Structural Equation Modeling Test of an Integrated Model of Spanish Youth’s Condom Use

Heterosexual transmission represents 26 % of newly diagnosed infection in Spanish youth. Behavioral change models have emphasized the influence of multiple variables to predict condom use behavior. The aim of this study is to examine how those variables are organized and which theory explains the condom use behavior better. A sample of 424 young heterosexuals (M age  = 20.62; SD = 2.16) filled out a battery of self-report questionnaires for assessing AIDS-related variables, personality traits and clinical variables (general, sexuality-related and health-related). A structural model was specified that included perceived pleasure and condom use self-efficacy as predictive variables. Depression and sexual compulsivity indirectly influence behavior. The final model accounted for 65.9 % of the variance in behavior. These results highlight the importance of cognitive and emotional variables as predictors of behavior (ex. expectations of pleasure and self-efficacy beliefs). This is important information for designing effective psychological interventions. La transmisión heterosexual representa el 26 % de los nuevos diagnósticos de VIH en los jóvenes españoles. Los modelos de cambio de comportamiento han enfatizado la influencia de múltiples variables para predecir el uso del preservativo. El objetivo del presente estudio es examinar cómo se organizan estas variables y qué teoría explica mejor esta conducta. Una muestra de 424 jóvenes heterosexuales (x edad  = 20.62; SD = 2.16) completó una batería de cuestionarios sobre Sida, rasgos de personalidad y variables clínicas (generales, sexuales y de salud). Los resultados obtenidos utilizando modelos de ecuaciones estructurales apoyan una relación directa entre uso del preservativo y autoeficacia y placer percibido. Las variables depresión y compulsividad sexual influyen indirectamente sobre la conducta. El modelo explica el 65.9 % de la varianza. Estos resultados destacan la importancia de las variables cognitivas y emocionales como predictoras del comportamiento. Esta información es importante para el diseño de intervenciones psicológicas eficaces.

Abstract Heterosexual transmission represents 26 % of newly diagnosed infection in Spanish youth. Behavioral change models have emphasized the influence of multiple variables to predict condom use behavior. The aim of this study is to examine how those variables are organized and which theory explains the condom use behavior better. A sample of 424 young heterosexuals (M age = 20.62; SD = 2.16) filled out a battery of self-report questionnaires for assessing AIDS-related variables, personality traits and clinical variables (general, sexuality-related and health-related). A structural model was specified that included perceived pleasure and condom use self-efficacy as predictive variables. Depression and sexual compulsivity indirectly influence behavior. The final model accounted for 65.9 % of the variance in behavior. These results highlight the importance of cognitive and emotional variables as predictors of behavior (ex. expectations of pleasure and self-efficacy beliefs). This is important information for designing effective psychological interventions.

