The perception of system and the reduction of uncertainty

ABSTRACT

Until the mid 1960's we had two separate streams of development of medical information systems. One proceeded from efforts under the Hill-Burton Program to develop management systems for hospitals and community health planning, with little stress placed initially on the interactive use of computers. The other, centered within medicine, and medical decision making, was from the outset an effort to exploit the growing potential of computers in clinical matters. Creation of the National Center for Health Services Research in 1968, was a key event in the merging of these two sets of activities, with formation of a program and study section in health systems development and a common source of funding for projects with either administrative or clinical significance, or both. The establishment of the National Center was proceeded by more than a decade of health services research with the objective of improved management systems. The concepts and systems growing out of this research were forerunners of many of the activities in modern medical informatics; they were foundations for the still developing techniques of resource allocation, staffing, scheduling, and for screening and diagnosis. A common theme runs through the early research and development. As a researcher one was faced with perplexing pictures of the behavior of both patients and the people and organizations serving them. Through direct observation of the central actions in patient care and support services, it was possible to perceive some pattern, order, or system in the demands stemming from illness and accident, and these patterns were inherently variable and apparently unpredictable. Whether physician, nurse or administrator, the provider was confronted with the need to make decisions under uncertainty. Having gained insight from research, the task was then one of finding ways to reduce uncertainty: by developing timely predictors of demands, by developing flexible administrative mechanisms to move resources to demands (or vice versa) and by reducing error in the assessment of patient condition. The early work in patient classification by level of dependency led to management systems for nurse staffing and patient placement. Refined by the inclusion of diagnosis as a variable patient classification led to mechanisms of resource allocation and reimbursement. In study of outpatient services, early work drew upon techniques of computer simulation to test alternative approaches to control of congestion and delay in outpatient services through computerized appointment scheduling, and records keeping systems. In applying statistical decision theory to disease screening early work led to awareness of the need to reduce uncertainty of testing, touching off a wave of efforts to base decision rules on multivariate analysis to increase sensitivity and specificity. The path to implementation of all these systems has led appropriately to computer based information systems. In retrospect, one can see in the decades of research and development a process of inexorable growth of intelligence in an evolving health system, growth in precision of sensing, analysis, communication and memory. In this perspective, the disparate research and development efforts begun years ago find meaning in the context of medical informatics, and the role of medical informatics itself can be better understood in the context of the evolving health system.

REFERENCES

• 1: Bailey, N.T.J.: A study of queues and appointment systems in hospital outpatient departments, with special reference to waiting times. J. Roy. Stat. Soc. 14, 185-199 (1952).
• 2: Coffey, R.M.: How a Medical Information System Affects Hospital Costs: The ~1 Camino Hospital Experience, NCHSR Research Summary Series, DHEW Publication No. (PHS) 80- 3265, March 1980.
• 3: Connor, R.J., Flagle, C.D., Hsieh, R.K.C., Preston, R.A., Singer, S.:Effective use of nursing resources: A research report. Hospitals 35, 30-39 (1961).
• 4: Dunn, P.F. , Flagle, C.D., Hicks, P.A.: The QUEUIAC: An electromechanical analog for the simulation of waiting line problems, Operations Research Vol 4, No. 6 December 1958.
• 5: Fetter, R.B., et al.: Case-mix definition by diagnosis-related groups. Med. Care 18, 1, 1980.
• 6: Flagle, C.D.: The Problem of organization for hospital inpatient care, in Management Sciences: Models and Techniques. C. W. Churchman and M. Verhulst (Eds.), Vol.2, pp. 275- 87. New York, Pergamon Press, 1960.
• 7: Flagle, C.D.: Allocation of medical and associated resources to the control of leprosy, (With M. F. Le ch at ) . Acta Hospitals 2, June, 1962; also published as Chapter 9 in Management Sciences in the Emerging Countries, N. Barich and H. Verhulst (Eds.). Oxford, Pergamon Press, 1965.
• 8: Flagle, C.D.: Statistical decision theory and the selection of diagnostic and therapeutic procedures in public health, (With M. Lechat), in Acres de la 3ieme Conference Internationale de Recherche Operationnel_le, G. Kreweras and G. Morlat (Eds.), pp. 194-202. London, English Universities Press, Ltd., 1964.
• 9: Flagle, C.D.: A decision theoretical comparison of three methods of screening for a single disease, in Proceedinqls 0 f the F i fth B e r k - elev Symposium on Mathematical Statisticsand Probability. L. LeCam and J. Neyman (Eds.)., Vol. 4, pp. 887-901. Berkeley: University of California Press, 1967.
• 10: Flagle, C.D.: Communication and control in comprehensive health planning. Ann. N.Y. Acad. sci. 161, 714-729 (1969).
• 11: Flagle, C.D.: The role of simulation in the health services. American Journal of Public Healt____~h, 60:2386-94, 1970.
• 12: Gabrielson, I.W., Soriano, A., Taylor, H.H., Flagle, C.D.: Analysis of congestion in an outpatient clinic. Final report on USPHS grant W-96. Baltimore: School of Hygiene and Public Health, The Johns Hopkins University 1962.
• 13: Gleser, M.A., Collen, M.F.: Towards automated medicaldecisions. Comput. Biomed. Res. 5, 180-189 (1972) .
• 14: Haldeman, J.C., Abdellah, F.G.: Concepts of progressive patient care. Hospitals 33: May 16 and June 1, 1959.
• 15: Lincoln, T.L., Parker, R.D.: M e d i c a 1 diagnosis using Bayes theorem. Hlth Serv. Res. 2, 34-45 (1967).
• 16: Miller, H.E., Pierce, F.A.: The implementation of nurse scheduling using mathematical programming. Proceedings, Forum on Nurse Staffing, National Cooperative Services Center for Hospital Management Engineering, New York, September 8-9, 1975. Nuffield Provincial Hospitals Trust: Studies in the function and design of hospitals. London: Oxford University Press 1955.
• 17: Thompson, J.D., Fetter, R. B., McIntosh, C. $., Pelletier,R. J.: Computer simulation of the activity in a maternity suite. In: Proceedings of the third international conference on operational research, Oslo, 1963. aris: Dunod 1964.

INDEX TERMS

Primary Classification:
  J. Computer Applications
  J.3 LIFE AND MEDICAL SCIENCES
      Subjects: Medical information systems

Additional Classification:
  H. Information Systems
  H.1 MODELS AND PRINCIPLES

  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.1 Applications and Expert Systems
          Subjects: Medicine and science
      I.2.3 Deduction and Theorem Proving
          Subjects: Uncertainty, "fuzzy," and probabilistic reasoning

  K. Computing Milieux
  K.2 HISTORY OF COMPUTING
      Subjects: Systems

General Terms:
Experimentation, Human Factors, Management