DEVELOPMENT AND RELIABILITY OF AN INSTRUMENT TO MEASURE PSYCHOSOCIAL DETERMINANTS OF SALT CONSUMPTION AMONG HYPERTENSIVE PATIENTS

Este estudio tuvo como objetivo presentar el análisis de validez de contenido y de confiabilidad de un instrumento para estudiar los factores determinantes del consumo de sodio entre hipertensos, basado en la extensión de la Teoría del Comportamiento Planificado. El instrumento fue sometido a validez de contenido por 3 jueces y a una prueba piloto con 5 sujetos. El instrumento final, compuesto por 3 comportamientos relacionados al consumo de la sal y por las variables psicosociales correspondientes, fue aplicado a 32 sujetos para evaluación de la consistencia interna y de la estabilidad temporal (intervalo de 15 días). Fueron observados coeficientes alfa de Cronbach>0,70 para la mayoría de las variables y coeficientes de correlación entra clases significativas, que apuntaron la estabilidad temporal de los conceptos mensurados. El instrumento desarrollado mostró evidencias de validez de contenido y de confiabilidad.


INTRODUCTION
C h a n g e s i n f o o d p a t t e r n s , w i t h l e s s vegetable consumption, increased use of meet and, recently, industrialized products (1) , have resulted i n s a l t i n t a k e a b o v e r e c o m m e n d e d l i m i t s i n different population groups (2)(3) .A recent study (3) found daily salt consumption between 13 and 17g in hypertensive patients, with 40 to 55% of this consumption deriving from addition to foods during or after their preparation, while consumption of high-salt foods was also frequent.This intake greatly exceeds the recommended limit of 6g of salt/day for normotensive people or 4g of salt/day for hypertensive persons (4) .
Excessive salt consumption, in turn, has been associated with severity of arterial hypertension and ventricular hypertrophy (5) , resulting in international recommendations for the use of reduced-salt diets as an important non-pharmacological intervention in hypertension prevention and treatment (6) .
Hypertensive patients are familiar with recommendations about the need to restrict salt consumption in their diet.Many report following lowsalt diets, as appointed in earlier research (7)(8) .The actual measurement of this behavior, however, indicates continued high consumption in that population, which confirms that food pattern changes are not easy to achieve and maintain.There is growing evidence that education programs have the highest chance of success if they are aimed at factors that operate on the motivation to act (9) .
In social psychology, there are conceptual models that serve as a base to understand and to and, consequently, of behavior (10) .Among theories aimed at understanding and predicting health behaviors, the Theory of Planned Behavior (11) -TPB has been widely used, including for food-related behaviors (12)(13) .According to TPB, the proximal The application of these motivational theories requires the use of instruments based on psychometric scales aimed at measuring the magnitude of factors related to the behavior under analysis and its correlations and, like any other instrument for measurement of psychosocial variables, need to demonstrate evidence of psychometric properties, such as reliability and validity (14) . Despite

Instrument development
The instrument was designed after an extensive literature review of research that used sociocognitive theories to study food behaviors and specific studies on factors interfering in salt consumption.TPB was used as the theoretical framework, in addition to other variables related to food behavior (self-efficacy, habit and environment) or specifically to salt consumption (food preferences, hedonic determinants and diet quality assessment).
Direct measurements of the involved constructs were carried out (11) .

Definition of psychosocial variables
Behavior was initially defined as: consider a low-salt diet as: consumption in the last 2 months of according to the following elements: action, target, context and time (11) .Starting from the behavior, assertions were formulated to measure intention and its determinants.

Content validity
The first version of the instrument was submitted to three judges with expertise on health behavior research and experience in measurement instrument validation, who assessed conceptual adequacy and clarity of the items.This version was applied to five subjects with characteristics similar to the target population, in order to identify words or questions that were difficult to understand or misunderstood (15) .Little understanding on the generic issue of sodium consumption was verified among subjects who did not prepare their own meals.
Moreover, a new discussion with the judges revealed that, as sodium consumption is a behavioral category with different actions (16) , it could not be defined by one single question, so that the behavior was divided into three distinct actions, as follows.

