Robust Trend Analysis in Predicting Histamine Provocation Concentration

Abstract

Trend analysis is used to predict the histamine provocation concentration PC₂₀ during the so-called challenge test, where increasing doses of the bronchoconstricting agent are given to a patient and the spirometric examination is subsequently performed. Classical least-squares analysis of trends as well as least absolute deviations analysis are compared. The results are interesting and promising from a statistical as well as medical point of view. The methods presented can be applied to many practical problems in medicine.

Key-Words

Histamine Test - L₁-estimation - Prediction - Provocation Concentration PC₂₀ - Trend Analysis

• REFERENCES

• 1 Pepys J, Hutchcroft BJ. Bronchial provocation tests in etiologic diagnosis and analysis of asthma. Amer Rev Resp Dis 1975; 112: 829-59.
• 2 Chai H, Farr RS, Froehlich LA. et al. Standardization of bronchial inhalation procedures. J Allergy Clin Immunol 1975; 56: 323-7.
• 3 SEPCR Working Group on Bronchial Hyperreactivity. In: Elser NM, Korrabiju KF, Quanjer PH. eds Guidelines for standardization of bronchial challenges with nonspecific bronchoconstricting agents. Bull Europ Physiopath Resp 1983; 19: 495-514.
• 4 Krusińska E, Liebhart J, Woźniak M. A trial of predicting the histamine provocation dose during challenge test. In: Abstracts of the 10th International Meeting on Clinical Biostatistics. Maastricht; The Netherlands: September 1989: 74.
• 5 Liebhart J, Krusińska E. Analysis of FEV₁ changes during challenge test as the First step for constructing subsequent histamine concentration model. In: Scientific Contributions of the Institute of Electrical Metrology, Technical University of Wroclaw Series: Conferences No 14, 1988. 31 69-73. (in Polish).
• 6 Späth H. Clusterwise linear least squares versus linear absolute deviations regression: a numerical comparison for a case study. In: Gaul W, Schader M. eds. Classification as a Tool of Research. Amsterdam: North-Holland; 1986: 413-22.
• 7 Kamburoff PF, Woitowicz HI, Woitowicz RH. Prediction of spirometric indices. Resp News Bull 1973; 17: 1.
• 8 Winer BJ. Statistical Principles in Experimental Design. London: McGraw Hill; 1970
• 9 Plackett RL. Regression Analysis. Oxford: 1960
• 10 Ralston A. A First Course in Numerical Analysis. New York: McGraw Hill; 1965
• 11 Antoch J, Bartkowiak A, Pekalska J. Computing L₁-norm, α-trimmed and a-win-sorized regression on the ODRA 1305 computer. Report N-159, Institute of Computer Science; University of Wroclaw: March 1986
• 12 Barrodale J, Roberts FDK. An improved algorithm for discrete L₁norm approximation. SIAM J Num Anal 1973; 10: 839-48.
• 13 Barrodale I, Roberts FDK. Solution of overdetermined system of equations in L₁-norm. CADM 1974; 19: 319-20.
• 14 Rao CR. Linear Statistical Inference and its Applications. New York: Wiley & Sons; 1965
• 15 Liebhart J, Karkowski Z, Liebhart E, Mafolepszy J. Error characterization in spirometry and an attempt of classification. Probl Techn Med 1986; 17: 113-23. (in Polish).
• 16 Senn SJ. The use of baselines in clinical trials of bronchodilators. Statist in Med 1989; 08: 1339-50.
• 17 Senn SJ. Oral discussion to [4]. The 10th International Meeting on Clinical Biostatistics. Maastricht; The Netherlands: September 1989