Akhtar S, Luby S P, Rahbar M H, Azam I
Department of Community Health Sciences, Medical College, The Aga Khan University, Karachi, Pakistan.
Southeast Asian J Trop Med Public Health. 2001 Jun;32(2):351-61.
This study was conducted on prison inmates in Sindh to determine whether HIV/AIDS related knowledge, attitudes and beliefs can predict their practices which risk HIV infection. A pre-designed questionnaire was administered in this cross-sectional study to collect the data on HIV/AIDS related knowledge, attitudes, beliefs, practices and demographic variables in a systematic sample of 3,395 prison inmates during July 1994. The data on responses of inmates to HIV/AIDS related knowledge, attitudes, and beliefs were analyzed and a clear interpretable factor structure emerged for each set of questions labeled as knowledge, attitude and beliefs. Similarly based on responses of inmates to practice questions, three factors emerged and were labeled as heterosexuality, homosexuality and drugs. The standardized factor scores of inmates for each of these six factors were computed and used in further analyses. Multiple linear regression analyses were carried out separately using heterosexuality, homosexuality and drugs factors score as dependent variables to identify if any of the independent variables (demographic variables, knowledge beliefs and attitude) predict these practice factors. The model for heterosexuality explained 23% of the variance and included HIV/AIDS related knowledge, beliefs, age, ethnicity and marital status and duration of imprisonment (F = 84.33, p < 0.001; R2= 23.0). The predictors in the model for homosexuality together explained 10% of the variance and included significant contribution by belief, martial status, ethnicity, education, age and duration of imprisonment (F = 24.76, p < 0.001; R2= 0.10). The model for drugs had significant contributions from HIV/AIDS related beliefs, marital status and ethnicity (F = 20.10, p < 0.001; R2= 0.03). Implications of prevention program based on these results are considered.
本研究针对信德省的监狱囚犯展开，以确定与艾滋病毒/艾滋病相关的知识、态度和信念是否能够预测其可能导致艾滋病毒感染的行为。在这项横断面研究中，于1994年7月对3395名监狱囚犯的系统样本进行了预先设计的问卷调查，以收集有关艾滋病毒/艾滋病相关知识、态度、信念、行为及人口统计学变量的数据。对囚犯关于艾滋病毒/艾滋病相关知识、态度和信念的回答数据进行了分析，针对每组分别标记为知识、态度和信念的问题，出现了清晰且可解释的因素结构。同样，根据囚犯对行为问题的回答，出现了三个因素，并分别标记为异性性行为、同性性行为和毒品。计算了这六个因素中每个因素的囚犯标准化因素得分，并用于进一步分析。分别以异性性行为、同性性行为和毒品因素得分为因变量进行多元线性回归分析，以确定是否有任何自变量（人口统计学变量、知识信念和态度）能够预测这些行为因素。异性性行为模型解释了23%的方差，包括与艾滋病毒/艾滋病相关的知识、信念、年龄、种族、婚姻状况和监禁时长（F = 84.33，p < 0.001；R2 = 23.0）。同性性行为模型中的预测因素共同解释了10%的方差，包括信念、婚姻状况、种族、教育程度、年龄和监禁时长的显著贡献（F = 24.76，p < 0.001；R2 = 0.10）。毒品模型中，与艾滋病毒/艾滋病相关的信念、婚姻状况和种族有显著贡献（F = 20.10，p < 0.001；R2 = 0.03）。基于这些结果对预防项目的影响进行了探讨。
