What this paper adds
INTRODUCTION
In the treatment of chronic no communicable diseases (NCDs), adherence to medication use has been associated with optimization of clinical outcomes, especially better disease control,1,2reduction of hospitalizations,1mortality,3 and health care costs.4However, the percentage of medication nonadherence remains high.
Medication adherence is one of the most complex self-care behaviors in the treatment of NCDs.5 According to the middle-rangeTheory of Self-care of Chronic Illness , self-care can be defined as a process of health maintenance by means of health practices and disease management, which can be applied to health and disease situations.6 In this context, the maintenance of self-care refers to behaviors to maintain well-being, health and physical and mental stability, such as smoking cessation, healthy food consumption, stress management, and medication adherence.
Nurses have a central role in promoting self-care,7especially regarding medication adherence, which implies an assessment and a decision on the need to intervene. However, the measure of this behavior has shown to be extremely complex.
There are several difficulties in measuring adherence.8,9 Although there is no consensus on a ”gold standard”,10 self-report instruments have been indicated as the best option for measurement of adherence, with regard to ease of implementation, low cost, flexibility (time and mode of administration), and low burden on the respondent. However, self-report instruments present potential disadvantages, especially social desirability and memory bias, which compromise the accuracy of this type of measure.11
Several self-report scales that measure adherence in chronic diseases are available in the literature.12-14 TheMorisky Medication Adherence Scale (MMAS) consisting of four items (MMAS-4) or of eight items (MMAS-8) are the ones of the most frequently used tools.15,16 However, reports of overestimation of adherence with the use of self-reported measures have been frequent.
In order to provide an accurate measure of medication adherence, theGlobal Evaluation of Medication Adherence Instrument(GEMA)17 was developed based on the previous study.11 The GEMA assumes that adherence is a complex and dynamic behavior, defined as ”taking medication for treatment, exactly as prescribed, which means, taking it every day, at the time and amount indicated, remembering the care needed when taking it, ”before and after meals and/or at bedtime”.17 This instrument proposes, in addition to the measurement of the proportion of doses, the assessment of self-care associated with medication intake. In addition, to reduce memory bias, GEMA proposes to access the memory to retrieve the proportion of doses taken in different periods, up to the period of interest for measurement: that is, the month prior to the interview.
This new instrument seeks to fill the gaps in the literature regarding the imperfections of self-report measures.18 Although the GEMA has been used in previous studies,17,19-21 to our knowledge, no study investigated its properties of the measurements. The objective of this study was to evaluate the measurement properties of the GEMA when administered to patients taking oral anticoagulants (OAC) in an outpatient follow-up. The feasibility, acceptability, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) were investigated; the validity of the convergent construct was tested with self-reported adherence measures and International Normalized Ratio (INR) stability.
METHODS
2.1 Design and setting
This was a methodological study conducted in an OAC outpatient clinic of a large university hospital, in the interior of the São Paulo state, Brazil.
2.2 Sample
The study included 127 adult patients taking OAC in an outpatient follow-up service. Patients who had been using OAC for at least six months were included. Patients whose OAC dosage was modified in the last month prior to the interview, who presented hemorrhagic or thromboembolic complications in the last three months, or who underwent surgery in the last six months prior to the interview were excluded.
2.3 Sampling process and sample size
The sample consisted of all the patients who met the inclusion criteria, from January to March of 2016. In the outpatient clinic, 200 patients were treated with OAC; of these, 177 agreed to participate; 50 were included in the pilot sample, and excluded from the final sample. Thus, the final sample consisted of 127 patients.
2.4 Data Collection Procedure
Data were obtained by means of interview, using instruments. The results of the last three INR dosages, and the individual therapeutic goal recommended for each patient, were obtained from the medical record.
2.5 Instruments
2.5.1 Instrument of Sociodemographic and Clinical Characteristics: was developed in a previous study22 and was submitted to content validity.