Connectionist Model to Help the Evaluation of Medical Equipment Purchasing Proposals

Abstract

There is in the developing world a great number of idle medical equipment, due to the absence of experienced professionals to conduct an effective purchasing plan in its several phases, including vendors proposals evaluation. As artificial neural networks are typically applied for pattern recognition and function approximation, it was developed a decision-making computational model, based on artificial neural networks, which entries were grades given to physical risk, cost and strategic importance to a chosen medical equipment. The outputs were also grades attributed by clinical engineers according to the importance of five factors (clinical, financial, quality, safety and technical) during the equipment evaluation. The use of the model’s outcome allows any clinical engineer to identify the proposal that best attend the health unit requirements. To validate this model, a national inquiry (32 clinical engineers) was conducted using an electronic chart that permitted to: (a) establish a major professional profile of the inquired engineers; (b) determine which were the most important criteria considered during a medical equipment procurement process and (c) generate 95 examples that were used to train, and to test, diverse types of artificial neural networks. Hence, to represent the knowledge of clinical engineers (for the evaluation process of purchasing proposals) who worked at public hospitals, with three to ten years of experience, the best results were encountered for an ensemble of 100 two-hidden-layers perceptrons trained with the Backpropagation algorithm. The neural networks responses presented average reliability superior than 85% in all cases studied. Therefore, a connectionist computational model can be useful during a decision making process to help hospital managers to choose an appropriate medical equipment.