OPADE: optimization of drug prescription using advanced informatics

Refining instructional text generation after evaluation

In this paper, we describe how user-adapted explanations about drug prescriptions can be generated from already existing data sources. We start by illustrating the two-step approach employed in the first version of the natural language generator and the limitations of generated texts, that we discovered through analytical and empirical evaluations. We claim that, although style refinement would be needed in these texts, particular care should be devoted to implementing some of the persuasion techniques that doctors employ in their explanations. This would require either thoroughly revising the text planning techniques employed or converting to a multistep generation architecture. We justify why we selected this second alternative and propose some heuristics to repair problems found in the first version of the generator. Some final considerations about the advantages of this approach and the possibility of generalizing it to other domains conclude the paper.

An efficient fault-tolerant order entry management information system based on special distributed client/server architecture

An automatic order entry system is very important for the processing of out-patient information, not only helping doctors to enter their orders directly but also reducing errors of communication. Many hospitals are anxious to set up a direct order entry system but are concerned about possible system failures. In this paper we report on an effective and efficient fault-tolerant order entry management system which satisfies the requirements for out-patient order entry. From the results of experiments on a prototype we found that the system was user friendly and reduced the time taken. Doctors are able to enter their orders more easily, accurately and quickly by selecting from the standardized and personalized menus to be found in the system.

An efficient fault-tolerant out-patient order entry system based on special distributed client/server architecture

An automatic order entry system is very important for processing out-patient information. This system not only helps physicians to enter their orders directly, but can also reduce order communication error and thus improve medical quality. Therefore, many hospitals have high aspirations to generate and implement direct order entry systems, but they are also concerned about the setbacks of system failure. In this paper, we present an effective and efficient fault-tolerant order entry system based on special distribution client/server architecture that satisfies the requirements of out-patient order entry very well. From the experimental results carried out on a prototype, we found that this system can improve the system response time of order entry and can also generate an operational method having a user friendly interface. The physicians can enter their orders easily, accurately, directly, flexibly and at a faster rate by making choices from standardized and personalized menus in this system.

Standardized structure and modular design of a pharmacokinetic database

BACKGROUND

The accumulated knowledge on drugs can be used for an individual drug dosage adjustment if it is placed at our disposal in an informatically structured form.

THEORY AND METHODS

We have started building up a pharmacokinetic database aimed at adjusting drug dosages, in exemplary form, to patients with renal impairment. Parameters needed for the three dosage adjustment rules (Dettli, Kunin, Holford) and the most general concept of pharmacokinetics constituted the theoretical basis.

TWO PROCESSES PERTAIN TO ALL DRUGS

Distribution and elimination. Total drug clearance and at least two parameters representing distribution and elimination processes are closely interdependent in mathematical terms (clearance = volume of distribution*rate of elimination). This relation yields the unifying concept that serves as a prerequisite for a structured recording of 30 assigned pharmacokinetic and pharmacodynamic parameters within an informatic database.

SOLUTIONS AND RESULTS

The information is retrieved and referenced from 2383 original publications by means of a standardized input module. The complete database at present contains 15,397 records for 1573 drugs. A programmed meta-analytic algorithm is used to calculate the statistical measures for the central value and variance--as available--from the pooled values of primary records. The statistically standardized parameters are extracted for 6601 pharmacokinetic parameters, and placed at the users disposal with the output module.

PRACTICAL UTILITY

Following meta-analysis, published pharmacokinetics can be used as statistical estimates of population parameters. The statistical estimates with variances permit an individual drug dosage adjustment by applying the Bayesian approach or neural networks.