e risk of exposure to Lyme disease in the province of Trento, Italian Alps, was predicted through the analysis of the distribution of Ixodes ricinus (L.) nymphs infected with Borrelia burgdorferi s.l. with a model based on bootstrap aggregation (bagging) of tree-based classifiers within a Geographical Information System. Data on I. ricinus density assessed by dragging the vegetation in 438 sites during 1996 were cross-correlated with the digital cartography of a GIS which included the variables altitude, exposure and slope, substratum, vegetation type and roe deer density. Ticks resulted more abundant at altitudes below 1,300 m a.s.l., in the presence of limestone and vegetation cover with thermophile deciduous forests with high densities of roe deer. A bootstrap aggregation procedure (bagging) was used to produce a model for the prediction of tick occurrence, which accuracy was tested on actual tick counts assessed by a further dragging campaign carried out during 1997 to determine infection prevalence and resulted in average 77%. Other tests of the model were made on other additional and independent data set. The prevalence of infection with Borrelia burgdorferi s.l, determined by PCR on 2,208 nymphs collected by dragging the vegetation during 1997 in 245 transects selected randomly within 8 areas where I. ricinus was predicted to occur by the model, was 17.5% and was positively correlated to tick abundance and roe deer density. These findings were used to relate the output of the bagged model (probability of tick occurrence) to the density of infected nymphs through a stepwise model selection procedure and thus to produce a GIS digital map of the probability distribution of infected nymphs in the Province of Trento at high resolution scale (50 m x 50 m cell resolution).The application of the bagging procedure increased the accuracy of the prediction made by a single classification tree, a well known classification method for the analysis of epidemiological data. Keywords: Ixodes ricinus, Lyme disease, GIS, Bagging, Tree-based Classifiers, risk prediction. Lyme disease is the most common bacterial infection transmitted by ticks in the boreal hemisphere (Gray et al. 1998b). In Europe, the tick Ixodes ricinus (L.) acts as the main vector of infection which is maintained by a series of competent reservoirs, at least 9 species of small mammals, 7 species of medium-sized mammals and 16 species of birds (Gern et al. 1998, Gray 1998a, Humair et al. 1999). In Italy, Lyme disease was first recorded in 1983. The causative agent Borrelia burgdorferi sensu lato (s.l.) and the genospecies recognized as pathogenic for humans, Borrelia burgdorferi sensu stricto (s.s.), Borrelia afzelii and Borrelia garinii (Johnson et al. 1984, Baranton et al. 1992, Canica et al. 1993) were isolated from humans and the tick I. ricinus (Burioni et al. 1988, Genchi et al. 1994, Bandi et al. 1997, Ciceroni and Ciarrocchi 1998, Cinco et al. 1998). Currently, the disease is endemic in at least six regions of central and north-eastern Italy (Liguria, Emilia- Romagna, Friuli Venezia Giulia, Veneto and Trentino-Alto Adige) with 1,171 cases recorded between 1986 and 1997, though this number is considerably underestimated (Pavan et al. 2000). The incidence rate of Lyme disease in the Trentino- Alto Adige region, one of the "hot-spots" of infection, was 8.02 cases/100,000 inhabitants estimated for the period 1986-1997 (Pavan et al. 2000). The mesoscale distribution of I. ricinus in the Province of Trento was determined by dragging the vegetation and by screening roe deer shot by hunters (Chemini et al. 1993, 1997). Environmental variables including altitude, exposure, vegetation type and geological substratum were used to develop a tree-based model for the assessment of the habitat patterns with the highest probability of tick occurrence (Merler et al. 1996). The model showed that altitude and geological substratum were the most impor...

Geographical Information Systems and Bootstrap Aggregation (Bagging) of Tree-Based Classifiers for Lyme Disease Risk Assessment in Trentino, Italian Alps

Rizzoli, Annamaria;Merler, Stefano;Furlanello, Cesare;
2002-01-01

Abstract

e risk of exposure to Lyme disease in the province of Trento, Italian Alps, was predicted through the analysis of the distribution of Ixodes ricinus (L.) nymphs infected with Borrelia burgdorferi s.l. with a model based on bootstrap aggregation (bagging) of tree-based classifiers within a Geographical Information System. Data on I. ricinus density assessed by dragging the vegetation in 438 sites during 1996 were cross-correlated with the digital cartography of a GIS which included the variables altitude, exposure and slope, substratum, vegetation type and roe deer density. Ticks resulted more abundant at altitudes below 1,300 m a.s.l., in the presence of limestone and vegetation cover with thermophile deciduous forests with high densities of roe deer. A bootstrap aggregation procedure (bagging) was used to produce a model for the prediction of tick occurrence, which accuracy was tested on actual tick counts assessed by a further dragging campaign carried out during 1997 to determine infection prevalence and resulted in average 77%. Other tests of the model were made on other additional and independent data set. The prevalence of infection with Borrelia burgdorferi s.l, determined by PCR on 2,208 nymphs collected by dragging the vegetation during 1997 in 245 transects selected randomly within 8 areas where I. ricinus was predicted to occur by the model, was 17.5% and was positively correlated to tick abundance and roe deer density. These findings were used to relate the output of the bagged model (probability of tick occurrence) to the density of infected nymphs through a stepwise model selection procedure and thus to produce a GIS digital map of the probability distribution of infected nymphs in the Province of Trento at high resolution scale (50 m x 50 m cell resolution).The application of the bagging procedure increased the accuracy of the prediction made by a single classification tree, a well known classification method for the analysis of epidemiological data. Keywords: Ixodes ricinus, Lyme disease, GIS, Bagging, Tree-based Classifiers, risk prediction. Lyme disease is the most common bacterial infection transmitted by ticks in the boreal hemisphere (Gray et al. 1998b). In Europe, the tick Ixodes ricinus (L.) acts as the main vector of infection which is maintained by a series of competent reservoirs, at least 9 species of small mammals, 7 species of medium-sized mammals and 16 species of birds (Gern et al. 1998, Gray 1998a, Humair et al. 1999). In Italy, Lyme disease was first recorded in 1983. The causative agent Borrelia burgdorferi sensu lato (s.l.) and the genospecies recognized as pathogenic for humans, Borrelia burgdorferi sensu stricto (s.s.), Borrelia afzelii and Borrelia garinii (Johnson et al. 1984, Baranton et al. 1992, Canica et al. 1993) were isolated from humans and the tick I. ricinus (Burioni et al. 1988, Genchi et al. 1994, Bandi et al. 1997, Ciceroni and Ciarrocchi 1998, Cinco et al. 1998). Currently, the disease is endemic in at least six regions of central and north-eastern Italy (Liguria, Emilia- Romagna, Friuli Venezia Giulia, Veneto and Trentino-Alto Adige) with 1,171 cases recorded between 1986 and 1997, though this number is considerably underestimated (Pavan et al. 2000). The incidence rate of Lyme disease in the Trentino- Alto Adige region, one of the "hot-spots" of infection, was 8.02 cases/100,000 inhabitants estimated for the period 1986-1997 (Pavan et al. 2000). The mesoscale distribution of I. ricinus in the Province of Trento was determined by dragging the vegetation and by screening roe deer shot by hunters (Chemini et al. 1993, 1997). Environmental variables including altitude, exposure, vegetation type and geological substratum were used to develop a tree-based model for the assessment of the habitat patterns with the highest probability of tick occurrence (Merler et al. 1996). The model showed that altitude and geological substratum were the most impor...
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11582/67
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
social impact