14 - HOSPITALS, ON-LINE SYSTEMS
by BERNARD A HODSON
Shortly after arriving at the University I was approached by a Mr. Tony Quaglia, the Administrator of a local hospital called the Victoria General Hospital. The hospital was very old and the decision had been made to build a new 200 bed hospital at a location not far from the University. He asked if I could review their plans and see if there might be some opportunity to introduce some computer operations when the hospital was built. To determine this I did a broad brush study of the hospital functions and wrote a report indicating that a considerable amount of automation was possible, outlining the activities that could be conducted. To my great surprise this document was requested around the world and we finished printing between 2000 and 3000 copies for distribution. It also triggered a number of requests around the world to present my views on hospital automation, taking me to several countries in Europe as well as throughout North America. The hospital decided to proceed with my ideas and funding was sought for the project.
When the project was first put forward for Federal Government support a violent "lobby" was established by certain groups in Winnipeg, in a de1iberate attempt to prevent the Victoria General Hospital getting the project (as it was not part of the "medical establishment" there). I am pleased to say that the Federal Government people saw the merit of the project and ignored the lobby. At one point one of the group went so far as to threaten harm to my career development at the University if I proceeded with the project.
After the project was approved this hostility continued and the Faculty of Medicine at the University of Manitoba turned down my request that they participate in the project. I felt from conversations I had with several people at the time, and since, that their refusal was essentially "political". In spite of thls a number of doctors ignored the politics and did, in fact, assist me with the project. The grant support I received from the Federal Government in Ottawa was used primarily to fund a project team who developed the detailed systems based on the generalised systems and software approach I had developed as part of my teaching and research activity, and through a considerable amount of personal time that I devoted to the project outside of my University work.
I used the software concepts in my graduate teaching program and several students developed simple, non hospital systems with the software. The same happened when I used the software at the Banff School of Advanced Management, and with Roman Catholic management organisations in the USA. In these latter cases the groups developed non hospital systems which operated from terminals attached in one case to my University computer, in the second case to a University computer in St. Lou1s.
By 1967 a major part of the systems and software research had been completed and it was a case of developing the considerable detail that remained to get the system into an operational status. We were plagued with the usual development problems common to projects of this magnitude but what hurt us most was the inability of IBM to supply us with satisfactory terminals, and software to run them. We obtained the terminals in operational condition only after appealing to the US president but the terminals supplied were inoperative for close to a year.
Because of this we had to develop our own terminal handling software, delaying overall software development by several months.
On the nursing side the project team consisted of three research oriented nurses, Kathy Goos (later Kathy McLaughlin), Margaret McCrady and Grace Alexander. Tony Quaglia and his assistant gave us input on the administrative side, and Dr. Bill Alexander (later Dean of Pharmacy at the University of Toronto) worked on the pharmacy side, also doing some of the early work on mixed drug interactions. Computer programs were developed by Ed Tretiak, who later became a lawyer, Don Costin, and some help from graduate students. On the medical side input was received from the Rosenfeld brothers, on staff at the Victoria General Hospital, and from three or four doctors on the Faculty of Medicine. In addition Dr Sukenik, from the Harvard Medical School and the Massachusetts General Hospital, flew to Winnipeg from Boston once a month to review our progress and give us advice. He was working in Boston on some of the early automation activity with Dr. Octo Barnett, a pioneer in the area. Kathy Goos and I flew to Boston to discuss our project with them, and also to review their approach. The Mass General people told us they envied our project because we were able to make decisions simply and quickly. They said that any decisions in Boston had almost to get the approval of the janitor, so bad was the political situation in the hospital. At the time of the visit Winnipeg had a major snowstorm and we returned home to walk through six feet high tunnels stretching from the roadway to the houses.
In the second year of the project I was invited to participate in a Federal Government study group, on what to do with scientific information in Canada. On the original approach to me it was indicated that only a small amount of time would be required. This was discussed with the University and it was felt by them that I should participate in the Federal study. It did not seem at the time that participation would detract from the detailed development phase of the hospital project. As it turned out this was a mistake. From the beginning of 1968, and on, the Federal study took up a considerable amount of time and I had to leave much of the detailed systems development work in the hands of project staff, a colleague giving some assistance from time to time, with minimal supervision by me.
I continued to do the detail software development because there was no other person capable of its development at that time. We had early on recognised the desirability of obtaining a full time project director but had been singularly unsuccessful in our attempt. However, it was felt that the hospital should assume the responsib1lity for writing the project report necessary to support a renewal of Federal support. I reviewed the report a day or two before its presentation to the Federal review committee and made the remark at the time that I was amazed at some of the technical detail that they had included in the report, but did not pursue the matter at that time.