Introduction
Epidemiological reports observe a progressive sexualization of the AIDS epidemic in Spain. Almost [1].
The unprotected sex put at risk for HIV infection, other sexually transmitted infections (STIs) and unwanted pregnancy [2][3][4][5]. Furthermore, safe sexual behavior (for example, abstinence, consistent condom use or mutual monogamy with an HIV-negative partner) is the only way to prevent sexual transmission of HIV and STIs. The researchers continue to analyze the factors that predict risk of sexual behaviors. Different theoretical models of behavior have been applied in the analysis of determinants of condom use in heterosexual relationships. The most commonly used have been: the Health Belief Model (HBM) [6], the Theory of Planned Behavior (TPB), developed from the Theory of Reasoned Action [7,8], the Social Cognitive Theory (SCT), updated version of the Social Learning Theory [9], or the Transtheoretical Model (TTM) [10]. These paradigms focus on different factors in attempting to explain the behavior, but all of them share a cognitive-social orientation.
From HBM, key elements to adopt preventive behaviors for HIV infection are that people perceive AIDS as a serious disease, perceive themselves at risk, know prevention mechanisms, are motivated to implement them and have the necessary resources to carry out protective behaviors [11]. Different studies have shown how the components of HBM are good predictors of behaviors related to HIV infection in youth. The high perception of vulnerability is related to delaying first sex and fewer sexual partners and also condom use [12]; the perceived benefits has been linked to condom use [13] and perceived barriers to unprotected sex [14].
On the other hand, the TPB has predicted 33 and 43 % of the intention to use condoms in young people [15][16][17], the intention to refuse sex with a new partner [18] or the intention to use contraception [19]. In a meta-analysis by Albarracin, Johnson, Fishbein and Muellerleile, a total of 96 studies using the TPB as a basic model to explain the condom use were examined [20]. The study shows that the behavior was related to behavioral intention (.45); behavioral intention was related to attitudes (.58) and the subjective norm (.39); attitudes were associated with behavioral beliefs (.56); norms were associated with normative beliefs (.46); and perception of control was associated with the behavioral intention (.45) and condom use (.25). In the implementation model it has been observed that the intention to use condoms predicts significantly its use during sex; the perception of control and attitudes are the best predictors preventive behavior; the subjective norm has more influence on the formation of intent to action on specific groups, such as adolescents [21][22][23].
The SCT proposes a model of human behavior integrated within a cultural context. Therefore, the behavior is the result of interaction between an auto-system, which allows to measure the control on own thoughts, feelings, motivations and actions, and external influences. This model promotes self-protective factors, enhance self-efficacy, and provide information to reinforce the benefits of healthy behaviors. Many of the preventive interventions in young people have taken this theoretical model [24,25]. It has also been used to explain the sexual risk behavior and promote condom use [26][27][28].
Finally, TTM is based on the premise that changes in behavior occur incrementally and through a predictable sequence of stages. In this regard, DiClemente and Prockaska observed that subjects who are placed in the precontemplation stage of change do not yet see themselves as having a problem, they are not thinking seriously about changing and tend to defend their current problem behavior [10]. It is in the contemplation stage when people are more aware of the personal consequences of their problem behavior and they spend time thinking about it. People value the pros and cons of modifying their behavior, so there is a greater emotional and cognitive implication to use condoms [29].
These models have a well-articulated set of theoretical constructs, which have facilitated the psychological interventions design to prevent HIV infection. But, in the first decade of the HIV epidemic, they have not achieved the necessary behavioral changes for primary prevention in young people. Efforts to increase the effectiveness of preventive interventions require a thorough understanding of the social, contextual and interpersonal determinants of risk behaviors for HIV infection [30]. The behavioral change models have emphasized the influence of specific factors as the level of information about HIV, attitudes towards condoms or beliefs related vulnerability to AIDS. But also the influence of other personality dimensions, clinical (for example, fear of negative evaluation, depression, self-esteem or worrying about health) or health-related factors (for example, health locus of control or value in health care) [31][32][33]. Abundant empirical literature suggests that there is some consensus on what variables are most important to explain the condom use behavior. But much less about how those variables are causally organized to predict behavior or what theory might be better than the other at explaining behavior. The reviewed studies have explained between 50 and 66 % of the variance of behavioral intention or condom use. They are mainly used by the TPB as a framework, although some of them have incorporated other variables as the dynamics of the relationship [22,[34][35][36][37][38][39].
The theoretical development of behavioral determinants of condom use has not been stopped. The search for approaches and scientific models that explain the behavior of condom use both individually and collectively continues. Therefore the main aims of this study are: to examine the influence of the variables included in the main theoretical models (HBM, TPB, SCT and TTM) and other dispositional variables in condom use among young people's heterosexual relationships, and test a structural equation model for predicting condom use behavior.

Methods Participants
A sample of 424 young heterosexual people were evaluated (60.4 % were women and 39.6 % were men). All participants had tertiary-level studies and a mean age of 20.62 years old (SD = 2. 16).
Given that the only safe behavior to prevent HIV infection entails the systematic use of the condom and that the remainder of response options would imply a certain amount of risk, it was decided to form two dichotomous groups with a equivalent distribution by sex and the mean age:

AIDS-Related Variables
• AIDS Prevention Questionnaire (CPS) This instrument was made up of 65 items that attempt to gather up the various components considered to be relevant in various HIV prevention models: HIV/AIDS knowledge, perceived susceptibility, perceived severity, fear of HIV infection, perceived condom use benefits and barriers, condom use self-efficacy, behavioral intention and behavior. The internal consistency and test-retest reliability of the data were acceptable, obtaining a Cronbach's Alpha value of .70 [40].