Behavior 1 -the use of less than 4g of salt (corresponding to 1 "flat" teaspoon of salt) when cooking one of your usual recipes for any meal;
Behavior 2 -To avoid adding table salt to the food you eat at meals (or the snacks) and Behavior 3to avoid the consumption of food with high salt content (i.e., snacks, industrialized foods, salty spices, etc…).
Another recommendation from the judges was to investigate behavior 1 among subjects who prepared their own meal only, as it would not be coherent to investigate determinants of behaviors the subject does not perform.As a small number of male patients prepare their meals, the recommendation was to restrict the investigation of behavior 1 to female patients, who are responsible for preparing meals.

Subjective norm -
Reliability assessment -The final version of the instrument was applied to 32 subjects for reliability analysis, according to homogeneity (internal consistency), using Cronbach's alpha and the stability criterion, with the test-retest.Alpha coefficients above 0.70 were considered indicators of high consistency (17)   .
Fifteen days after the initial application, the instrument was again applied to the same 32 subjects to verify agreement between answers at both times, analyzing the intra-class correlation coefficient (ICC) and the weighted Kappa coefficient for one item in the food preferences, a nominal variable.

RESULTS
Study participants in this phase were 32 hypertensive patients with a mean age of 55. 13   (±7.86)years, 50% of whom were women, 53.1% married, with an average 7.2 (±3.62) years of education, and 14.7 (±11.6) years as the mean length of hypertension diagnosis.

Internal consistency
Table 1 shows the mean total and item scores, as well as Cronbach's alpha coefficients for each construct related to the three behaviors.Scores for most variables were above 4.0 for the three behaviors, suggesting that the study group is favorable to a low-salt diet.One exception was habit (behaviors 1 and 2), a variable that obtained lower mean scores than the other variables, suggesting that participants were accustomed to automatically add salts when preparing or to already prepared foods, without thinking about it.
In the reliability analysis, most variables for behaviors 1, 2 and 3 reached Cronbach's alpha coefficients above 0.70, except for subject norm, which scored between 0.58 and 0.67.For the three behaviors, removing item four of the subjective norm resulted in a substantially better coefficient.

Stability (test-retest)
Table 2 presents ICC values for each psychosocial variable related to the three behaviors and for the additional general variables, as well as the Kappa coefficient for the nominal variable.ICC for variables of behaviors 1 and 2 were >0.44, indicating temporal stability.In both behaviors, the intention construct revealed a trend towards temporal stability.
In behavior 3, all variables presented significant coefficients, indicating temporal stability, which was also observed for the general questions, with ICC levels of more than 0.45 and a weighted Kappa coefficient of 0.64.