As it turned out my "amazement" was well founded. Apparently someone had obtained a copy of a report prepared by Lockheed Missiles and Aerospace for the famous Mayo Clinic in Rochester, Minnesota, (even to this day I have not had an opportunity to see and review the Lockheed report) and they had included several pages of the report (without asking prior perm1ssion) in the Victoria General Hospital subm1ss1on to Ottawa. On the review team from Ottawa was a consultant from the US, a person who had been involved in the writing of the Mayo report. While I question the inclusion of this person in the review team (because of a conflict of interest) the decision by the Federal Government to curtail funding of the project was the only dec1sion that they could make, and I fully concurred with their decision. Ironically I had a phone call later from one of the US consultants colleagues to say that my systems and software concepts were quite sound and they were interested in what arrangements could be made to use the software as a base for a medical 1nformat1on system network being considered in California.
After leaving the University I rewrote the software completely for a different computer, incorporating a number of additional features. I also continued with some development work in the systems area, particularly related to admitting and discharge functions, and to patient accounting, which had not been a significant part of the Victoria Hospital system.
At a demonstrat1on of this updated system to the American Hospital Association Convention in Houston (which was supported in part by new Federal Government funds) the system was seen and reviewed by a considerable number of people, including senior people from medical groups in Canada as well as by such groups as Lockheed, IBM, Burroughs etc. The response to the demonstration was most favourable, the system itself being in advance of anything else shown at the convention. Dr. Goyette, (of the Federal Department of National Health and Welfare) reviewed the demonstration, (which was offered bilingually) and indicated his pleasure at seeing this development taking place in Canada.
Although the project had its share of problems it achieved considerable suocess. Several prominent medical groups reviewed it in detail. As special guest for lunch with the Governor of King's College Hospital in London, England, the Governor told me that his reviewers were amazed that "so much had been accomplished in so short a time with so few people and so little money" (it was of course due to my generalised concepts, later to become GENETIX). Similar sentiments were expressed by groups in France, Holland, the USA and several other countries. It confirmed that while my ideas on generalised programs were radical, they nevertheless were feasible and did in fact work in the real world.
In particular the software concepts were very closely scrutinised by experts in the field and almost always with favourable resulting comments. Computer manufacturers reviewed it very carefully and indicated they would like to use the software. Several major companies reviewed it and gave very favourable comments. One company said that they had m a world wide survey of software for time-sharing networks and that my product was the only one coming close to a reasonable system for practical operations. Similar favourable sentiments were received from many other groups.
Since then, of course, I have developed additional systems and software concepts which makes the work done at that time seem rather trivial in comparison. Dr.C H Walton, the Chairman of the Executive Council of a major Winnipeg based Clinic said at the time, referring to the hospital project, "We are fortunate in Manitoba in having one of the most advanced of such centres in Canada and it has developed a continent-wide reputation as a leader in its field".
Later we did a "broad brush" study of the Clinic automation needs, as well as giving a series of seminars to their medical staff. To refresh your mind, in the broad brush study approach I developed we attempt to see the whole picture and analyse on a broad basis everything that is happening in respect of information flow , decision flow, and materials movement. Up to this time most "Organisation and Methods" people tended to tackle only one or two operations in an environment, (and regrettably this still is the tendency). Advanced systems studies, however, cover the total field of the environment and this is what we mean by the term "Broad" in "Broad Brush Study".
In such a study we try to determine the general picture - the information interchange of each particular section and how each section inter-relates with each other. This was done extensively at the Victoria General Hospital and we were probably at least two years ahead of anyone on the North American Continent at that time because of it, a view confirmed by visitors from, and visits to various hospitals and installations in North America and Europe.
Most hospital systems up to this time (there weren't many in actuality) had concentrated on the rather mundane patient billing and insurance activity. The one I did, as a result of the Broad Brush Study, linked the physician with the Labs, X-ray, Pharmacy, and other patient oriented activity, all tied together through the physician's order. This was the key to our success, based on the starting point of the doctor's order for the patient.
With the assistance of the Provincial Government we made application for development funds to the Federal government in Ottawa, who were good enough to grant us development money through the Department of National Health. At this point there has been approximately 30 to 40 million dollars spent across the continent on programs that had gone nowhere, one reason being they were not geared to development, $1,000,000 on one project alone, that got nowhere. By today's standards these numbers may seem trivial.