Personality Traits
• Revised NEO Personality Inventory (NEO-PI-R) The questionnaire is structured on the five dimensions of the Five-Factor Model (Neuroticism, Extraversion, Openness to Experience, Agreeableness and Conscientiousness). The version used for this study is the Spanish adaptation. The questionnaire contains 240 items that are answered on a five-point Likert scale ranging from strongly agree (1) to strongly disagree (5) [41,42].

Clinical Variables
• Rosenberg Self-Esteem Scale (RSE) It is a unidimensional instrument, made from a phenomenological conception of self, which measures the respect and acceptance of people to themselves. A 10-item scale whose items are answered using 4-point Likert scale format ranging from 1 (strongly agree) to 4 (strongly disagree). The Spanish version used has an internal consistency of .85 and .88 [43]. • Beck Depression Inventory II (BDI-II) It is a 21-item self-report instrument intended to assess the existence and severity of symptoms of depression as listed in the DSM. There is a four-point scale for each item ranging from 0 to 3, sorted from lowest to highest severity. It has a high coefficient alpha (.89) in the Spanish version used [44,45]. • Penn State Worry Questionnaire (PSWQ) It is a 16-item measure of trait anxiety designed to assess the general tendency to experience the experience of worry. It is a 5-point scale ranging from 1 (not at all typical of me) to 5 (very typical of me). The experimental version used is an adaptation to the worry about health with a internal consistency of .90 [46,47]. • The Brief Fear of Negative Evaluation Scale (BFNE) It is the measure most commonly used to determine the degree to which people experience apprehension at the prospect of being negatively evaluated. It contains 12-items to which respondents rate the degree to which each statement applies to them on a 5-point Likert scale ranging from 1 (not at all characteristic of me) to 5 (extremely characteristic of me). The Spanish version used has a internal consistency of .90 [48,49].

Sexuality-Related Clinical Variables
• Sexual Sensation Seeking Scale (SSSS) The scale is an 11-item, Likert-type measurement that asks respondents to endorse the extent to which they agree with a series of statements related to personality disposition sensation seeking. The answers vary in a range from 1 (not at all like me) to 4 (very much like me). The version used for this study is the Spanish adaptation which has an internal consistency of .70 [50,51]. • Sexual Compulsivity Scale (SCS) The scale is a 10-item, Likert-type measurement that asks respondents to endorse the extent to which they agree with a series of statements related to sexually compulsive behavior, sexual preoccupations, and sexually intrusive thoughts. The answers vary in a range from 1 (not at all like me) to 4 (very much like me). The version used for this study is the Spanish adaptation which has an internal consistency of .84 [50,52]. • Sexual Pleasure/Affinity The original version of the measure was composed of 7 items that assessed the perceived pleasure of sexual behaviors related to varying degrees of risk for HIV infection along a 5-point scale, ranging from 1 (Not at all pleasurable) to 5 (Extremely pleasurable). In the adaptation to heterosexual population, some items were added. Therefore, the scale was composed of 11 items [50].

Health-Related Clinical Variables
• The Health Locus of Control Scale (HLC) An instrument that measures generalized expectancies regarding locus of control related to health. It is a 11-item in a 6-point Likert format, ranging from 0 (strongly disagree) to 5 (strongly agree). The experimental version was used [53,54]. • Health Self-Care Scale (unpublished) A 11-item which represent efforts to maintain self-care in different areas of health, (for example, physical exercise, check-ups, etc.).
It is a numerical rating scale for self-reporting of self-care from 0 (I never care about it) to 10 (I always care about it). The overall alfa de Cronbach obtained, .686, indicates a good internal consistency of the questionnaire.

