gms | German Medical Science

50. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds)
12. Jahrestagung der Deutschen Arbeitsgemeinschaft für Epidemiologie (dae)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie
Deutsche Arbeitsgemeinschaft für Epidemiologie

12. bis 15.09.2005, Freiburg im Breisgau

Distributions of health care utilization in childhood and adolescence

Meeting Abstract

  • Andrea Stroux - Institut für Medizinische Informatik, Biometrie und Epidemiologie, Charité Berlin - CBF, Deutschland
  • D. Staab - Lehrstuhl für Computerlinguistik, Universität Heidelberg, Deutschland
  • M. Demleitner - Klinik für Pädiatrie, Charite Berlin - CVK, Deutschland
  • B. Falk - Lehrstuhl für Computerlinguistik, Universität Heidelberg, Deutschland
  • U. Grittner - Institut für Medizinische Informatik, Biometrie und Epidemiologie, Charité Berlin - CBF, Deutschland
  • S. Oberwöhrmann - Lehrstuhl für Computerlinguistik, Universität Heidelberg, Deutschland
  • P. Martus - Institut für Medizinische Informatik, Biometrie und Epidemiologie, Charité Berlin - CBF, Deutschland

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. Deutsche Arbeitsgemeinschaft für Epidemiologie. 50. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds), 12. Jahrestagung der Deutschen Arbeitsgemeinschaft für Epidemiologie. Freiburg im Breisgau, 12.-15.09.2005. Düsseldorf, Köln: German Medical Science; 2005. Doc05gmds423

Die elektronische Version dieses Artikels ist vollständig und ist verfügbar unter: http://www.egms.de/de/meetings/gmds2005/05gmds252.shtml

Veröffentlicht: 8. September 2005

© 2005 Stroux et al.
Dieser Artikel ist ein Open Access-Artikel und steht unter den Creative Commons Lizenzbedingungen (http://creativecommons.org/licenses/by-nc-nd/3.0/deed.de). Er darf vervielf&aauml;ltigt, verbreitet und &oauml;ffentlich zug&aauml;nglich gemacht werden, vorausgesetzt dass Autor und Quelle genannt werden.


Gliederung

Text

Introduction

Little is known about the distribution of health care utilization within the group of children and adoslescents, and about which medical and psycho-sociological factors are associated with high or low utilization. Often, the definition of ”high utilizers” in health care is based on percentiles or arbitrarily fixed numbers of medical consultations, and the underlying distributions of health care utilization are ignored.

Material and Methods

In the initial stage of the BMBF project 01GL0010 ”Identification and characterization of children and adolescents with high health care system utilization in Berlin, and health political recommendations”, panel doctor services‘ data from 255126 individuals up to 18 years old (7.7 ± 4.8 years; 48.9% girls) out of Berlin and surroundings have been investigated with respect to frequencies of medical consultation during one year (3rd quarter 2000 to 2nd quarter 2001). Distributions of these frequencies were identified separately for several age groups using finite mixture analysis. Only those cases were included into the analyses who were seen by a doctor at least once a year, leading to truncated distribution models.

Results

Depending on age group, distributions were identified as mixtures of geometric and/or zero-truncated poisson distributions. In the lower age groups up to eleven years, health care utilization was characterized by two qualitatively different distributions (geometric and truncated poisson), which did not allow for differentiation into “high” and “low” utilizers. In the age groups from twelve years up, mixtures of two geometric distributions could be identified, indicating different populations of health care utilization among adoslescents. Given these distributions, ROC-curves can be determined which allow to quantify the accuracy of classification into normal or high utilizer group, depending on the chosen cutpoint (number of medical consultations per year) (Figure 1 [Fig. 1]).

Discussion

In early childhood, no evidence for seperate groups of “high” and “low” utilizers could be found, for they seem to originate from the same (geometrically distributed) population.

If appropriate, as in the group of children from twelve years of age up, a distribution-based definition of health care utilization allows to specify ”diagnostic accuracy” for the identification of high utilizers in terms of sensitivity and specificity.

Thus, finite mixture analysis may be an appropriate alternative to arbitrarily fixing percentiles or numbers of medical consultations for the definition of “high utilizers”.


References

1.
Leu RE. Determinanten der Nachfrage nach medizinischen Leistungen. In: Brennecke R, Schach E, eds. Ambulante Versorgung: Nachfrage und Steuerung. Berlin: Springer; 1987.
2.
Smith RC, Gardiner JC, Armatti S, et al. Screening for high utilizing somatizing patients using a prediction rule derived from the management information system of an HMO: a preliminary study. Med Care 2001;39(9):968-78
3.
Titterington D, Smith A, Markov U. Statistical Analysis of Finite Dristibutions. New York: John Wiley; 1985