Volume 70 October 1977

701

Section of Epidemiology & Community Medicine President A A Adelstein MD

Meeting 10 February 1977

Health Information Systems Dr A S Haro Department of Planning and Evaluation, National Board of Health, Siltasarrenkatu 18a, Helsinki 53, Finland)

Health Information Systems and Users' Needs In spite of the continuous technological development which is characteristic of the present day, there seems to be a general feeling that health technology is not being developed and applied in harmony with the social function of health services. The administration and management of health-related activities have not 'exploded' to the same degree. A summary of world opinion was expressed at the World Health Assembly in 1973 where it was concluded that, both in developed and in less developed countries, it was commonly felt that health services are not fulfilling expectations. Attention should be drawn especially to the recommendation to apply modern scientific management to health administration. The need to improve information was particularly stressed: 'Health service information systems, as they now exist, are often ineffective and inefficient despite the clear need for them ... Health information systems need to be designed once again, starting from basic principles' (WHO 1973).

What is an Information System ? Modern 'systems management' or 'management by systems' (Johnson et al. 1973) focuses attention on the ultimate goals of management as a whole. The element that makes interaction possible is information, understood in its widest possible sense. Information systems are designed to ensure that management has relevant information when making decisions. It is erroneous to think that information systems

depend only on measurements. Management is also based on value judgements, intuitive predictions and subjective experiences. An information system does not mean solely statistical information. In principle a health information system incorporates all possible channels that can convey answers to the questions put by the decision makers. The establishment, maintenance and development of an information system is an important and highly specialized function. It is revealing, in this respect, that there is a tendency to use the expression 'information service' to describe a unit which performs this function (Bodenham & Wellman 1972). Routine statistics, registers and reliable opinions are natural components of the information system, which also includes published information about research and surveys. All have their merits and limitations, and the use which they are expected to serve dictates which are the most suitable (Haro 1972).

What Should be Known? The reliability and usefulness of statistical information are often criticized. It is difficult to find any article about statistics which does not incorporate statements such as: no data should be collected unless they have a known use. Statements of this kind are truisms, but not very helpful ones, because it is rarely stated what 'known use' means in practice. There exist few analyses that show what gaps in information are reflected in the goals of health policy and in what way better information would change the order of priorities. Shortage of useful information is one of the few valid excuses for mismanagement and erroneous decisions. Gaps in knowledge can be bridged by determined action and political will, but the decisions which guide such action will only be as sound as the information upon which they are based. A prerequisite of efficient management is a reasonable

702

Proc. roy. Soc. Med. Volume 70 October 1977

amount of information concerning problems, resources, possibilities of intervention and the environment within which specific actions have to be taken (WHO 1971). Seen through the eyes of responsible authorities the information services should especially cover areas where the degree of uncertainty and the risk of making erroneous decisions are greatest; for example, where the situation is problematical or where a mistake would have severe consequences. Information should be relevant, reliable, understandable and up to date, free of the limitations usually ascribed to statistical services. In subsequent paragraphs information services are discussed from two different standpoints. The content of information, i.e. what should be known, is seen as the basic problem and discussed first. It is easier to decide how the information needed should be collected, processed and communicated. General Strategy Mapping the whole area of what should be known is not a helpful exercise if no advice is given on how to proceed in practice. It is, therefore, necessary to devote some thought to general strategy. In principle, decision-making is a process which converts information into instructions for the management system. The possibility of not acting should always be included. To do nothing means that, on the basis of the available facts, we decide that there is no reason for any action. This is correct if we know that continuation without change is justified. A laissez-faire attitude, however, without being properly informed, may sooner or later lead to critical situations. In this connexion people other than formal administrations should also be counted among the decision-makers. A decision-maker can be anybody who has, or should have, the possibility of forming an opinion concerning the situation and, at least indirectly, of participating in the decision-making process. At the theoretical level it is easy to agree that proper information is essential. But when it is necessary to define what is needed or what should be known, consensus is difficult to achieve. There are many reasons for this. One is that different personalities are to be found among decisionmakers. Active and emotionally motivated persons are not fond of measurements that are not directly helpful. An experienced official does not feel any temptation to collect data that he has not needed in the past, but a manager without much experience behaves in a different way; general curiosity may well describe his attitude and there are those who compensate for their limitations by overdoing research and the compilation of data. Another source of difficulty stems from the differences between the optimum requirements of an information service and those of the organi-

