Objectives:To examine the relationship of social capital with anxiety and depression symptoms among urban elderly in China.Methods:A household survey by using self-administered questionnaires was conducted in 3 communities in Jinan,Shandong Province and the data of 925 people aged 60 years or over were collected.The social capital indicators according to its operational definition and Zung's Self-rating Depression Scale (SAS) and Zung's Self-rating Depression Scale were used.The social capital factors were extracted by using exploratory factor analysis and were dichotomized to conduct logistic regression analysis.Results:The average standard scores of SAS and SDS were (35.8 ±6.9) and (41.9 ±8.7),respectively.The proportion of anxiety symptom was 4.4％ and that of depression symptom was 19.0％.Logistic regression analysis showed that perceived social support,perceived trust,reciprocity and safety and social participation were associated with self-rating depression(Ps＜ 0.05).Perceived social support (OR ＝ 0.29,95 ％ CI:0.20-0.44),perceived trust,reciprocity and safety (OR ＝0.49,95％ CI:0.33-0.72) and social participation (OR ＝ 0.66,95％ CI:0.45-0.98) were protective factors of depression.No significant association between social capital factors and anxiety was found.Conclusion:It suggests that social capital could be a target for elderly people's mental health promotion in China.Additional efforts should be taken on enhancing the elderly people's perceived support,trust and safety from their families and communities,as well as providing more opportunities of social activities in the neighborhoods.

Objective:To analyze and evaluate the items of the general module of multidimensional health measurement instruments system for elderly patients with chronic disease (MHIEC-GM)(V1.0) based on classical test theory (CTT) and item response theory (IRT).Methods:A self-administered survey of 2 375 elderly patients with chronic diseases was conducted using the MHIEC-GM(V1.0) scale. The quality of items was analyzed using the variability method, correlation coefficient method, factor analysis method, and Cronbach's α coefficient method based on CTT. The difficulty coefficient, differentiation coefficient, and information content of items were analyzed using the IRT analysis software MULTILOG 7.03.Results:The CTT results showed that the standard deviations of all nine items were greater than 0.9. The results of the correlation coefficient method, the factor analysis method and the Cronbach′s α coefficient method showed that except for the GMI9 item, the correlation coefficients between the other items and the total scale were all greater than 0.5, and the factor loads were all greater than 0.5, and the Cronbach′s α coefficients after the deletion of the items were all less than 0.763. Based on the results of the four analysis methods, except for the GMI9 item, all the other items had good characteristics. The IRT results showed that the information contents of all items were >0.333 (5/15) except for the GMI4, GMI6, and GMI8, and the differentiation coefficient of each item was 1.44. The items that difficulty coefficient met the [-4, 4] inclusion criteria were the GMI2, GMI3, GMI5, GMI7 and GMI9, showing a monotonically increasing trend with increasing difficulty level. The items of GMI1, GMI4, GMI6, and GMI8 needed further improvement.Conclusion:Most of items of the MHIEC-GM(V1.0) scale have good psychometric properties, but some items need to be further revised and validated according to the characteristics of chronic diseases in the elderly.