Desenvolvimento e validação de um instrumento para mensuração do medication literacy em idosos.

Abstract

Medication literacy (ML) encompasses the skills to access, understand and properly use medication information. Aging-related changes impact ML, making older adults more prone to drug-related problems. This scenario highlights the need to measure ML to assess its effects on self-care, safety, and effectiveness in the use of medication by older adults. Thus, the aim of this thesis was to develop and validate an instrument to measure ML in older adults. The instrument was developed in multiple stages. First, a scoping review was performed to identify the definitions, instruments and factors associated with ML. The scoping review results were used to develop the conceptual model of ML, which was submitted to a panel of experts to evaluate the importance of its elements using a Likert-5 scale. The conceptual model obtained comprised four dimensions (functional, communicative, critical, and numeracy), five subdimensions (access, understand, communicate, evaluate, and calculate), the determinants and objectives of ML. An item bank was created to generate the preliminary version of the instrument, which was submitted to content validation. A panel of experts evaluated the relevance and clarity of the items using a Likert-3 scale and classified them per the theoretical subdimensions of ML. The instrument was modified and applied to 30 older adults for semantic analysis. New adjustments were made, giving rise to the pilot version of the instrument, containing 33 items distributed in eight fictitious medication-use scenarios. The pilot version was applied to 228 older adults (≥60 years) selected in the community and in two outpatient services of two teaching hospitals in Belo Horizonte for the evaluation of psychometric properties. One item presented high values of skewness (-2.5) and kurtosis (4.5) in the descriptive analysis of the items and was excluded from the instrument. Four, two and one-factor solutions were tested in exploratory factor analysis, which pointed to the one-dimensionality of the instrument; all items had a factor loading >0.30 in the one-factor solution. A three-parameter logistic model was adopted in the item response theory; one item had discrimination scores <0.65 and another three had guessing scores >0.35, which were excluded from the instrument. The final version, with 28 items, had excellent internal consistency (Kuder-Richardson formula 20 = 0.90); its scores were highly and directly associated with health literacy and educational level, pointing to its convergent validity. A preliminary standardization of the scores was carried out using the percentile norm, classifying the scores as low, medium, and adequate level of ML. Finally, the results of this thesis indicate that the instrument is psychometrically adequate to measure the ability of older adults to access, understand, communicate, calculate, and evaluate medication-related information. The instrument can be used for pharmacoepidemiological research and clinical practice, providing information to propose actions for the safe and effective use of medications in this population and the promotion of healthy aging.