Kosten der Schizophrenie

Strukturgleichungsmodelle als Instrument der Entscheidungshilfe bei gesundheitsökonomischen Fragestellungen

verfasst von

J. Clouth, A. Uber, J. M. Graf von der Schulenburg

Abstract

Objective: The goal of the study is to bring out advantages of structural equation models over conventional methods in health economic decision-making processes. Methods: In this study, causal relations of direct and indirect costs are investigated in schizophrenic subjects. The structural equation models are based on data obtained from a health care analysis. X² test, degrees of freedom, P value, goodness of fit and adjusted goodness of fit are used to evaluate the structural equation models. Results: The structural equation model demonstrates that no less than 23 percent of direct costs for the treatment of schizophrenia are accounted for by the two factors severity of disease and time. The severity of disease carries more weight than the time factor. The standardized path coefficient is 0.33. The time factor has less effect than the severity of disease with a standardized regression coefficient of 0.27. The double arrow between the factors severity and time describes a correlation which is 0.26. Based on both inference statistical and descriptive parameters, all of which are within the nominal range, the structural equation model for direct costs can be confirmed with the data set on hand. With use of this model, 51 percent of indirect costs are accounted for by the two factors severity and time. As for indirect costs, the time factor of 0.63 carries more than twice as much weight as severity exerting effects with a standardized path coefficient of 0.25. This also explains the correlation between the residuum of the Global Assessment Scale (GAS) and the residuum of the time factor. There is a relationship between the residuum of GAS, which measures the degree of severity in the long term, and the residuum of the time factor. It may be conjectured that another factor is hiding behind this correlation, that is early retirement. In this model, too, all parameters are within the nominal range. Conclusions: The example of costs of schizophrenia has demonstrated the portrayal and quantification of networking structures. Due to the portrayal of structures and quantification of causal relationships, structural equation models provide substantial support for decision-making processes.

Externe Organisation(en)

Eli Lilly and Company

Typ

Artikel

Journal

Gesundheitsokonomie und Qualitatsmanagement

Band

8

Seiten

120-124

Anzahl der Seiten

5

ISSN

1432-2625

Publikationsdatum

04.2003

Publikationsstatus

Veröffentlicht

Peer-reviewed

Ja

ASJC Scopus Sachgebiete

Health policy

Ziele für nachhaltige Entwicklung

SDG 3 – Gute Gesundheit und Wohlergehen

Elektronische Version(en)

https://doi.org/10.1055/s-2003-39145 (Zugang: Geschlossen)