It is a general problem for software developers that the world will not stand still while a timely concept is converted into marketable software. This is particularly true in areas such as medical informatics where concepts evolve rapidly and manpower is limited.
The PRAKTICE system described in Gabrieli’s brief paper accepts clinical notes in a loosely structured natural language form. These are then analyzed and stored in an encoded form suitable for subsequent analysis.
It is clear from the text and the references that Gabrieli has been working on clinical records for nearly two decades. He is not alone in his ambition to codify medical text (see Mishelevich’s 1972 review of three experimental systems [1]).
As Gabrieli points out, the choice of coding scheme is a critical factor. The project dismissed existing schemes such as the Systematized Nomenclature of Medicine and the International Classification of Diseases as internally inconsistent, and a new hierarchical code was developed.
In 1978 Major et al. showed that medical terminology has a rich semantic structure that cannot be captured in a single hierarchical classification [2]. The simple example in Gabrieli’s paper does not show how the PRAKTICE coding scheme overcomes this problem.
The goals of the system were ambitious when they were set in the 1970s; it is a pity that the project does not appear to have benefited from the publications that have appeared since then.