2.2. Study population
According to the study by Pelch et al. [9] and considering an alpha coefficient of 0.05 and beta coefficient of 0.2 and a correlation of 0.18 between social role quality and physical health and well-being, the sample size was estimated as 240 according to the following formula. To increase the accuracy by 10% and consider potential sample loss, the sample size was increased to 250.
Sampling took place through the convenience method until the determined sample size was reached.
Inclusion criteria
Being an Iranian woman aged 18–45 years with AIDS and having health records in the select center, a CD4 less than 500 (as per the center’s treatment protocol), minimum reading and writing literacy, a sexual partner, and no other self-reported psychological or physical illnesses.
Exclusion criteria
Incomplete completion of the questionnaires and the self-reported use of psychotropic substances during the study, which did not allow proper responding to the study questions.
2.3. Data collection and definition of terms
Data were collected using five questionnaires, including the Psychosocial Adjustment to Illness Scale (PAIS), the Perceived Social Support Scale (PSSS), the Multidimensional Health Locus of Control Scales, the Illness Perception Questionnaire (IPQ), the Experiences in Close Relationships (ECR) Scale and a socio-demographic checklist.
To examine social roles, which are defined as the product of personal and group efforts by individuals to reap the benefits of their accumulated roles and prevent many social problems, such as the feeling of social alienation [12], three questionnaires were used, namely the ECR, PSSS and a socio-demographic checklist. In this study, physical health was assessed using two questionnaires, namely the Multidimensional Health Locus of Control Scales and the IPQ.
Socio-demographic checklist
A researcher-made checklist including items on age, education, occupation, marital status, insurance status, number of children, and method of affliction were completed for all the participants.
Psychosocial Adjustment to Illness Scale
The Psychosocial Adjustment to Illness Scale (PAIS) consists of 46 items in seven domains, including health care orientation, vocational environment, domestic environment, sexual relationships, extended family relationships, social environment, and psychological distress. Scoring is on a 4-point scale from zero to three. The sum of the scores of each component was divided by the number of items in that component and the mean was considered that component's compatibility score, and the sum of all the scores was divided by the total number of items and the mean total score was taken as the total score. The validity and reliability of this tool was calculated in Iran in 2013 and the scale’s Cronbach's alpha was calculated as 0.94 [13]. The present study assessed the reliability of the tool in the target group and the Cronbach's alpha coefficient for the total score of psychosocial adjustment was calculated as 64.1.
Experiences in Close Relationships (ECR) Scale
The Experiences in Close Relationships (ECR) is a 36-item tool that measures adult attachment in two dimensions (18 items each), namely attachment-related anxiety and attachment-related avoidance in close relationships. The items are answered on a Likert scale from ‘strongly agree’ to ‘strongly disagree’ (1–7), and each person receives a score from 6 to 42 in the avoidance dimension and 3 to 21 in the anxiety dimension of each domain of relationship. The first four items are reverse scored. The validity and reliability of this questionnaire have been assessed in Iran by Seyedi et al. in 2013 [14]. The present study also assessed the reliability of this questionnaire in the target group and its Cronbach's alpha coefficient was found as 63.3% for anxiety and 84% for avoidance.
Perceived Social Support Scale
The Perceived Social Support Scale (PSSS) is a 12-item tool designed by Zimmett et al. (1988) to assess perceived social support from three sources, namely the family, friends, and significant others. To obtain the total score, the scores of all the scale items are summed up and divided by their number –that is, 12 [15]. The present study assessed the reliability of the questionnaire in the target group and the Cronbach's alpha for social support was reported as 0.91.
Illness Perception Questionnaire
The Illness Perception Questionnaire (IPQ) was designed by Broadbent et al. in nine subscales based on the revised form of this questionnaire, and its first eight items are scored from 1 to 10 while item nine is open-ended and asks about the three major causes of the disease in respective order. Each subscale measures one component of illness perception. In Iran, the validity and reliability of this questionnaire were examined by Bazzazian. The Cronbach’s alpha coefficient of the questionnaire was reported as 0.8 and its retest reliability coefficient after a six-week interval was reported as 0.42 to 0.75 for the different items [16]. The present study examined the reliability of this questionnaire in the target group and reported a Cronbach’s alpha coefficient of 62.4 for the entire scale.
Multidimensional Health Locus of Control Scales
The Multidimensional Health Locus of Control Scales was developed by Wallston and DeVellis in 1978. This questionnaire has 18 items in three subscales (internal locus of health control, powerful others locus of control and chance health locus of control). Each scale consists of six items, which are measured using a 6-point Likert scale (‘strongly disagree’ =1 point to ‘strongly agree’ =6 points). In Iran, this scale was first translated and localized by Meshki, Ghafrani Pour, Azad Fallah and Hajizadeh (2007), and the reported coefficients were 70% for internal health locus of control, 75% for powerful others locus of control and 69% for chance locus of control [17]. The present study re-evaluated the reliability of each dimension in the target group and reported Cronbach’s alpha coefficients of 47.5 for the internal locus, 59.2 for the powerful others locus and 58.6 for the chance locus.
Procedures
An ethics code was obtained from the Ethics Committee of Alborz University of Medical Sciences (ABZUMS.REC.1397.195) and the Ethics Committee of the Counseling Center for Behavioral Diseases of Imam Khomeini Hospital in Tehran (IR.TUMS.VCR.REC.1398.435).
First, the researcher visited the counseling center of Imam Khomeini Hospital and identified the eligible candidates and briefed them on the study objectives. Written consent was then obtained from them if they wished to participate in the study. The study questionnaires were then distributed among them to answer. The researcher was present at the center during this stage and answered any potential questions the respondents had. If the questionnaires could not be completed in a single session, the subjects were invited for a follow-up visit. There was no obligation or imposition to complete the questionnaires in the first session and they could be completed at any time during the research. A separate space was allocated in the center for this purpose, so that the questionnaires could be completed without stress and others’ presence.
Statistical analyses
This study examined the fit of a conceptual model of path analysis (Fig. 1) to determine the concurrent relationship of psychosocial adjustment variables, social roles (experiences in close relationships, perceived support and socio-demographic characteristics), and physical health (locus of health control, illness perception). First, the normality of the quantitative variables was assessed by the Kolmogorov-Smirnov test. Path analysis is a generalization of the normal regression that, in addition to expressing the direct effects, also demonstrates the indirect effects and the effect of each variable on the dependent variables and can be used to provide a rational interpretation of the observed relationships and correlations. Data were analyzed in Lisrel-8.8 and SPSS-16 software. The correlation results were analyzed using Pearson’s correlation coefficient and the path analysis was expressed as beta. The level of significance was set at T-values > 1.96.