Journal Information
Vol. 13. Issue 5.
Pages 391-398 (September - October 1999)
Vol. 13. Issue 5.
Pages 391-398 (September - October 1999)
Open Access
Los modelos multinivel o la importancia de la jerarquía
Visits
10189
E. Sánchez-Cantalejo
Corresponding author
esc@easp.es

Emilio Sánchez-Cantalejo Ramírez. Escuela Andaluza de Salud Pública. Campus Universitario de Cartuja, Apdo. de Correos 2070. 18080 Granada.
, R. Ocaña-Riola
Escuela Andaluza de Salud Pública (Granada)
This item has received

Under a Creative Commons license
Article information
Resumen

Una parte importante de la investigación sanitaria proporciona bases de datos en las que se puede establecer alguna estructura jerárquica. Así, los individuos estudiados, unidades muestrales de nivel 1, pueden pertenecer a grupos o unidades mayores, por ejemplo, la zona geográfica donde viven, el médico que los atiende, el hospital que los asiste, etc., las unidades de nivel 2. La posible mayor homegeneidad entre individuos de un mismo grupo respecto a individuos de distintos grupos invalidaría la hipótesis de independencia necesaria para poder utilizar los modelos tradicionales de regresión. Los modelos multinivel, también denominados modelos jerárquicos, permiten solventar esta dificultad al distinguir los distintos niveles jerárquicos de las predictoras, separando la variabilidad de los individuos objeto de estudio de la de los grupos a los que pertenecen. Aunque estos modelos se desarrollaron en la pasada década, especialmente en el campo de la educación, en los últimos años se ha puesto en evidencia su aplicabilidad en la investigación sanitaria. Este trabajo describe algunos modelos multinivel sencillos, discute sus ventajas sobre los métodos clásicos y presenta algunos ejemplos de aplicación tanto en la investigación epidemiológica como en la de servicios sanitarios.

Palabras clave:
Análisis Multinivel
Modelos Estadísticos
Análisis de Regresión
Análisis Multivariante
Análisis Estadístico
Summary

Many researchers in Public Health have data bases with a hierarchical structure. The studied patients (level 1) can be nested in groups, i.e., district, doctor, hospital, etc. (level 2). It is possible that patients in the same group be similar, so traditional regression models can not be used because the hypothesis of independent observations is not satisfied. A Multilevel Analysis, using hierarchical models, can be a solution for this problem; these models take into account the distribution of the data at different levels to estimate two types of variability: one due to individuals in the study and another due to the groups in which patients are nested. These types of models were applied in education in the last decade, however they have been recently applied in Health Research. This paper is a review about multilevel analysis. A discussion about hierarchichal models versus traditional regression models is presented and some applications in Epidemiology and Health Research are showed.

