implementation of the IS be as already indicated "q", and non-occurrence of such effects be "1-q". Assume also that the chances of the negative effect occurring are believed to be 100 q out of 100. If the negative effects do occur, the consequences (non-desirable effects on the production of health services, and consequently unhealth for the patients) can be judged by the decision-makers as equivalent to a loss of N. Moreover, an organization s well being (WB) during the future in question will be related (positively) to the development of an optimal system, which is expected in the absence of negative effects. Thus, the influence of the negative effect N is thus equivalent to a reduction in well being equal to (WB-N). If we now use U (WB) with probability 1-q to express the utility of the organizations well being with an appropriate system, then we can describe the utility of a non-optimal system over the future period as U (WB-N) with probability q. If for instance q=0.5 there is 50% chance of that either WB or WB-N occur. Representing this expected level as E(U) (expected utility), we obtain: E(U)=U(WB) (1-q)+U(WB-N)q Eq.4 The decision-maker, in such circumstances may be confronted with the subject choice among alternatives with different expected benefits of levels of risk and confronted with a choice between the certain value of the IS and a gamble with possible uncertain outcomes. 16
Decision-makers that wish to avoid risk so as to avoid the possibility of losses might in such circumstances, select a project with a lower risk of losses. Consequently, in situations of high risk for obtaining a non-optimal IS, a second best allocation of the capital investment will be preferred (investing in participation). Even if this action affects the total amount of consumption of resources (time, personnel involvement, and or money) that is available in the short-run, the decision-maker will chooses the riskless solution to maximize the expected utility of the IS. This preference results in a willingness to take risk premium e.g. participation insurance, that will ensure that the final benefits outcome will exceeded their cost, reflecting in this manner an adjustment of the rate of returns for the perceived risk 17