zation as a whole. Often one sector is undeniably efficient but consumes resources which, if used differently, could markedly improve the overall achievement. A well-known example is the situation in which, owing to sophisticated curative services, too little is left for the preventive activities which are more important to the effectiveness of the health activities as a whole. This is reflected in the respective resources devoted to the gathering of information (Haro 1976). An unavoidable difficulty relates to the fact that the life of organizations fluctuates. During quiet periods, without major managerial problems, information needs are minimal. But critical situations occur in which organizational survival is directly related to proper information. In such situations measurements are needed of aspects of the organization that are hardly relevant during quiet periods. Traditionally, decisions concerning investment in information services require formal ratification by senior administrators. This is correct, but most such administrators have qualifications other than competence in the use of statistics, planning, research and modern methods of management. In practice this often leads to another problem: a relative negligence, which is reflected in the isolation of the information services from the real activities of the organization and from responsible management. This attitude reflects past circumstances, when time was not a scarce resource and attitudes were against change. Of course there are areas where information services can be properly organized without close cooperation with the decision-makers; but numerous information needs cannot be known and properly met without continuous cooperation between the information services and experts who know the problems. Models To analyse and to describe in detail decisions as they are made in real life is impossible, but much can be achieved if the more important components are selected and presented symbolically as a model. The construction of a model has much to recommend it. Interrelationships can be seen more clearly; implicit values and assumptions have to be made explicit, which means that they can be critically analysed; and a model indicates what information is vital and what the consequences are of using estimates instead of measurements. The expression 'mental model' is used to denote a model that is multidisciplinary and exists only in the imagination of decision-makers. Evaluative thinking relies partly on experience, partly on facts, but to a great extent on intuition. In many instances the model is merely a set of ideas of what might happen if something is done or not done. However, such mental models are realities in

Section ofEpidemiology & Community Medicine relation to decisions- the more important these are, the more the decision is based on abstractions. On the other hand, 'formal models' are constructive elements in mental models and may influence and improve the understanding of the reality. It can be said that health-policy decisions are compromises between interest groups or their representatives, who have differently composed mental models. The previous discussion on systems, information and models leads to the key issue of this paper. One can postulate that an intentionally collected item of information primarily serves one purpose: to be used in some model which makes evaluative considerations possible. Only exceptionally will this be a formal model to which quantitiative measurements can be fitted. In most instances the basic models are mental. Models are in this case not ends in themselves but only methods to help in the real problem: how to establish the information required. The focus of interest is not the formulation of decisions but the choosing of the items of information or indicators needed for forming opinions concerning the problem. For this limited purpose relatively crude models are serviceable. But even such models can be constructed only by persons who know the problem. In most cases this requires teamwork from experts. Valid models cannot be made by someone who is unable to go deeply enough, for example, an expert in statistical methodology alone. Programmes and projects can be analysed by experts who have in mind alternative actions. Planning is usually based on some model which indicates the problems and, accordingly, the information needed. Managerial decisions can be simulated in order to find out what information seems to be relevant. Politicians, decision-makers, administrators, experts and planners are together responsible for devising models. Statistical and data-processing experts are, of course, natural participants but their responsibilities start when the previous groups have reached a consensus on what needs to be known. To provide needed measurements for all possible models is not feasible and technical feedback often influences the modelling activities. But from the technical point of view the difficulties are only relative and can be solved. It is customary to speak about planning in relation to programmes or activities, but in principle planning can also be planned. This is easily understood if the planning process is divided into its natural components. There are two main types of planning. We can plan having something new in mind; examples of this can be seen in the field of building construction and in all health programmes that aim to solve a defined problem. Another objective of planning is to improve or

703

control something that already exists. These two types of planning are based on different models and require different kinds of information. The 'goal-attainment model' is interested in effectiveness, which can be assessed, for example, on the basis of epidemiological measurements. For many practical management purposes, a 'control model' which focuses attention on efficiency, on the correct use of resources, is at least as relevant. Finance, personnel, services produced and other measurements, which may be uninteresting to an epidemiologist, are valid indicators in this model, which describes the activity as a continuous service to be controlled. In real life both these aspects are usually influential but to a various degree. The control model has a tendency to refer to a static pattern of activity which is uninfluenced by changes in resources, techniques, manpower, or costs. A more goal-oriented way of planning is typical of a society or organization which is aiming to introduce innovations. Content and Method It is not meaningful to discuss the content of an information system without considering the methodological aspects of data collection. An information system that is intended to serve all possible planning activities and decision making of an organizational unit has an obvious tendency to be based on individualized data (ID) systems (databanks, registers &c.) in which information concerning cases, events, persons &c. is stored. The basic idea is to make possible the processing of the data in such a way that they can be regrouped to fit any tabulation or representation required. In practice most ID systems rely on modern technology and this gives some freedom for the planning of the information content. Variables, which are difficult to handle in traditional forms of report, can be included and to some extent the work of the informants can be made less exacting. The central element of a routine statistical system is a questionnaire or form which summarizes the work done by the informants, normally over a definite period of time. The national figures are compilations of the summaries at lower levels. Such an aggregated data (AD) system has some advantages, for example, the direct costs of the system are relatively small, and it does not require much technical expertise. Relatively long timeseries are usually available, but the system is rigid and cannot give answers to questions formulated differently. A great deal of information is lost when the data are aggregated. This limits the value of the AD system, especially in planning and scientific management. There are sectors in which such a system serves quite well; in other fields, the value of statistics is limited. The provision of information, for manage-

