Medical Blog

5. DISCUSSION AND CONCLUSIONS: THE FEASIBILITY OF CURRENT
APPLICATIONS AND MEDICAL CARE UNIT


5.1 The application care unit

In a recent presentation, R. Baud emphasizes the advantages of logic programming as one key bases of rapid prototyping applications so-called care unit: a single high level language Abstraction will ultimately navigate easily between alternatives, before settling.
Experience with Prolog is quite conclusive in that regard, if one knows how to tap into a language procedural as C when needed for specific tasks. The portability of this language is satisfactory. Its effectiveness is sufficient if we take some precautions.
The paradigm of object orientation has the advantage of allowing a division of information and functions into distinct units called objects. Such a quality is needed as soon as the application size becomes important. It is even a precondition in complex situations. It must be remembered that many promising small expert systems have never been able to grow and be useful for this reason mainly. The programming logic has not yet acquired its standards in engineering software. It is argued here that the object paradigm is an element of the solution to this problem. The basic characteristics are:

• notion of object class
• notion of methods attached to objects and other procedural attachments
• concept type and a corresponding enforcement
• notion of generic types
• concept of multiple inheritance of a class from multiple other
• notion of class invariants
The notion of genericity is an abstraction of language that is deemed essential for an environment rapid development.
The simple inheritance between classes is considered as definitely inadequate: it does not correspond to the nature of the application. It is imperative to have a functionality of multiple inheritance to "stick" to the semantics of the problem addressed.

The realization of the object paradigm as described herein was made by one of the authors in Prolog on station UNIX work

5.2 The multi-windowing.

Observing the work of the nurse, having regard to patients under its charge, is characterized by a series of specific and ad hoc. Moreover, every action for a patient is often conducted in parallel to similar actions for other patients. This analysis shows quite clearly a need to converse with the computer using several parallel paths, each of which may in turn increase, can be interrupted, then leave. The ability to open multiple windows on a single screen work is absolutely necessary. Similarly, a widescreen 1000 x 1000 bitmap is there a standard to be considered with conviction.
The multi-windows are only good if they are made on a configuration suitable for securing adequate response time related to the second maximum in the key management window. So you need a Fast Ethernet network, a server disk to clean the unit of care (and not shared with other units) and a powerful processor on each station (at least today M68020).
Such a system will certainly be used by several nurses or doctors, during the same period. Windowing techniques have already proven themselves and it is easy to imagine several users coming and going to the terminal, finding their windows as they recognize an icon in the corner of the screen, opening with the mouse and resuming a process where they left off a few minutes before and after satisfaction they can leave the terminal by closing their windows if necessary. For data longer confidential, it must be put in place a system of time-outs and passwords to ensure the rights access.
In essence, this multi-windows combined with the practice of rapid prototyping, tends to spread widely. During the session SCAMC-1987 in Washington DC several presentations were made in this meaning. Include the application prototype CWS (Clinical Workstation Project) from Johns Hopkins. Multiples windows used are: "Mainscreen" window commands, the window kardex of patients, last window is used for the "unit summary". The merits are manifested at the outset of those speed Action that is to say, saving time in the practical application combined with improved dramatically the reliability of the data compared to the usual practice of applications on screen 24 lines x 80 columns.

5.3 The tools of the doctor's workstation Care Unit

If the question asked by the doctor is usually along the following lines: "Is there in the medical literature and in the archives of our institution, a patient of the same features as mine? What was he doing? -
What has happened to him? ", then the doctor's main tools are in addition to the facilities of order and receipt results of the HIS (to be seen in a window, eg, the window DIOGENE 'transactional system), the windows in the literature such as Medline (Paperchase or via DataStar), the window devoted to statistical research on computerized records of patients, the equivalent of 'CLINQUERY' of C. Safran, research recommendations and listings of clinical attitudes to support hypermedia.
A preliminary prototype test was performed on a workstation. The same approach also lends itself to taken into account equally well to ambulatory medicine than that of the nursing care plan.

6 – CONCLUSION

We showed that the penetration of computers in the nursing unit for specific applications and Autonomous nurses is a major task that was finally reached by technical means available on the market today.
Our description of a typical configuration is not a coincidence, but rather corresponds to the needs clearly recognized. We do not believe in intermediate solutions with limited equipment. It will anyway several years to run such applications, and by then the facilities described as required will be widely available on the market, scope of investment resources of hospitals.
Meanwhile, it is fair to mention here that most of the problems discussed throughout this paper were the subject of a research project submitted to the committee AIM (Advanced Informatics in Medicine) of the Community European P. Degoulet et al, under the acronym HELIOS (Hospital Environment Language Within a year Object Information System) with the collaboration of CAP-SESA, Heidelberg and Krebsforschungzentrum Computing Centre Hospital of Geneva.