Procedure
The sample collection process was carried out in a period of about 2 years. The development of this study is conducted in the framework of a larger research project that aims to bring preventive HIV strategies adapted to the characteristics of the population to which they are addressed, so that more efficient and effective interventions are achieved for behavioral changes in young people.
Each year the Unit sets up informative tables and panels on campus on World AIDS Day. Interested persons were given the opportunity to participate in studies carried out by the Unit. The first step was to contact via email or phone the youth, who had given us their data, to inform them about the objectives and procedure of the present study. After signing the informed consent, they completed the questionnaires (approximately 90 min). In the present study we applied the guidelines of the Spanish data protection law known as Ley Orgánica de Protección de Datos (LOPD) and the Declaration of Helsinki, to guarantee the confidential nature and treatment of the data obtained and to protect the ethical principles for research involving human beings.

Analysis of Data
To select the criterion variable and to carry out the statistical analyses, biological risks of the practices evaluated were taken into account (oral sex, vaginal penetration and anal penetration), as well as their frequency. It was chosen as a dependent variable (DV) condom use in vaginal intercourse, due to its high frequency because it represents one of the highest risk sexual behaviors for HIV infection. A dichotomous variable from single-item AIDS Prevention Questionnaire is coded: How often have you used a condom in vaginal intercourse? So the dependent variable (DV) of study belonged or not to a risk group for HIV infection. The value 1 corresponds to the risk group (RG), that is, not use condoms consistently during sex (never, sometimes and quite often). While value 0 represents norisk group (NRG), who themselves have used condoms consistently.
Following the differential study, a multivariate analysis was carried out using a multiple logistic regression analysis. Logistic regression was the chosen analytical method for two reasons: (a) The conditions of multivariate normality, homoscedasticity and linearity are not required, and (b) the model may incorporate independent variables of different types [55]. The Enter method, in which all variables in a block are entered in one step, was used in order to find the best predictors.
Then, structural equation modeling (SEM) was used. The statistical program used is the EQS 6.1 which enables robust analysis method assuming that a normal multivariate distribution is not followed. Maximum likelihood estimation of missing data was used and thus robust estimation of standard errors was conducted for tests of fit and significance of the paths. It is recommended that social research use the following absolute fit indices (

Results
At the moment of evaluation, 63.2 % (n = 268) of the sample were having sexual intercourse with a steady partner. Fifteen per cent (n = 40) of these young people reported being unfaithful to their partners with others and almost half of them did not use a condom (n = 17). While 22.4 % of youth report a single sex partner within their live, 13.7 % report two, 14.9 % report three, and 49 % report four or more (M = 4.59, SD = 4.93). The prevalence of consistent condom use in vaginal intercourse was 39.3 % and of no-consistent condom use was 60.7 %. When it asked about the intention to use condoms in future sexual intercourse, increased intention to use condoms always (64.4 %) and decreased intention to inconsistentuse (35.6 %).

Preliminary Analysis
A multiple regression logistic analysis with all the above variables was performed with the enter method. The value of R square Naglekerke indicates that the proposed model accounts for 46.4 % of the variance of the dependent variable. This equation generated ten explanatory variables which were age, depression (BDI's punctuation), sexual pleasure without condom (Sexual pleasure/affinity factor), and fear of HIV infection (CPS item) as risk factors. And Compliance (facet of Agreeableness), Competence (facet of Conscientiousness), sexual pleasure with a condom (Sexual pleasure/affinity factor), condom use self-efficacy (CPS factor), behavioral intention (CPS item), interference (SCS dimension) as protective factors. Hosmer-Lemeshow test did not obtain statistical significance (v 2 = 8.031; p = .430), indicating a goodness of fit of the model. In general, a good classification result is obtained with an average of 76.4 % of classifications being correct. The results are better with regard to sensibility, given that in the RG 82 % of subjects are correctly classified; the results are considerably worse in relation to specificity, as here there is a correct classification of 67.8 % of the participants that belonged to the NRG. The variables age and sexual pleasure without condom multiplied by 20 and 18 %, respectively, the risk of not using a condom during sex. Moreover, the personal interference reduced it by 73.5 % (see Table 1).
Perceived pleasure with a condom (B = .573, p \ .05) and without condom (B = -.439, p \ .05), and self-efficacy (B = .371, p \ .05) were predictors of behavior in the model. There is an indirect effect of depression on behavior through self-efficacy (B = -.357, p \ .05), also of sexual compulsivity on behavior through perceived pleasure without a condom (B = .298, p \ .05).

