Modelling tick bite risk by combining random forests and count data regression models

English version

Autoři: Irene Garcia-Marti ^aff001; Raul Zurita-Milla ^aff001; Arno Swart ^aff002
Působiště autorů: Department of Geo-Information Processing, Faculty of Geo-Information and Earth Observation (ITC), University of Twente, Enschede, the Netherlands ^aff001; Centre for Infectious Disease Control, National Institute for Public Health and the Environment (RIVM), Bilthoven, the Netherlands ^aff002
Vyšlo v časopise: PLoS ONE 14(12)
Kategorie: Research Article
prolekare.web.journal.doi_sk: https://doi.org/10.1371/journal.pone.0216511

Souhrn

The socio-economic and demographic changes that occurred over the past 50 years have dramatically expanded urban areas around the globe, thus bringing urban settlers in closer contact with nature. Ticks have trespassed the limits of forests and grasslands to start inhabiting green spaces within metropolitan areas. Hence, the transmission of pathogens causing tick-borne diseases is an important threat to public health. Using volunteered tick bite reports collected by two Dutch initiatives, here we present a method to model tick bite risk using human exposure and tick hazard predictors. Our method represents a step forward in risk modelling, since we combine a well-known ensemble learning method, Random Forest, with four count data models of the (zero-inflated) Poisson family. This combination allows us to better model the disproportions inherent in the volunteered tick bite reports. Unlike canonical machine learning models, our method can capture the overdispersion or zero-inflation inherent in data, thus yielding tick bite risk predictions that resemble the original signal captured by volunteers. Mapping model predictions enables a visual inspection of the spatial patterns of tick bite risk in the Netherlands. The Veluwe national park and the Utrechtse Heuvelrug forest, which are large forest-urban interfaces with several cities, are areas with high tick bite risk. This is expected, since these are popular places for recreation and tick activity is high in forests. However, our model can also predict high risk in less-intensively visited recreational areas, such as the patchy forests in the northeast of the country, the natural areas along the coastline, or some of the Frisian Islands. Our model could help public health specialists to design mitigation strategies for tick-borne diseases, and to target risky areas with awareness and prevention campaigns.

Klíčová slova:

Public and occupational health – Forests – Machine learning algorithms – Machine learning – Land use – Netherlands – Decision trees – Statistical distributions

Zdroje

1. EEA EEA. The impacts of urban sprawl. 2006.

2. EEA EEA. Landscape Fragmentation in Europe. IlpoeUni-StuttgartDe. 2011.

3. Tack W. Impact of Forest Conversion on the Abundance of Ixodes Ricinus Ticks. Department of Forest and Water Management, Department of Biomedical Sciences. Ghent University. 2013.

4. Uspensky IV. Blood-sucking ticks (Acarina, Ixodoidea) as an essential component of the urban environment. Entomol Rev. 2017;97: 941–969. doi: 10.1134/S0013873817070107

5. Allan BF, Keesing F, Ostfeld RS. Effect of Forest Fragmentation on Lyme Disease Risk. Conserv Biol. 2003;17: 267–272. doi: 10.1046/j.1523-1739.2003.01260.x

6. Mulder S, van Vliet AJH, Bron WA, Gassner F, Takken W. High risk of tick bites in Dutch gardens. Vector Borne Zoonotic Dis. 2013;13: 865–871. doi: 10.1089/vbz.2012.1194 24107214

7. Hall JL, Alpers K, Bown KJ, Martin SJ, Birtles RJ. Use of Mass-Participation Outdoor Events to Assess Human Exposure to Tickborne Pathogens. 2017;23: 463–467.

8. Sandifer PA, Sutton-grier AE, Ward BP. Exploring connections among nature, biodiversity, ecosystem services, and human health and well-being: Opportunities to enhance health and biodiversity conservation $. Ecosyst Serv. 2015;12: 1–15. doi: 10.1016/j.ecoser.2014.12.007

9. Ehrmann S, Liira J, Gärtner S, Hansen K, Brunet J, Cousins SAO, et al. Environmental drivers of Ixodes ricinus abundance in forest fragments of rural European landscapes. BMC Ecol. 2017; 1–14.

