Dorota Myszkowska1, Renata Majewska2. 1. Department of Clinical and Environmental Allergology, Jagiellonian University Medical College, Cracow, Poland. 2. Chair of Epidemiology and Preventive Medicine, Jagiellonian University Medical College, Cracow, Poland.
Abstract
INTRODUCTION AND OBJECTIVES: It is important to monitor the threat of allergenic pollen during the whole season, because of practical application in allergic rhinitis treatment, especially in the specific allergen immunotherapy. The aim of the study was to propose the forecast models predicting the pollen occurrence in the defined pollen concentration categories related to the patient exposure and symptom intensity. MATERIAL AND METHODS: The study was performed in Cracow (southern Poland), pollen data were collected using the volumetric method in 1991-2012. For all independent variables (meteorological elements) and the daily pollen concentrations the running mean for periods: 2-, 3-, 4-, 5-, 6- and 7 days before the predicted day were calculated. The multinomial logistic regression was used to find the relation between the probability of the pollen concentration occurrence in the selected categories and meteorological elements and pollen concentration in days preceding the predicted daily concentration. The models were constructed for each taxon using data in 1991-2011 (without 1992 and 1996 due to missing data in these years) and 1998-2011 pollen seasons. RESULTS: The days classified among the lowest category (0-10 PG/m3) (pollen grains/m 3 of air) dominated for all the studied taxa. The percentage of the obtained predictions of the pollen occurrence fluctuated between 35-78% which is a sufficient value of model predictions. Considering the studied taxon, the best model accuracy was obtained for models forecasting Betula pollen concentration (both data series), and Poaceae (both data series). CONCLUSIONS: The application of the recommended threshold values during the predictive models construction seems to be really useful to estimate the real threat of allergen exposure. It was indicated that the polynomial logistic regression models could be a practical tool for effective forecasting in biological monitoring of pollen exposure.
INTRODUCTION AND OBJECTIVES: It is important to monitor the threat of allergenic pollen during the whole season, because of practical application in allergic rhinitis treatment, especially in the specific allergen immunotherapy. The aim of the study was to propose the forecast models predicting the pollen occurrence in the defined pollen concentration categories related to the patient exposure and symptom intensity. MATERIAL AND METHODS: The study was performed in Cracow (southern Poland), pollen data were collected using the volumetric method in 1991-2012. For all independent variables (meteorological elements) and the daily pollen concentrations the running mean for periods: 2-, 3-, 4-, 5-, 6- and 7 days before the predicted day were calculated. The multinomial logistic regression was used to find the relation between the probability of the pollen concentration occurrence in the selected categories and meteorological elements and pollen concentration in days preceding the predicted daily concentration. The models were constructed for each taxon using data in 1991-2011 (without 1992 and 1996 due to missing data in these years) and 1998-2011 pollen seasons. RESULTS: The days classified among the lowest category (0-10 PG/m3) (pollen grains/m 3 of air) dominated for all the studied taxa. The percentage of the obtained predictions of the pollen occurrence fluctuated between 35-78% which is a sufficient value of model predictions. Considering the studied taxon, the best model accuracy was obtained for models forecasting Betula pollen concentration (both data series), and Poaceae (both data series). CONCLUSIONS: The application of the recommended threshold values during the predictive models construction seems to be really useful to estimate the real threat of allergen exposure. It was indicated that the polynomial logistic regression models could be a practical tool for effective forecasting in biological monitoring of pollen exposure.
Authors: Igor Kaidashev; Halyna Morokhovets; Viktoriia Rodinkova; Lawrence DuBuske; Jean Bousquet Journal: World Allergy Organ J Date: 2020-09-30 Impact factor: 4.084
Authors: Jean Bousquet; Josep M Anto; Isabella Annesi-Maesano; Toni Dedeu; Eve Dupas; Jean-Louis Pépin; Landry Stephane Zeng Eyindanga; Sylvie Arnavielhe; Julia Ayache; Xavier Basagana; Samuel Benveniste; Nuria Calves Venturos; Hing Kin Chan; Mehdi Cheraitia; Yves Dauvilliers; Judith Garcia-Aymerich; Ingrid Jullian-Desayes; Chitra Dinesh; Daniel Laune; Jade Lu Dac; Ismael Nujurally; Giovanni Pau; Robert Picard; Xavier Rodo; Renaud Tamisier; Michael Bewick; Nils E Billo; Wienczyslawa Czarlewski; Joao Fonseca; Ludger Klimek; Oliver Pfaar; Jean-Marc Bourez Journal: Clin Transl Allergy Date: 2018-09-17 Impact factor: 5.871