704

Proc. roy. Soc. Med. Volume 70 October 1977

ment and for more general descriptive purposes, requires the integration or linkage of data. This can be achieved by systematically applying common classifications and uniform definitions in all relevant subsystems. In theory such requirements can easily be fulfilled, but in practice only limited combinations are feasible. Codes that attempt to satisfy the specific needs of different sections of society escalate towards impracticability. Personal linkage is not always favoured by the public and the systematic use of personal identification is possible only in advanced and well-organized information systems (Nordic Medical Statistics Commission (NOMESCO) 1973). In Finland the use of derived personal identification numbers is widespread and covers many different aspects of life. In the health field we have been relatively liberal in this respect and have stressed more the possible benefits to the care of illnesses and their prevention than theoretical misuses for criminal purposes. In a regionalized health service system it is relatively easy to have information about a person's health. Hospital medical records are not kept in bank vaults and the address is known to everybody. We have tried to solve these problems more informally by ethical codes rather than by strict legislation and inspection systems as in some of our neighbour countries. A more reasonable way to obtain integrated information is by means of specially planned sample surveys. These make it possible to cross sectional boundaries and to link different types of information to an individual respondent or family; for example, health and factors in the past and present environment, both physical and social. A single survey enables an analysis to be made of present circumstances but trends are often more important. A natural solution to this limitation can be obtained by repeated surveys. The problems dictate whether the same sample or another comparable sample of basic informants or objects of measurements has to be studied. A before-andafter strategy is well suited to situations where new legislation, great innovation, a marked increase in resources or some other unique event can serve as a dividing line (Purola et al. 1974). Conclusion Large and complicated problems cannot be handled without trying to form a bird's-eye view of the situation. Problems can be seen differently but some generalizations are justified. One of these is the fact that statistical services for health purposes have not developed as fast as could have been expected, bearing in mind the central position of health among social concerns as well as in national figures of expenditure. The most recent WHO conference of National Committees on Vital and Health Statistics (WHO 1974) reported: 'it would

be necessary in most countries to add substantially to existing systems and also to modify them to make them more responsive to the working needs of users'. It is also well known that national and international cooperative activities, aiming to produce an accepted set of health indicators and social indicators, have not in general been a great success and that slow progress is typical (Land 1975). Very much the same can be said about the UN projects for developing statistical systems like sociodemographic statistics (SSDS) and environmental statistics (SES) (UN 1975). It is reasonable to ask what is wrong. There are numerous explanations, such as attitudes, economic burdens and education, but it can be postulated that the main reason is too little attention to the conceptual problems. The goal should be a system which really serves and produces purposeful information. A great deal of valuable information concerning health and health problems is derived from the records of physicians and other experts giving services to individual patients. It has been said that a physician is a 'man whose profession means making decisions' (Bernard 1976). The first stage of his work, which for him is crucial, is the collection of data, i.e. signs and symptoms derived from consultation and from clinical, radiological and biological tests, and the comparison of the results with accepted standards or norms. In principle there should be no basic difference in the attitudes in relation to the role of information between those who make decisions concerning an individual and those who make decisions at the collective level. In practice this is seldom true. Most physicians have not been trained in the art of making administrative decisions. On the other hand, the decision-makers do not always properly motivate the basic informants when requiring data for collective purposes. If this is not done, it is difficult for the latter to understand why a considerable part of the brief patient contact should be devoted to nonessential duties. Much statistical information directly aids the delivery of medical care to individual patients and indirectly modifies clinical decisions. Attitudes and decisions can explain a lot but it is hardly reasonable simply to wait until these are changed. A more active policy is justified. I think that more attention to methods which initiate a closer dialogue between users and producers of information is in principle the correct solution to our problem. As a first step we can recommend a mental model, a relatively soft and rounded statement concerning goals and policies. Decisions may need a consensus of action but for planning, and especially for information-service planning, much can be achieved if the most typical and relevant interrelationships and basic values are made ex-

