Adaptation and evaluation of the measurement properties of the Brazilian version of the Self-efficacy for Appropriate Medication Adherence Scale1

Objectives: to undertake the cultural adaptation of, and to evaluate the measurement properties of, the Brazilian version of the Self-efficacy for Appropriate Medication Adherence Scale in coronary heart disease (CHD) patients, with outpatient monitoring at a teaching hospital. Method: the process of cultural adaptation was undertaken in accordance with the international literature. The data were obtained from 147 CHD patients, through the application of the sociodemographic/clinical characterization instrument, and of the Brazilian versions of the Morisky Self-Reported Measure of Medication Adherence Scale, the General Perceived Self-Efficacy Scale, and the Self-efficacy for Appropriate Medication Adherence Scale. Results: the Brazilian version of the Self-efficacy for Appropriate Medication Adherence Scale presented evidence of semantic-idiomatic, conceptual and cultural equivalencies, with high acceptability and practicality. The floor effect was evidenced for the total score and for the domains of the scale studied. The findings evidenced the measure's reliability. The domains of the Brazilian version of the Self-efficacy for Appropriate Medication Adherence Scale presented significant inverse correlations of moderate to strong magnitude between the scores of the Morisky scale, indicating convergent validity, although correlations with the measure of general self-efficacy were not evidenced. The validity of known groups was supported, as the scale discriminated between "adherents" and "non-adherents" to the medications, as well as to "sufficient dose" and "insufficient dose". Conclusion: the Brazilian version of the Self-efficacy for Appropriate Medication Adherence Scale presented evidence of reliability and validity in coronary heart disease outpatients.


Data collection procedure
The data were obtained by the researcher, individually, in a private environment, in accordance with the stages shown below. -Second stage: the Brazilian version of the SEAMS was reapplied (retest) in a proportion of the subjects who participated in the application (test), in similar conditions, with an interval of fifteen days between the first and second application. In this stage, only those participants whose return was arranged in the abovementioned service took part (n=34).

Data collection instruments
Instrument for sociodemographic and clinical characterization: the instrument constructed and subjected to content validity in a previous study was used (11) .
Definition of the drug therapy evaluated: the drug therapy evaluated was related to reduction in CHD's morbidity and mortality -lifesaving therapy -(that is, ACE-inhibitors, ARBs, Beta blockers, antiplatelet drugs and statins) and two other drugs which improve the signs and symptoms associated with coronary heart disease (that is to say, digitalis, diuretics and nitrates). interruption of the use of the drug as a result of perceiving improvement, and interruption of the therapy due to perceiving worsening in the clinical situation (12) .
The Brazilian version of the Morisky scale will be used (13) .
In the Brazilian version, a Likert-type response scale was used, of 4 to 5 points, varying from (1) Never to (5) Daily; (1) Never to (5) Always and (1) Never to (4) Always. The sum of the responses to the four items generates a score between 4 and 18; higher scores indicate low adherence; lower scores, high adherence. and association with time markers (fasting, breakfast and lunch), are not in accordance with the medical prescription. The participants classified in Group I were considered "adherent" and those classified in the other groups, as "nonadherent" (15) .
-General Perceived Self-efficacy Scale (GSE): an instrument created by Schwarzer and Jerusalem (16) , which is unidimensional and made up of 10 items, which refer to how to deal with success in a specified situation.
The participant responds to the instrument through a five point Likert response scale which varies from 1 (totally disagree) to 5 (totally agree). The total score has a variation from 10 to 50. A high score signifies a high perception of self-efficacy. The version adapted to Brazilian Portuguese was used (17) . (little confident), and 3 (very confident). The total score, which consists of the sum of the responses, can vary between 13 and 39; the higher the score, the greater the self-efficacy for adherence to the drug treatment (9) . This measures the proportion of judges who are in agreement regarding the items and general aspects evaluated (10) . The items' relevance and representativity was evaluated, through a Likert-type scale with scores varying between 1 and 4 (1= not relevant or not representative, 2= requiring major revision in order to be representative, 3= requiring minor revision in order to be representative, 4= relevant or representative).

Analysis of the data
The CVI was calculated through the sum of agreement of the items which received scores of "3" or "4", divided by the total number of responses. The items with scores of "1" or "2" were revised. participants who responded to all the items (18) . The floor effect, which is equivalent to the 10% of the scale's worst possible results, and the ceiling effect, which corresponds to the 10% of the scale's best possible results, were evaluated (19) .
-Evaluation of reliability: the Cronbach alpha coefficient was used to calculate the internal consistency, with a Cronbach alpha of >0.70 being established as evidence of satisfactory internal consistency (20) . In order to evaluate the stability of the measure, the Intraclass Correlation Coefficient (ICC) was used, with ICC >0.7 being considered satisfactory (21) .  It was hypothesized that the participants classified as "nonadherent" and "insufficient dose" would present the lower self-efficacy for medication adherence, according to the proportion of medication adherence.
A level of significance of 5% was adopted.

Ethical aspects
The study was approved by the university's

Practicality, acceptability and ceiling and floor effects
The results suggest that the Brazilian version of the SEAMS is an instrument which is easy to apply, with a  Table 2).    Table 4 -Spearman correlation coefficients between the scores of the Brazilian versions of the Self-efficacy for Appropriate Medication Adherence Scale, the Morisky Self-Reported Measure of Medication Adherence Scale and the General Perceived Self-efficacy Scale (n=147). Campinas, SP, Brazil, 2014 *Self-efficacy for Appropriate Medication Adherence Scale (SEAMS); †General Perceived Self-efficacy Scale (GSE); ‡r= correlation coefficient.

Discussion
In this study, the cultural adaptation of the SEAMS efficacy for a specific behavior -medication adherence.
However, it is emphasized that negative correlations of moderate to strong magnitude were observed between the SEAMS and the MMAS-4, which suggests convergent construct validity (9) .
In relation to the validity of known groups, it was observed that the dimensions and total score of the SEAMS discriminated between CHD patients classified as "adherent" and "non-adherent". Therefore, the sensitivity of the SEAMS, in the detection of differences between the groups, suggests that this instrument may be responsive, that is, capable of measuring changes in self-efficacy for medication adherence, over time. Data were not found in the literature relating to the validity of known groups of the SEAMS.
Self-efficacy is an important construct which can, partly, explain the behavior of medication adherence in CHD patients, as well as being particularly relevant as it is potentially modifiable (7) , being able to be the basis for the development of interventions related to behavioral change (24) .
The measurement provided by the SEAMS has potential applications for clinical practice and for research. with a low self-efficacy (25) . for this reason (26) . Although an objective measurement of medication adherence was not used, it is emphasized that more than one self-reported measure was used, with a view to obtaining a more accurate evaluation of medication adherence.
As a result, this study provides a tool with evidence of reliability and validity for measuring self-efficacy, for medication adherence, which could be useful in the evaluation of this construct, after nursing interventions directed towards the improvement of self-efficacy for medication adherence.

Conclusion
This study provides evidence that the Brazilian version of the Self-efficacy for Appropriate Medication Adherence Scale (SEAMS) is an instrument which is easy to understand, and whose measurement properties are