Full text is only aviable in PDF
Biblografía
[1.]
G. Box.
Statistics and quality improvement.
J Royal Stat Soc, Series A, 157 (1994), pp. 209-229
[2.]
H. Goldstein, D. Spiegelhalter.
League tables and their limitations: statistical issues in comparasions of institutional perfomance (with discussion).
Royal Statis Soc, Series A, 159 (1996), pp. 385-443
[3.]
C.W. Buck.
Prenatal and perinatal causes of early death and defect.
Preventive and Community Medicine, 2d ed.,
[4.]
M. Blaxter.
Health and Lifestyles.
Rooutledgr, (1990),
[5.]
V. Carstairs, R. Morris.
Deprivation: explaining differences in mortality between Scotland and England and Wayes.
BMJ, 299 (1989), pp. 886-889
[6.]
M. Gronick.
Medicare patients: regional differences in length of hospital stay, 1969–1971.
Soc Sec Bull, 40 (1975), pp. 22-41
[7.]
J.E. Wennberg, J.L. Freeman, W.J. Culp.
Are hospital services rationed in New Haven or over-utilised in Boston?.
Lancet, 337 (1987), pp. 1185-1189
[8.]
K. McPherson, J.E. Wennberg, O.E. Hovind, P. Clifford.
Small-area variation in the use of common surgical procedures: an international comparasion of New England, England and Norway.
N Engl J Med, 307 (1982), pp. 1310-1314
[9.]
D.G. Altman, S.N. Goodman.
Transfer of technology from statistical journal to biomedical literature. Past trends and future predictions.
JAMA, 272 (1994), pp. 129-132
[10.]
H. Morgenstern.
Ecologic studies in epidemiology.
Annu Rev Public Health, 16 (1995), pp. 61-81
[11.]
W.S. Robinson.
Ecological correlations and the behavior of individuals.
Am Sociol Rev, 15 (1950), pp. 351-357
[12.]
H. Goldstein.
Multilevel statistical models.
2.a ed., Arnold, (1995),
[13.]
N.T. Longford.
A fast scoring algorithm for maximum likelihood estimation in unbalanced mixed models with nested random effects.
Biometrika, 74 (1987), pp. 817-827
[14.]
P.J. Diggle, K.Y. Liang, S.L. Zeger.
Analysis of longitudinal data.
Oxford Univ Press, (1994),
[15.]
D. Clayton.
Generalized linear mixed models.
Markov Chain Monte Carlo in Practice,
[16.]
S. Richardson.
Développements récents de la biostatisque.
Rev Epidem et Santé Publ, 44 (1996), pp. 482-493
[17.]
T.A.B. Snijders, R.J. Bosker.
Model variance in two-levels models.
Sociol Methods Research, 22 (1994), pp. 342-363
[18.]
H.J. Sixma, P.M.M. Spreeuwenberg, M.A.A. van der Pasch.
Patient satisfaction in the general practitioner: a two-level analysis.
Medical Care, 36 (1998), pp. 212-229
[19.]
C.A. Gatsonis, A.M. Epstein, J.P. Newhouse, S.L. Normand, B.J. McNeil.
Variations in the utilization of coronary angiography for elderly patients with an acute myocardial infarction: an analysis using hierarchical logistic regression.
Medical Care, 33 (1995), pp. 625-642
[20.]
M.J. Daniels, C.A. Gatsonis.
Hierarchical polytomous regression models with application to health services research.
Stat Med, 16 (1997), pp. 2311-2325
[21.]
F. Berrino, M. Sant, A. Verdecchia, R. Capocaccia, T. Hakulinen, J. Estève.
Survival of cancer patients in Europe: The EUROCARE study. IARC Sci Publ No 132.
IARC, (1995),
[22.]
D.K. Stangl, J.B. Greenhouse.
Assesing placebo response using bayesian hierarchical survival models.
Lifetime Data Anal, 4 (1998), pp. 5-28
[23.]
N. Sastry.
A multilevel hazards model for hierarchically clustered data: model estimation and an application to the study of child survival in northeast Brazil.
RAND Working Paper Series, (1995), pp. 95-115
[24.]
T. Snijders.
Analysis of longitudinal data using the hierarchical linear model.
Quality and Quantity, 30 (1996), pp. 405-426
[25.]
C. Duncan, K. Jones, G. Moon.
Health-related behaviour in context: a multilevel modelling approach.
Soc Sci Med, 42 (1996), pp. 817-830
[26.]
I.H. Langford, G. Bentham.
Regional variations in mortality rates in England and Wales: an analysis using multi-level modelling.
Soc Sci Med, 42 (1996), pp. 897-908
[27.]
W.M. Mason.
Comment.
J. Eduacational and Behavioral Statistics, 20 (1995), pp. 221-227
[28.]
A.V. Diez-Roux.
Bringing context back into epidemiology: variables and fallacies in multilevel analysis.
Am J Public Health, 88 (1998), pp. 216-222
[29.]
M von Korff, T. Koepsell, S. Curry, P. Diehr.
Multi-level analysis in epidemiologic research on health behaviors and outcomes.
Am J Epidemiol, 135 (1992), pp. 1077-1082
[30.]
P. O'Campo, A.C. Gielen, R.R. Faden, X. Xue, N. Kass, R.R. Wang.
Violence by male partners against women during the childbearing year: a contextual analysis.
Am J Public Health, 85 (1995), pp. 1092-1097
[31.]
C. Duncan, K. Jones, G. Moon.
Do places matter? A multi-level analysis of regional variations in health-related behaviour in Britain.
Soc Sci Med, 37 (1993), pp. 725-733
Copyright © 1999. Sociedad Española de Salud Pública y Administración Sanitaria
Download PDF
Idiomas
Gaceta Sanitaria
Article options
Tools
es en

¿Es usted profesional sanitario apto para prescribir o dispensar medicamentos?

Are you a health professional able to prescribe or dispense drugs?