In some ways our project was bringing the mystique of medicine and the mystique of computer science together - like attracting poles. As our project and the one developed by Lockheed for the Mayo Clinic were very similar I will describe what we did in more detail. In fact, when our project was mentioned in the Press Lockheed phoned me and asked if our project was for real (so many projects at that time were pie in the sky theoretical nonsense). Lockheed visited me on three occasions and I visited them in Sunnyvale, California, and we exchanged views. It was very interesting to find that we had both gone through almost the same thinking process and arrived at virtually the same conclusions. Both our approaches concentrated on the doctors' order. As this was a very early "on-line" system I will give it in some detail.
The system was designed to use a large scale IBM 360/65 computer on a real time basis. This in fact means that the computer was to be used to monitor the hospital's functions twenty-four hours a day, seven days a week. It did not mean, however, that the computer would be in actual use for the hospital functions every minute of the twenty-four hours. If such were the case, no 250 bed hospital (the size of the hospital being piloted) could have afforded the system. It was estimated that less than one hour per day of actual computer time would in fact be used, and that a good portion of this could be used on the night shift. Had a small computer been used, it would have had to be used by the hospital alone or in conjunction with one or two other hospitals, but the unit cost of operation would have been considerably higher.
The objective of the system was to capture all data once only, as follows:
- Patient admitting data.
- The recording of Doctor's Orders and the inclusion of these in the patient's record, together with the necessary operations to transfer these Doctor's Orders Into action.
- The capture of the results of laboratory tests, of X-rays and of other services, including them in the patient record.
- The ancillary capture of data needed to run the hospital but which generally is not captured in the patient's record. This includes such data as personnel time records, personnel data, the receipt of goods from Purchase Orders, and similar items.
At each Nursing Station there was a computer monitor, as there were also at various other locations throughout the hospital. These monitors were to be used by Ward Clerks, Nurses and Doctors, after suitable identification, for communicating with the various departments of the hospital and in particular, to have access to the patients' medical records. A typical sequence of activity is now presented. After suitable identification of the user a base screen would be presented on the monitor such as:
- DOCTORS ORDERS
- PATIENT HISTORY
- PHYSICAL EXAM
- LAB REPORT
- ADMIT PATIENT
If a 1 was selected the system would check that a nurse was at the terminal (wishing to report on a patient) and present a screen showing categories of doctors' order, such as:
- ALIMENTARY SYSTEM
- CARDIOVASCULAR SYSTEM
- EMOTIONAL STATUS
- GENERAL SIGNS AND SYMPTOMS
- GENITO-URINARY SYSTEM
- INTRA VENOUS THERAPY
- LOCOMOTOR SYSTEM
- NERVOUS SYSTEM
- PT. HAVING SURGERY
If the nurse now selected 2 she/he would indicate they wished to report on the cardiovascular system for a patient, and the following would appear:
- PULSE AND HEART RATE
- LYMPH NODES
Selection of 3 by the nurse might then lead to presentation of a display on bleeding, and a selection from that screen would complete the nurses input. At some point, usually at the beginning of the sequence, the patient would have been identified.
The computer software, in the meantime, has been keeping track of the sequence selections, and a summary of the selections made is presented on the screen for the nurse to verify. If verified then the transaction is processed. In the above "nurses note" sequence the nurses observation would be placed on the patient record. Other types of sequence might involve preparation of a prescription by the pharmacy, a request for an X-ray or physiotherapy, these latter requests only being allowed to a physician.
The hospital activity was broken down to a sequence of screens similar to those just described, and allowed any type of transaction for the patient to be handled quite quickly, once the staff had become familiar with the process.
It soon became evident that the structure developed for use by doctors and nurses, and the technique for handling the structure, were useful in other areas. The hospital system had shown its usefulness in handling supply orders, doctors' orders, nurses' notes on patient progress, medication orders, admitting and discharge functions.
This meant that it could now be considered for a variety of other applications outside the hospital, and was further proof that the Hodson-Turing concept of a single piece of software to handle all applications was quite practical. Systems work in a school division showed its potential usefulness in an educational system, and it was also shown very useful for teaching the concepts of information systems at all levels.
As the displays are being presented and choices being made, certain checks were being carried out by the computer. For instance certain responses must be followed by certain other responses (e.g. a drug order must have a route (oral, intravenous) indication, treatment must have an associated time). Any contravention of these rules caused an error message to be presented. Similarly any data that was manually entered (e.g. a lab test result, a patient name on admission) would have data validity checks. Such things as alphabetic checks, range checks and conditional checks are carried out on all data before acceptance. This was accomplished by associating with each monitor display an internally stored edit table. Also internally stored was a destination table which was used when the transaction was accepted as correct. This served to transfer the transaction message to another monitor (e.g. a supply order to Central Supply, a drug order to the Pharmacy) and to transfer a copy as appropriate to patient file.