Extended Abstract
1. Objectives
Aging is a global phenomenon that will soon be addressed among the most critical social and welfare challenges of developing countries. In Iran, the aging process of the population is also inevitable. Demographic projections in 1420 indicated an increase in the number of older adults aged ≥65 years by 12 million [1]. The most important indicator of health and longevity is the hope of recovery and life expectancy. These important indicators are affected by many factors and reflect the socioeconomic, cultural, and health status of each society [2]. Numerous older adults in Iran are struggling with biopsychological health problems [3]. This study attempted to provide the Structural Equation Modeling (SEM) of life expectancy in the elderly considering their anxiety sensitivity, social support, and pain perception.
2. Methods & Materials 
This was a descriptive-analytical study with a cross-sectional design conducted in 2016. The study population consisted of all home resident older men and women aged ≥60 years in Isfahan City, Iran. Of whom, 340 subjects were selected as the study samples. They completed the following questionnaires: 
The 15-item Life Expectancy questionnaire by Khalilian Shalamzari et al. (2013): Its internal consistency was calculated as 0.97 using Cronbach’s alpha coefficient. The 12-item Perceived Social Support Scale by Zimet et al. (1988): It has three subscales addressing three sources of support; family, friends, and significant others. Its subscales’ Cronbach’s alpha was reported as 0.89, 0.86, and 0.82, respectively.
The 16-item Anxiety Sensitivity Index by Reiss, Peterson, Gursky, and McNally (1986): The test-retest reliability coefficient of 0.75 for two weeks and 0.70 for 3 years intervals indicate the reliable structure of this scale. The psychometric properties of its Persian version were examined by Bayrami et al., quoted by Narimani and Sharbati (2015). They tested its internal consistency, test-retest reliability, and split-half reliability, and reported the values of 0.93, 0.95, and 0.97, respectively. The 14-item Pain Perception Scale by Yousefi et al. (2016): For its internal consistency, the Cronbach’s alpha coefficient of 0.90 was obtained. The collected data were analyzed in SPSS and AMOS using Pearson’s correlation coefficient, regression analysis, and SEM.
3. Results
The study participants were 144 older men and 196 older women. Their Mean±SD age was 66.91±7.44 years. Most of them aged 60-65 years (47.9%), while (26.8%) aged 66-70 years, and only (3.2%) aged ≥85 years. Furthermore, (52.6%) were educated below high school, (25.9%) had a high school diploma, and (21.5%) held academic degrees. Moreover, (56.6%) were married, (26.8%) widows/widowers, (5.6%) divorced, and (2.1%) of the subjects were single. In addition, (36.2%) of the subjects had 3-4 children, (30.3%) had 1-2 children, and 4.1 had no children. Furthermore, (46.8%) of the samples were housekeepers, (31.8%)were retired, and (24.4%) were employed. In total, (48.5%) of the study participants reported chronic health problems, (29.7%) reported acute health problems, and (21.8%) reported no health problems.
The Mean±SD scores of life expectancy, pain perception, anxiety sensitivity, and social support were 50.51±5.71 years, 39.48±15.71, 40.20±12.59, and 63.91±14.62, respectively. Regarding the anxiety sensitivity dimensions, the Mean±SD scores of physical concern, cognitive concern, and social concern were 20.54±9.60, 10.06±7.69, and 3.25±3.01, respectively. Regarding the social support dimensions, the Mean±SD values of perceived support from family, friends, and significant others were 61.69±5.12, 20.19±5.46, and 22.06±6.83, respectively.
According to Table 1, the values of fit indices were close to 9; thus, they were at appropriate levels, which indicate the goodness fit of the proposed model. This model was not significant at P<0.05 (X2=9.82, df=104). This finding suggests that the proposed model is not significantly different from the significant model (Figure 1). This result confirms our hypothesis that the SEM of life expectancy is possible based on anxiety sensitivity, social support, and pain perception. The achieved results only suggested a significant positive correlation between life expectancy and social support and its three dimensions (P<0.01). The cognitive concern (anxiety sensitivity) had a significant negative relationship with life expectancy (P<0.05). Moreover, pain perception, overall anxiety sensitivity, and its two dimensions of physical and social concerns had no significant relationship with life expectancy.
4. Conclusion
The obtained results revealed that life expectancy was significantly related to social support and its dimensions. This finding was consistent with the results of Shakerinia [3] and Souliotis et al. (2016). Our results regarding the significant negative relationship between cognitive concern subscale (i.e. anxiety sensitivity) and life expectancy were in line with those of Movahedi et al. and Ben-Arye and associates [5, 6]. There was a significant positive relationship between pain perception and anxiety sensitivity and its dimensions. Moreover, there was a significant negative relationship between pain perception and social support and its dimensions. These results were consistent with those of Vesal and colleagues.

Furthermore, our results were consistent with those of Davis et al. regarding the significant relationship between life expectancy and pain perception, anxiety sensitivity, and its physical and social concerns dimensions. Unexpectedly, social support more effectively reduced perceived pain and anxiety sensitivity, compared to life expectancy. Increasing perceived social support may reduce pain and anxiety and improve hope and life expectancy in the elderly.

Ethical Considerations
Compliance with ethical guidelines
All ethical principles were considered in this study. Participants were free to leave the study at any time. They all were informed about the research objectives and confidentiality of their information.
Funding
This study was extracted from a master thesis of Mahnoosh Kiani in Psychology approved by Khorasgan Branch of Islamic Azad University in Isfahan.
Authors' contributions
All authors contributed in preparing this article.
Conflicts of interest
The authors declared no conflict of interest.