Medication Error Due to Drug Name Similarity: An algorithmic approach to Orthographic Similarity

Numerous studies have identified drug name similarity as a significant source of medication error. One of the approaches employed to minimise this problem has been to apply algorithms designed to act on strings of letters. Two sets of Spanish drug name pairs were used. Eight algorithms reported in the literature were applied to establish the degree of similarity between the different names. Statistical methods were used to assess algorithm performance. The main element of analysis was V of Cramer. Simple mathematical and logical operations were applied to the algorithms in order to enhance differences between the two sets of drug name pairs studied and thus optimise the final results. The results confirm that the algorithms differentiated between the two sets of drug names, indicating that they could serve as a useful tool to identify potential confusion between drug names that could lead to medication errors.