Section ofEpidemiology & Community Medicine

plicit. One of the positive aspects of a formally accepted plan is the fact that it introduces a standardized way of thinking about the goals of the organization. A dialogue needs contributions from all participants. The decision-makers are primarily responsible for the content - what should be known - other experts for methods and techniques. In order to participate in the dialogue, the health statistician and other experts should learn to understand the problems of decision-makers. The information produced should naturally be reliable and relevant but these terms are not, in this connexion, identical with scientific exactness and scientific value. Measurements at the right time, even if relatively crude but in principle correct, are often more helpful than exact results available later. Discussion of the optimal organizational structure is one where very different opinions can be defended. Introducing the concept of 'system' to information services can be helpful here. The goal is not a unit, office or department but a functionally coordinated system which fulfils the expectations of decision-making bodies. An information system reflects the basic principles of a management system and the two cannot be discussed in isolation. Requirements set upon the quality, type and timeliness, with participation or dialogue, dictate to a marked extent the structure and location of information services in the organizational hierarchy. In each organization or society decision making is a specific process, and planning is accordingly different. A planning process, which is shown to be suited to one society, cannot serve as well in another, except when the decision-making procedure as a whole is modified accordingly. The same is in principle true of information services, but there are numerous principles which are valid in all circumstances (WHO 1974). Evalution in the health field is generally difficult and in most instances based on subjective impressions. This is of courses true in relation to the value of information. My personal impression is that it is difficult to find any activity in the field of health where a relatively small input would give such a valuable output as in the planned development of information services. REFERENCES Bernard J (1976) In: Decision Making and Medical Care: Can Information Science Help? Ed F T de Dombal & F Gremy. North Holland, Amsterdam; p 3 Bodenham K E & Wellman F (1972) Foundations for Health Service Management. OUP, London Harb A S (1972) Methods of Information in Medicine 11, 1 (1976) In: Cross-National Sociomedical Research: Concepts, Methods, Practice. Ed M Pflanz & E Schach. Thieme, Stuttgart; p49

705

Johnson R A, Kast F E & Rosenzweig J E (1973) The Theory and Management of Systems. 3rd edn. McGraw-Hill Kogakusha, Japan Land K C (1975) International Social Sciences Journal 27, 7 Nordic Medical Statistics Commission (1973) Planning Information Services for Health Administration, Decision Simulation Approach. Mimeo, Stockholm Purola T, Kalimo E & Nyman K (1974) Health Services Use and Health Status Under National Sickness Insurance. A: I 1/1974. Social Insurance Institution, Helsinki World Health Organization (1971) WHO Technical Reports Series No. 472 (1973) Report of the Technical Discussions at the Twenty-Sixth World Health Assembly, Geneva. A 26 (1974) WHO Technical Reports Series No. 559 United Nations (1975) Towards a System of Social and Demographic Statistics (ST/ESA/STAT/SER F/18). United Nations, New York

Dr M A Heasman (Information Services Division, Common Services Agency, Scottish Health Service, Edinburgh) Health Information Services: a View from Scotland

Health information services are first and foremost a service - if they are not that then they are nothing; and in considering their structure and purpose this has to be kept in the forefront of the mind. Their task is to provide those with an interest in the problems of morbidity, mortality and delivery of health services with the data they require, whether this be for planning, administration, management, epidemiological or socioeconomic research, for education, political interest or just sheer curiosity. There may be different priorities on these various uses of health information, and these priorities may vary from time to time, but all have to be met. There is now a move, particularly at an international level, to extend the definition of a health information system to include all types of information used in health service management, but this paper considers only the statistical and computing services as suggested by Knox et al. (1972) and by Bodenham & Wellman (1972). The paper also draws upon my experience in Scotland. I shall first discuss some matters of general concern before going on to consider the information services as organized in that country. Routine Statistics and the Needfor Flexibility Despite the modern tendency to dispute the accuracy and usefulness of routinely collected official statistics, they form the cornerstone of health information services as they have done since the time of Farr. Vital statistics, data on hospital inpatient treatment, cancer registration and nu-

Health information systems and users' needs.

Volume 70 October 1977 701 Section of Epidemiology & Community Medicine President A A Adelstein MD Meeting 10 February 1977 Health Information Sys...
878KB Sizes 0 Downloads 0 Views