A design and study of computer search system of drug drug interactions based on the hospital formulary of a specific hospital

<p>Detection of drug drug interactions in the medication order is one of the important tasks of pharmacists and drug information specialists. A handy personalcomputer system was designed for a practical use. Drugs dealt with are those listed in the Hospital Formulary of Kobe-city Central Hospital. A computer used is PC-9801VX and an application program was written by BASIC. Data files created were a random Drug Master File and a random Interaction Master File. Fields of Drug Master File consist of registration number, brand name, first three letters of brand name, generic name, therapeutic category code, drug classification code, single or combination, and ingredients (in case of combination). Interaction Master File consists of registration number, drug classification code, clinical significance, treatment, and source code. Various expressions of drug interaction were grouped and coded into 79 types and 88 types with an addition of clinical symptoms. In order to make the computerization possible for an expression such as "an interaction between drug A and drug-group B", an original coding system and original logic system were developed and implemented. Data addition and deletion as well as the interaction search were executed promptly by making an access to data files directly and/or making a cross-over access among random files. Search time was examined by changing number of drugs. Five trials were performed for each case and the average of % standard deviation was 31%. As far as the search time spent is concerned it was found that it took only a moment (4 seconds on the average) when an interaction of two drugs were searched. When five drugs were searched the search time was 27 seconds on the average. The search time increased non-linearly as the number of drugs increased. This non-linearity is due to the fact that a two-way search was performed for a single interaction. In other words, when number of drugs, n, is examined, n(n-1) times searches are performed. In the actual situation of pharmacy practice number of drugs searched would be two, and it can be said that this system operates promptly and feasibly in such an occasion. In a rare case a search up to 15 drugs or so might be necessary, and the search time of 15 drugs was 189 seconds. The authors believe that this time is within an acceptable limit. Expressions used in literature for drug interaction were not scientifically controlled and there exists, in addition, discrepancy in expression for a single drug interaction of Drug A and B depending upon which literature (Drug A or B) was examined. A universal and uniform system of expression for drug interaction is urged to be established for the purpose not only of clinical use but also of computer application. In this system search was performed from both sides (Drug A and B) as a unique trial and both search results were shown on the display. Problems of drug interaction should not be treated just as an issue of discussion or argument, but they should be treated as problem which should be solved and coped with clinically. In order to correspond to those requirements, record fields for clinical advice were installed and significance of a particular drug interaction in clinical setting was taken into consideration. The three-stage rating system of P. Hansten was adopted and processed into the system. This trial as well as an implementation of "how to treat and cope with" information into the system, we believe, are unique and useful in problem solving of drug interaction. Issue of drug interaction has not yet been apprehended thoroughly not only in the public but also in the professional field of health. Computerization in this area of drug interaction is indispensable due to the complicated and kaleidoscopic nature of information. In order to investigate and solve various problems and enigmas existing in the information activity of drug interaction</p><p>(View PDF for the rest of the abstract.)</p>