Potential prescription patterns and errors in elderly adult patients attending public primary health care centers in Mexico City

INTRODUCTION Six out of every 10 elderly persons live in developing countries. OBJECTIVE To analyze and assess the drug prescription patterns and errors in elderly outpatients attending public health care centers in Mexico City, Mexico. MATERIALS AND METHODS A descriptive and retrospective study was conducted in 2007. Fourteen hundred prescriptions were analyzed. Prescriptions of ambulatory adults aged >70 years who were residents of Mexico City for at least two years were included. Prescription errors were divided into two groups: (1) administrative and legal, and (2) pharmacotherapeutic. In group 2, we analyzed drug dose strength, administration route, frequency of drug administration, treatment length, potential drug-drug interactions, and contraindications. Variables were classified as correct or incorrect based on clinical literature. Variables for each drug were dichotomized as correct (0) or incorrect (1). A Prescription Index (PI) was calculated by considering each drug on the prescription. SPSS statistical software was used to process the collected data (95% confidence interval; p <0.05). RESULTS The drug prescription pattern in elderly outpatients shows that 12 drugs account for 70.72% (2880) of prescribed drugs. The most prescribed drugs presented potential pharmacotherapeutic errors (as defined in the present study). Acetylsalicylic acid-captopril was the most common potential interaction (not clinically assessed). Potential prescription error was high (53% of total prescriptions). Most of the prescription errors were due to omissions of dosage, administration route, and length of treatment and may potentially cause harm to the elderly outpatients. CONCLUSIONS A high number of potential prescription errors were found, mainly due to omissions. The drug prescription pattern of the study population is mainly constituted by 12 drugs. The results indicate that prescription quality depends on the number of prescribed drugs per prescription (p < 0.000).