Quantal Response Regression : Ein Verfahren zur statistischen Analyse einer multifaktoriellen Genese

Die Arbeit beschreibt ein multivariates Verfahren zur statistischen Analyse des Zusammenhanges zwischen dem Eintreffen (Y=l) oder Nichteintreffen (Y=0) eines Ereignisses als binäre abhängige Variable und einem Vektor unabhängiger Variablen, die sowohl qualitativer als auch quantitativer Natur sem können. Als Resultat erhält man die Koeffizienten einer Linearkombination der unabhängigen Variablen, die ähnlich wie die partiellen Regressionskoeffizienten einer multiplen Regression zu beurteilen sind, wobei jedoch die Zielgröße die logit-transformierte Wahrscheinlichkeit P(Y=1) für das Eintreffen des Ereignisses darstellt. Die Methode eignet sich insbesondere zur Analyse des Zusammenhangs zwischen dem Auftreten einer Krankheit und den Risikofaktoren, für diese Krankheit. Dies wird anhand von zwei epidemiologischen Beispielen aus Prospektivstudien über koronare Herzkrankheiten dargestellt.

This paper describes a multivariate method for the statistical analysis of the relationship between the occurrence (Y=l) or non-occurrence (Y = 0) of an event as a binary dependent variable and a vector of independent variables. This vector may be of qualitative or quantitative character. The results are the coefficients of a linear combination of the independent variables which may be interpreted similarly to the partial regression coefficients in multiple regression. However, the response variable represents the logit-transformed probability P(Y=1) of the occurrence of the event. The method is especially suitable for the analysis of the relationship between the occurrence of an illness and the risk factors for this illness. This is demonstrated by means of two epidemiological examples out of prospective studies on coronary heart disease.

Schlüssel-Wörter

Logistische Risikofunktion - multivariate Analyse - Risikofaktoren - Epidemiologie

Key-Words

Logistic Risk Function - Multivariate Analysis - Risk Factors - Epidemiology

• Literaturverzeichnis

• 1 Armitage P. Statistical Methods in Medical Research. Blackwell: Oxford and Edinburgh; 1971
• 2 Armitage P, McPherson C. K. Statistical Studies of Prognosis in Advanced Breast Cancer. J. chron. Dis 20 1969; 343-360.
• 3 Cornfield J, Gordon T, Smith W. W. Quantal Response Curves for Experimentally Uncontrolled Variables. Bull. int. statist. Inst 38 1961; 97-115.
• 4 Coronary Drug Project Research Group. Factors Influencing Long-term Prognosis after Recovery from Myocardial Infarction — Three-year Findings of the Coronary Drug Project. J. chron. Dis 27 1974; 267-285.
• 5 Cox D. R. Analysis of Binary Data. Methuen; London: 1970
• 6 Dyer A. R, Stamler J, Berkson D. M, Lindbebg H. A. Relationship of Relative Weight and Body Mass Index to 14-Year Mortality in the Chicago Peoples Gas Company Study. J. chron. Dis 28 1975; 109-123.
• 7 Fletcher R, Powell M. J. D. A Rapidly Convergent Descent Method for Minimization. Comput. J 6 1963; 163-168.
• 8 Gordon T. Hazards in the Use of the Logistic Function with Special Reference to Data from Prospective Cardiovascular Studies. J. chron. Dis 27 1974; 97-102.
• 9 Halperin M, Blackwell W. C, Verter J. I. Estimation of the Multivariate Logistic Risk Function: A Comparison of the Discriminant Function and Maximum Likelihood Approaches. J. chron. Dis 24 1971; 125-158.
• 10 Harding le Riche W, Milner J. Epidemiology as Medical Ecology. Livingstone: Edinburgh and London; 1971
• 11 Kahn H. A, Herman J. B, Medalie J. H, Neufeld H. N, Riss E, Goldbourt U. Factors Related to Diabetes Incidence : A Multivariate Analysis of two Years Observation on 10,000 Men. J. chron. Dis 23 1971; 617-629.
• 12 Kleinbaum D. G, Kupper L. L, Cassel J. C, Tyroler H. A. Multivariate Analysis of Risk of Coronary Heart Disease in Evans County, Georgia. Arch. int. Med 128 1971; 943-948.
• 13 Truett J, Cornfield J, Kannel W. A Multivariate Analysis of the Risk of Coronary Heart Disease in Framingham. J. chron. Dis 20 1967; 511-524.
• 14 Walker S. H, Duncan D. B. Estimation of the Probability of an Event as a Function of Several Independent Variables. Biometrika 54 1967; 167-179.
• 15 Walker S. H, Duncan D. B. Report of the Committee for the Assessment of Biometric Aspects of Controlled Trials of Hypoglycemic Agents. J. Amer, med. Ass 231 1975; 582-608.