Discussion
In an attempt to advance the search for a conceptual design integration to enable effective prevention campaigns and programs to prevent HIV among heterosexual youth, two issues were raised. What variables act as risk and protective factors in condom use? And, how are they organized? These questions guided our research design.
To answer the first question a regression analysis with all variables was performed. Most of the variables that predict the non-use of condoms were associated with emotional and affective factors except age, which is a socio-demographic variable. The older youth use fewer condoms consistently during sex. It seems logical to think that there is greater probability of steady relationships; therefore people could use other contraceptives methods that do not prevent STIs. Different studies in Spain suggest the existence of the phenomenon called monogamy not 123 protective [2][3][4][5]. They reported about negative beliefs that arise when using a condom with a steady partner (e.g., mistrustfulness, lack of love for the other person). On the other hand, the experience of fear of a HIV infection and depressive traits appear as risk factors. Authors confirmed the inverted U-shaped relationship between negative emotions and preventive behaviors [11,58]. The experience of negative emotions may partially affect the processes of self-regulation and, thus, this can interfere with the ability to initiate sexual activity, to refuse unwanted sexual activity and to negotiate wanted sexual relationship [39,59]. Similarly, because this is a correlational study, negative emotions can also occur as a result of having sexual risk behavior. The short-term benefit to get pleasure prevails over the long term cost of a possible disease. In the cost-benefit balance, these youth attach greater importance to achieve short-term benefit in the form of sexual pleasure than the possible negative future consequences their behavior may cause. Perhaps these people underestimate the risk; they are more susceptible to the reinforcing effects of pleasurable stimuli considered [60,61]. The psychological characteristics that have appeared as protective factors are related to cognitive, motivational and behavioral variables. The SCS's interference dimension made significant contributions to understanding the decision of safer sexual behavior. As Carnes pointed out, sexual compulsives often have sexual feelings and cognitions of great intensity and frequency [62]. Interpersonal interference of compulsive behavior probably facilitates problem awareness and therefore the risk awareness. Social consequences of certain sexual behaviors motivate the consistent use of condom. Moreover, the expectation for physical sexual pleasure with condom was associated with practicing safe sex. Perhaps, safe sex practices are psychologically reinforced by one's perceived sexual pleasure. Self-efficacy beliefs not only affect how well individuals motivate themselves for use of condoms, it also affects the choices they make at important decisional points. According to Bandura, ''if self-efficacy is lacking, people tend to behave ineffectually, even though they know what to do'' [9]. Alike, as experience in condom use increases, the skills also increase, which will surely have an impact on improving self-efficacy expectations. Finally, the two facets which were related to safe sex outcome, compliance and competence, are included in the domain of Agreeableness and Conscientiousness. People with high interpersonal skills, with tendency to think before acting, that consider potential consequences, are less likely to participate in risky sexual behaviors [63][64][65]. On the other hand, the importance of condom use self-efficacy has been demonstrated in many studies [27].
Preliminary results suggest that there are two profiles of young people. If it is understood that the decision to use a condom in a sexual relationship seems to be a type of psychosocial stress, it is possible to apply the model by Lazarus and Folkman. Transactional approach defines as person-environment transactions the stressful experiences. When youths are faced with a condom-use decision, they evaluate the significance and potential threat (primary appraisal). The secondary appraisals address what one can do about the situation. It seems that youth who consistently use a condom employ coping problem-focused strategies, and youth who do not consistently use a condom, employ coping emotion-focused strategies [66].
Our second question responds to how all these significant variables are organized in a explanatory model of youth's condom use. Behavior was composed by self-reported condom use and behavioral intention because they have appeared closely related [20,21,29,36]. The behavior was directly predicted by perceived sexual pleasure and condom use self-efficacy. Depression and sexual compulsivity have an indirect influence on behavior.
Self-efficacy is the primary variable of SCT but also appears in other theoretical models (HBM, TPB as control beliefs, or TTM). Self-efficacy emphasizes the individual and reflects a person's level of confidence in his or her ability to control the environment. A person, whose cognitive self-evaluation or judgement of their capabilities is high, will tend to use a condom with greater confidence. Empirical evidence has demonstrated that people with high levels of condom use self-efficacy make successful decisions regarding sexual health, both ability to use condoms as ability to reject sexual risk behaviors [22,23,[26][27][28][29]67]. Self-efficacy is not a static characteristic, it can be altered by internal personal factors in the form of cognitive and affective events. So that, those who score higher on measures of self-efficacy show substantially fewer symptoms of depression. Self-efficacy acts as a buffer against negative feelings that is the effect of stressful life events on depressive symptoms is mediated through the impact of stressful life events on self-efficacy [68,69].
The expectancies about sexual pleasure is a variable associated with the HBM, and subsequently to the TPB. Pleasure is the main element of human sexual motivation [70]. Condoms can interpose a mechanical barrier, limit physical contact, reduce tactile sensation, attenuate heat transduction, or affect other aspects of sexual functioning masculine and feminine [71]. Studies support the hypothesis that persons who believe condoms interfere with pleasure or reduce pleasure, or who rate condom-protected sex as less enjoyable or pleasurable than unprotected sex may be less likely to use condoms in practice, and conversely [72][73][74][75]. This double profile emphasizes the role of cognition or emotion as important factors in the decision process [5,33,[76][77][78]. When making decisions, there exists an imaginary balance between a desire for immediate gratification and delayed gratification. Impulsive people give into temptation and desire for immediate gratification, where the longer term consequences might be ignored or not evaluated. Non-impulsive people reflect the benefits of delayed gratification and ability to wait for these benefits, perhaps to avoid undesirable consequences [79].

