Publicatie
Publication date
Are influenza surveillance data useful for mapping presentations?
Uphoff, H., Stalleicken, I., Bartelds, A., Phiesel, B., Kistemann, B.T. Are influenza surveillance data useful for mapping presentations? Virus Research: 2004, 103(1-2), 35-46
Download the PDF
Geographical information system (GIS) based on mappings of influenza data are rare
(http://www.b3e.jussieu.fr.80/sentiweb/fr) and influenza data are commonly aggregated for rather large areas (http://www.eiss.org, http://oms2b3e.jussieu.fr/FluNet). The most limiting factors for the use of morbidity-data from practices in GIS-based mappings are differences which are not related to morbidity. These differences may be due to consultation behaviour, interpretation of the case definition, age distribution of patients and other reasons. In order to reduce the impact of these non-morbidity related differences on the interpretation, the data of many practices are usually pooled and consequently rather large areas are presented. Extracting and harmonising the signals for increased morbidity from practices is a presupposition for mapping with a sufficient geographical resolution. The possibility to harmonise by reducing those confounding differences on a practice level is investigated. Different harmonisation methods were applied to data from Germany where acute respiratory infections (ARI) per consultations are registered and from The Netherlands were influenza like illnesses (ILI) per population are registered. The harmonisation of the indices between countries was achieved by scaling them in relation to the level of the index representative for the peak activity during a usual influenza epidemic. The Kriging method is applied as a means of spatial prediction for the influenza data. The preliminary results are discussed with respect to resulting mappings. (aut. ref.)
(http://www.b3e.jussieu.fr.80/sentiweb/fr) and influenza data are commonly aggregated for rather large areas (http://www.eiss.org, http://oms2b3e.jussieu.fr/FluNet). The most limiting factors for the use of morbidity-data from practices in GIS-based mappings are differences which are not related to morbidity. These differences may be due to consultation behaviour, interpretation of the case definition, age distribution of patients and other reasons. In order to reduce the impact of these non-morbidity related differences on the interpretation, the data of many practices are usually pooled and consequently rather large areas are presented. Extracting and harmonising the signals for increased morbidity from practices is a presupposition for mapping with a sufficient geographical resolution. The possibility to harmonise by reducing those confounding differences on a practice level is investigated. Different harmonisation methods were applied to data from Germany where acute respiratory infections (ARI) per consultations are registered and from The Netherlands were influenza like illnesses (ILI) per population are registered. The harmonisation of the indices between countries was achieved by scaling them in relation to the level of the index representative for the peak activity during a usual influenza epidemic. The Kriging method is applied as a means of spatial prediction for the influenza data. The preliminary results are discussed with respect to resulting mappings. (aut. ref.)
Geographical information system (GIS) based on mappings of influenza data are rare
(http://www.b3e.jussieu.fr.80/sentiweb/fr) and influenza data are commonly aggregated for rather large areas (http://www.eiss.org, http://oms2b3e.jussieu.fr/FluNet). The most limiting factors for the use of morbidity-data from practices in GIS-based mappings are differences which are not related to morbidity. These differences may be due to consultation behaviour, interpretation of the case definition, age distribution of patients and other reasons. In order to reduce the impact of these non-morbidity related differences on the interpretation, the data of many practices are usually pooled and consequently rather large areas are presented. Extracting and harmonising the signals for increased morbidity from practices is a presupposition for mapping with a sufficient geographical resolution. The possibility to harmonise by reducing those confounding differences on a practice level is investigated. Different harmonisation methods were applied to data from Germany where acute respiratory infections (ARI) per consultations are registered and from The Netherlands were influenza like illnesses (ILI) per population are registered. The harmonisation of the indices between countries was achieved by scaling them in relation to the level of the index representative for the peak activity during a usual influenza epidemic. The Kriging method is applied as a means of spatial prediction for the influenza data. The preliminary results are discussed with respect to resulting mappings. (aut. ref.)
(http://www.b3e.jussieu.fr.80/sentiweb/fr) and influenza data are commonly aggregated for rather large areas (http://www.eiss.org, http://oms2b3e.jussieu.fr/FluNet). The most limiting factors for the use of morbidity-data from practices in GIS-based mappings are differences which are not related to morbidity. These differences may be due to consultation behaviour, interpretation of the case definition, age distribution of patients and other reasons. In order to reduce the impact of these non-morbidity related differences on the interpretation, the data of many practices are usually pooled and consequently rather large areas are presented. Extracting and harmonising the signals for increased morbidity from practices is a presupposition for mapping with a sufficient geographical resolution. The possibility to harmonise by reducing those confounding differences on a practice level is investigated. Different harmonisation methods were applied to data from Germany where acute respiratory infections (ARI) per consultations are registered and from The Netherlands were influenza like illnesses (ILI) per population are registered. The harmonisation of the indices between countries was achieved by scaling them in relation to the level of the index representative for the peak activity during a usual influenza epidemic. The Kriging method is applied as a means of spatial prediction for the influenza data. The preliminary results are discussed with respect to resulting mappings. (aut. ref.)