Aplicación de análisis de redes para la elaboración de perfiles epidemiológicos en estudios sanitarios
Description
Resumen. Habitualmente en los estudios epidemiólogicos se suele trabajar con variables binarias que muestran la presencia de determinadas enfermedades, las que a su vez se asocian con otro conjunto de enfermedades, denominadas comorbilidades,
y que también se miden a través de variables binarias y que en general se asumen como factores de riesgo de las primeras. En el ámbito de estos estudios existen situaciones donde se manejan enfermedades no transmisibles (ENT), en particular
en salud bucal, donde ambos tipos de variables pueden ser intercambiables en cuanto
a quien hace el rol de factor de riesgo.
Teniendo en cuenta esta situación se propone usar el análisis de redes para la determinación de tipologı́as de encuestados en base a los atributos binarios, de forma de obtener perfiles epidemiológicos bien diferenciados.
Los datos utilizados corresponden a un estudio en personas que demandan atención en la Facultad de Odontologı́a-Udelar durante el perı́odo 2015-2016. A través del análisis de redes (AR), a partir de las variables se construye la matriz de ad-
yacencias sobre la que se aplican una baterı́a de métricas (closeness, betweenness, modularity, clustering) sobre los nodos y enlaces, que permite detectar comunidades.
Las comunidades creadas mediante el AR a través del uso de diferentes algoritmos de búsqueda, como el de fast greedy o de random walk o de particionadoespectral, se usan para evaluar el cambio en las proporciones de las variables analizadas (patologı́as o factores de riesgo) cuando se consideran en forma global y al interior de cada comunidad.
Usually in epidemiological studies it is usual to work with binary variables that show the presence of certain diseases, which in turn are associated with another set of other diseases, called comorbidities, and that are also measured through binary variables and that are generally assumed as risk factors of the former. Within the scope of these studies, there are situations where noncommunicable diseases (NCDs) are managed, particularly in oral health, where both typesof variables can be interchangeable as to who plays the role of risk factor.
Taking into account this situation, it is proposed to use network analysis to develop typologies of respondents based on binary attributes, in order to obtain well-differentiated epidemiological profiles.
The data used correspond to a study in people who demand attention in the Facultad de Odontologı́a -Udelar during the period 2015-2016. Through the social network analysis (SNA), the adjacency matrix is constructed from the variables, on
which several metrics (closeness, betweenness, modularity, clustering) are applied on the nodes and links, which allows the detection of communities. The communities created through the AR using different search algorithms, such
as fast greedy, random walk or spectral partitioning, are used to evaluate the change in the proportions of the variables analyzed (diseases or risk factors) considered in global level and inside of each community.
Files
Puebla_2020_capi7.pdf
Files
(1.3 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:67b77206711ca630ab061643ca746034
|
1.3 MB | Preview Download |
Additional details
Additional titles
- Translated title (En)
- Application of network analysis for the development of epidemiological profiles in health studies
References
- Bhupathiraju SN, T. K.,Coronary heart disease prevention: Nutrients, foods, and dietary patterns., Clin Chim Acta, 412(17-18):pp.1493–514,2011.
- Bonacich, P.,Power and centrality: A family of measures. American Journal of Sociology, pp.5:1170,1987
- Bonacich, P. and Lloyd, P., Eigenvector-like measures of centrality for asymmetric relations. Social Networks, pp.23:191 – 201, 2001.
- Borgatti, S. P., Everett, M. G., and Johnson, J. Analyzing Social Net- works,SAGE Publications Ltd., 2013.
- Brandes, U. A faster algorithm for betweenness centrality, The Journal of Mathematical Sociology, pp. 25(2):163–177,2001.
- Brandes, U. and Erlebach, T.,Network analysis: methodological founda- tions. Number 3418 in LCNS, Tutorial. Springer, Berlin ; New York, OCLC: ocm58474176, 2005.
- Butts, C. T.,sna: Tools for Social Network Analysis.R package version 2.4.,2016.
- Cook, N., Cutler, J., Obarzanek, E., Buring, J., Rexrode, K., and SK, K. , Long term effects of dietary sodium reduction on cardiovascular disease outcomes: observational follow-up of the trials of hypertension prevention (tohp). British Medical Journal, 334(7599):,pp.885–8.,2007
- Csardi, G. and Nepusz, T., The igraph software package for complex network research. InterJournal, Complex Systems:1695,2006.
- Food and Agriculture Organization of the United Nations, Fats and fatty acids in human nutrition. Report of an experte consultation 10-14 november 2008, FAO,2010.
- Huang, Z. (1997) A fast clustering algorithm to cluster very large categorical data sets in data mining. in kdd: Techniques and applications. Technical report, World Scientific.,1997.
- Kolaczyk, E. and Csárdi, G. Statistical analysis of network data with R.Springer, New York, 2014.
- Kolaczyk, E. and Csárdi, G. sand: Statistical Analysis of Network Data with R, R package version 1.0.3, 2017.
- Luke, D. A user's guide to network analysis in R. Springer, 2015.
- Newman, M. E. J. Assortative mixing in networks. Phys. Rev. Lett., 89:208701, 2002
- R Core Team R: A Language and Environment for Statistical Computing.R Foundation for Statistical Computing, Vienna, Austria, 2016.
- Sabidussi, G. The centrality index of a graph. Psychometrika, 31(4):, pp.581– 603.,1966.
- Skapino, E. y Alvarez Vaz, R. Prevalencia de factores de riesgo de enfer- medades crónicas no transmisibles en funcionarios de una institución bancaria del Uruguay. Revista Uruguaya de Cardiologı́a, 31:,pp. 246 – 255.,2016.
- Wasserman, S. and Faust, K. Social network analysis: methods and ap- plications. Number 8 in Structural analysis in the social sciences. Cambridge University Press, Cambridge ; New York.,1994.
- Weihs, C., Ligges, U., Luebke, K., and Raabe, N.. klar analyzing german business cycles.In Baier, D., Decker, R., and Schmidt-Thieme, L., editors, Data Analysis and Decision Support, pp. 335–343, Berlin. Springer-Verlag., 2005.
- World Cancer Research Fund International (2007). Food, nutrition, physical activity and the prevention of cancer: a global perspective. Technical report, World Cancer Research Fund, American Institute for Cancer Research., 2007.