Limitations
Some limitations need to be addressed. First, we have taken as criterion variable the condom use in vaginal intercourse. As indicated earlier, the reasons are that it is one of the riskier practices and at the same time it is one of the most  (Item-3 If I have to suggest to a partner to use a condom, I have/should fear of rejection by him/her; Item-5 I am sure I would remember to use condoms although I have used alcohol or other drugs; Item-6 I feel/would feel uncomfortable when put on a condom or put it on my partner); Sexual Compulsivity (Item-3 My desires to have sex have disrupted my daily life; Item-6 I find myself thinking about sex while at work; Item-8 I have to struggle to control my sexual thoughts and behavior; Item-9 I think about sex more than I would like to; Item-10 It has been difficult for me to find sex partners who desire having sex as much as I want to); Sexual Pleasure without condom (Item-1 Vaginal sex without condom; Item-3 Receptive anal intercourse without condom; Item-10 Receptive oral-genital sex without condom); Sexual Pleasure with condom (Item-2 Vaginal sex with condom; Item-7 Insertive oral-genital sex with condom) frequent practices enjoyed by young people. However, it would be important to analyze whether the same results are replicated in other sexual practices and risk situations, for example oral sex, anal intercourse, a steady partner, a casual partner and under the effect of drugs. Second, the results should be generalized with caution, due to the specific sociocultural characteristics of the participants (e.g., high level of education) and they were interested in AIDS-related research. Future studies should include more representative samples of youth to test the universality of the mechanisms found in this study. Finally, prediction is used in the context of a cross-sectional study that may not take into consideration changes over time.

Implications and Contribution
Nevertheless, the innovative aspect of the current study lies in the evaluation of a comprehensive model for condom use in young heterosexuals. Findings contribute to the understanding of the role of each component and their possible integration into a unified explanatory framework. It has been shown that the behavior of condom use is mainly influenced by the appraisal of self-competence to use condoms and expectancies about sexual pleasure with/ without condoms, also mentioning the distal role of depressive symptoms and sexual compulsivity. Therefore, two types of young people are intuited within the reflectivity-impulsivity continuum. The components of psychological interventions aimed at risk groups should focus on the work of beliefs about sexuality (love, romance, pleasure, etc.), positive attitudes towards condom use, focusing on the sensory and sensual aspects of themselves, selfregulation and management of risk decision making.