Neighborhood clustering of non-communicable diseases: results from a community-based study in Northern Tanzania

BMC Public Health

Research Article

Uptal D. Patel¹  John W. Stanifer¹  Nathan Thielman²  Joseph R Egger³  Elizabeth L. Turner⁴ 

[1] Department of Medicine, Duke University, DUMC Box 3182, 27710, Durham, NC, USA;Duke Global Health Institute, Duke University, 27710, Durham, NC, USA;Duke Clinical Research Institute, Duke University, DUMC Box 3646, 27710, Durham, NC, USA;Duke University Medical Center, Box 3182, 27710, Durham, NC, USA;Department of Medicine, Duke University, DUMC Box 3182, 27710, Durham, NC, USA;Duke Global Health Institute, Duke University, 27710, Durham, NC, USA;Duke University Medical Center, Box 3182, 27710, Durham, NC, USA;Duke Global Health Institute, Duke University, 27710, Durham, NC, USA;Duke Global Health Institute, Duke University, 27710, Durham, NC, USA;Department of Biostatistics and Bioinformatics, Duke University, DUMC Box 2721, 27710, Durham, NC, USA;

关键词: Chronic kidney disease;    Cluster design;    Design effect;    Epidemiology;    Intra-cluster correlation;    Non-communicable disease;    Sub-Saharan Africa;    Variance;   

DOI  :  10.1186/s12889-016-2912-5

 received in 2015-10-08, accepted in 2016-03-01,  发布年份 2016

来源: Springer

【 摘 要 】

BackgroundIn order to begin to address the burden of non-communicable diseases (NCDs) in sub-Saharan Africa, high quality community-based epidemiological studies from the region are urgently needed. Cluster-designed sampling methods may be most efficient, but designing such studies requires assumptions about the clustering of the outcomes of interest. Currently, few studies from Sub-Saharan Africa have been published that describe the clustering of NCDs. Therefore, we report the neighborhood clustering of several NCDs from a community-based study in Northern Tanzania.MethodsWe conducted a cluster-designed cross-sectional household survey between January and June 2014. We used a three-stage cluster probability sampling method to select thirty-seven sampling areas from twenty-nine neighborhood clusters, stratified by urban and rural. Households were then randomly selected from each of the sampling areas, and eligible participants were tested for chronic kidney disease (CKD), glucose impairment including diabetes, hypertension, and obesity as part of the CKD-AFRiKA study. We used linear mixed models to explore clustering across each of the samplings units, and we estimated absolute-agreement intra-cluster correlation (ICC) coefficients (ρ) for the neighborhood clusters.ResultsWe enrolled 481 participants from 346 urban and rural households. Neighborhood cluster sizes ranged from 6 to 49 participants (median: 13.0; 25th–75th percentiles: 9–21). Clustering varied across neighborhoods and differed by urban or rural setting. Among NCDs, hypertension (ρ = 0.075) exhibited the strongest clustering within neighborhoods followed by CKD (ρ = 0.440), obesity (ρ = 0.040), and glucose impairment (ρ = 0.039).ConclusionThe neighborhood clustering was substantial enough to contribute to a design effect for NCD outcomes including hypertension, CKD, obesity, and glucose impairment, and it may also highlight NCD risk factors that vary by setting. These results may help inform the design of future community-based studies or randomized controlled trials examining NCDs in the region particularly those that use cluster-sampling methods.

【 授权许可】

CC BY   
© Stanifer et al. 2016