DISCUSSION
This study aimed to test the psychometric properties of an instrument that was developed to study the determinants of sodium intake among hypertensive patients.The content validity phase indicated that salt consumption should be assessed not merely from the perspective of one single behavior, but according to three distinct behaviors that result in final dietary salt consumption.Sodium is present in foods in natura, in the salt added when preparing meals, in the saltshaker used on the table and, in high concentrations, in industrialized foods (18) .This particular characteristic of sodium consumption entails additional complications, also to estimate consumption in quantitative terms (3) .
In general, the data also evidence that the measures of the psychosocial variables that constitute the intention to perform the three behaviors, as well as the intention measure, are reliable according to the criteria of homogeneity and stability.For some of them, certain considerations are due about their psychometric performance.In the three behaviors, a low item-total correlation was observed for item 4 of the subjective norm, I feel social pressure to [behavior 1, 2 or 3], indicating that, although conceptually adequate, the patients did not interpret it consistently with the other items in the construct.Hence, in subsequent study phases, the deletion of this item must be considered when assessing subjective norm for the three behaviors.
As for temporal stability, all concepts measured revealed stability in the examined interval, which is important for future research aimed at analyzing the natural evolution of these perceptions, or at assessing the effects of educative interventions to modify them (12)   .
The development of scales to measure factors that support salt consumption behavior, mainly in hypertensive patients, who significantly benefit from the adoption of a low-salt diet, entails different consequences for nurses' educative practices.Reliable and valid measures are necessary to allow nurses to design and assess orientation programs based on social and behavioral theories (10) , with greater chances of success.
Before using this instrument, however, salt c o n s u m p t i o n n e e d s t o b e d i a g n o s e d : i t s quantification, followed by the identification of sources contributing to excessive consumption, like in earlier studies (3) .The next step is to diagnose the psychosocial determinants of each behavior.
Although these behaviors are complementary, factors supporting their practice can differ.Hence, each behavior can require a specific intervention on its determinant factors.Therefore, the identification of what factors motivated a person to adopt a lowsalt diet or not provides support to guide the elaboration and practice of interventions (10) .Besides, predictive factors can differ depending on populations or population subgroups.
In this sense, it is expected that applying scales is useful to detect such differences, guide specific interventions and, also, to serve as a reference to assess the efficacy of interventions put in practice in different population subgroups (19) .
In conclusion, data obtained in this study provided support for content validity and reliability according to internal consistency and stability criteria of the instrument to measure determinants of sodium consumption among hypertensive patients, given their behaviors of adding salt while preparing their meals, adding salt to prepared foods and consuming highsalt foods.This instrument is expected to be a useful diagnostic tool for health education intervention.

p r e d i c t h e a l t h b e h a v i o r
s .I n g e n e r a l , s t r o n g variations are found between subjects who adhere or not to certain behaviors, particularly in terms o f c o g n i t i v e f a c t o r s , d e s c r i b e d a s t h e m o s t important determinants of the motivation to act of volitional behavior is one's intention to engage in that behavior.Intention is determined by three variables: the attitude related to the b e h av i o r, s u b j e c t i ve n o r m s a n d p e r c e i v e d b e h av i o ra l c o n t r o l ( P B C ) (11) .F u r t h e r, f o r t h e behaviors that are not completely volitionally controlled, the TPB incorporates perceptions of control over performance of the behavior as an additional behavior predictor.The key idea of the model can be expressed by the equation: B ≈ I (+PBC) ≈ A + SN + PBC (where: B = behavior; I = intention; A = attitude; SN = subjective norm; PBC = perceived behavioral control).
g.: I feel weird if I don't do it) were reformulated for the sake of easier understanding.The final instrument was pretested again, involving five subjects from the target population, revealing that the understanding of behaviors and related variables had improved c o n s i d e ra b l y, w i t h n o n e e d f o r a d d i t i o n a l modifications.Instrument The instrument consisted of three behaviors related to salt consumption and their respective psychosocial variables: intention, attitude, subjective norm, perceived behavioral control, self-efficacy and habit (the latter for behaviors 1 and 2 only) as well as general variables, taken from other models or described as salt consumption-related.Behavior -The three behaviors were measured through the following question: in the last 2 months, which alternative best describes your behavior of [description of behaviors 1, 2 or 3] (1) all days of the weeks, (2) most days of the week, (3) sometimes, (4) rarely, (5) never.To calculate the final score, answer scores are inverted to maintain uniformity in the interpretation of scores (the higher the score, the more favorable to the behavior).Intention -intention was measured by six items on five-point Likert scales for each behavior, (e.g.: I intend to [description of behavior] in the next 2 months definitely not[1] -definitely[5]).
Ethical aspects -The project and the free and informed consent term, which participants signed before the interview, were approved by the local Research Ethics Committee (CEP No 563/2006).

Table 2 -
Mean scores of variables on test and retest and temporal stability analysis of behavioral variables 1,