Mining Sequential Patterns in Pharmacy Database Using Genetic Algorithm

Medication management is an important process in pharmacy field. Prescribing errors occur upstream in the process, and as such, their effects can be perpetuated, and sometimes even exacerbated, in subsequent steps. Prescription errors are an important issue for which conflicts with another prescribed medicine could cause severe harm for a patient. Former knowledge of prescriptions can reduce these errors. The discovery of such knowledge requires one of data mining techniques, such as sequential pattern. Sequential pattern mining is an important technique which describes potential sequenced relationships among medicine prescriptions in pharmacy database. In addition, time is the most important factor, in pharmacy, due to the shortage of pharmacists and to contain the cost of healthcare delivery. Since sequential pattern algorithms takes a long time to find the exact optimal rules while evolutionary algorithms can find good rules in a short time, then Genetic Algorithm (GA), which is an evolutionary algorithm, can be used to discover sequential pattern rules in pharmacy database. In this paper we will use Genetic algorithm to assess patient prescriptions based on former knowledge of series of prescriptions in order to extract sequenced patterns and predict unusual activities to reduce prescribing errors