Manda Samuel O M, Lombard Carl J, Mosala Thabang
Biostatistics Unit, South African Medical Research Council, Pretoria, South Africa.
Geospat Health. 2012 May;6(2):221-31. doi: 10.4081/gh.2012.140.
An analysis of the ecological association between the human immunodeficiency virus (HIV) and syphilis was undertaken using joint mapping modelling based on data from South African national HIV and syphilis sentinel surveillance surveys carried out between 2007 and 2009. The syphilis prevalence, taken as proxy for sexual behaviour and increased HIV transmission, was first used with health district-level deprivation and population density as a covariate in a HIV prevalence spatial regression model and, secondly, together with HIV as a bivariate outcome. HIV is more highly prevalent in deprived and populated areas than elsewhere, while syphilis has a high prevalence in less deprived and less populated areas. Spatially, the HIV prevalence was lowest in the southwestern and highest in the northeastern parts of the country. This was in discordance to the syphilis prevalence, which revealed negative correlations with HIV prevalence. Considerable variations across the districts remained after adjusting for the contextual covariate factors. Divergent spatial patterns between HIV and syphilis were identified, regarding both observed and unobserved covariate effects. The differential disease-specific spatial prevalence patterns may point to inconsistent successes in interventions between the two diseases. Overall, the results emphasize the need to develop and test plausible aetiological hypotheses relating to ecological correlations and causes of the disease-specific interjectory between the districts.
利用基于2007年至2009年期间南非全国艾滋病毒和梅毒哨点监测调查数据的联合映射模型,对人类免疫缺陷病毒(HIV)与梅毒之间的生态关联进行了分析。梅毒患病率被用作性行为和艾滋病毒传播增加的替代指标,首先在艾滋病毒患病率空间回归模型中与卫生区层面的贫困程度和人口密度作为协变量一起使用,其次与艾滋病毒作为双变量结果一起使用。艾滋病毒在贫困和人口密集地区的患病率高于其他地区,而梅毒在贫困程度较低和人口较少的地区患病率较高。在空间上,该国西南部的艾滋病毒患病率最低,东北部最高。这与梅毒患病率不一致,梅毒患病率与艾滋病毒患病率呈负相关。在对背景协变量因素进行调整后,各地区之间仍存在相当大的差异。在观察到的和未观察到的协变量效应方面,艾滋病毒和梅毒之间存在不同的空间模式。不同疾病特定的空间患病率模式可能表明两种疾病干预措施的成功情况不一致。总体而言,结果强调需要制定和检验与生态关联以及各地区疾病特定轨迹的原因相关的合理病因假设。