10. Oechslin CP, Heutschi D, Lenz N, Tischhauser W, Péter O, Rais O, et al. Prevalence of tick-borne pathogens in questing Ixodes ricinus ticks in urban and suburban areas of Switzerland. Parasites and Vectors. 2017;10: 1–18.

11. Olivieri E, Gazzonis AL, Zanzani SA, Veronesi F, Manfredi MT. Seasonal dynamics of adult Dermacentor reticulatus in a peri-urban park in southern Europe. Ticks Tick Borne Dis. 2017;8: 772–779. doi: 10.1016/j.ttbdis.2017.06.002 28647128

12. Didyk YM, Blaňárová L, Pogrebnyak S, Akimov I, Peťko B, Víchová B. Emergence of tick-borne pathogens (Borrelia burgdorferi sensu lato, Anaplasma phagocytophilum, Ricketsia raoultii and Babesia microti) in the Kyiv urban parks, Ukraine. Ticks Tick Borne Dis. 2017;8: 219–225. doi: 10.1016/j.ttbdis.2016.10.002 27923669

13. Kowalec M, Szewczyk T, Welc-Falȩciak R, Siński E, Karbowiak G, Bajer A. Ticks and the city—Are there any differences between city parks and natural forests in terms of tick abundance and prevalence of spirochaetes? Parasites and Vectors. 2017;10: 1–19.

14. Santos AS, de Bruin A, Veloso AR, Marques C, Pereira da Fonseca I, de Sousa R, et al. Detection of Anaplasma phagocytophilum, Candidatus Neoehrlichia sp., Coxiella burnetii and Rickettsia spp. in questing ticks from a recreational park, Portugal. Ticks Tick Borne Dis. 2018;9: 1555–1564. doi: 10.1016/j.ttbdis.2018.07.010 30097348

15. Paul REL, Cote M, Le Naour E, Bonnet SI. Environmental factors influencing tick densities over seven years in a French suburban forest. Parasites and Vectors. 2016;9: 1–10.

16. Szekeres S, Docters van Leeuwen A, Tóth E, Majoros G, Sprong H, Földvári G. Road-killed mammals provide insight into tick-borne bacterial pathogen communities within urban habitats. Transbound Emerg Dis. 2018. doi: 10.1111/tbed.13019 30230270

17. Szekeres S, Docters van Leeuwen A, Rigó K, Jablonszky M, Majoros G, Sprong H, et al. Prevalence and diversity of human pathogenic rickettsiae in urban versus rural habitats, Hungary. Exp Appl Acarol. 2016;68: 223–226. doi: 10.1007/s10493-015-9989-x 26613759

18. Kiewra D, Stefańska-Krzaczek E, Szymanowski M, Szczepańska A. Local-scale spatio-temporal distribution of questing Ixodes ricinus L. (Acari: Ixodidae)-A case study from a riparian urban forest in Wrocław, SW Poland. Ticks Tick Borne Dis. 2017;8: 362–369. doi: 10.1016/j.ttbdis.2016.12.011 28089124

19. LoGiudice K, Ostfeld RS, Schmidt K a, Keesing F. The ecology of infectious disease: effects of host diversity and community composition on Lyme disease risk. Proc Natl Acad Sci U S A. 2003;100: 567–71. doi: 10.1073/pnas.0233733100 12525705

20. Eisen RJ, Eisen L, Girard YA, Fedorova N, Mun J, Slikas B, et al. A spatially-explicit model of acarological risk of exposure to Borrelia burgdorferi-infected Ixodes pacificus nymphs in northwestern California based on woodland type, temperature, and water vapor. Ticks Tick Borne Dis. 2010;1: 35–43. doi: 10.1016/j.ttbdis.2009.12.002 20532183

21. Gassner F, Hansford KM, Medlock J. Greener cities, a wild card for ticks? In: Braks MA, van Wieren SE, Takken W, Sprong H, editors. Ecology and prevention of Lyme borreliosis. Wageningen Academic Publishers; 2016. pp. 187–203.

22. De Keukeleire M, Vanwambeke SO, Somassè E, Kabamba B, Luyasu V, Robert A. Scouts, forests, and ticks: Impact of landscapes on human-tick contacts. Ticks Tick Borne Dis. 2015;6: 636–644. doi: 10.1016/j.ttbdis.2015.05.008 26055232

23. Zeimes C, Olsson GE, Hjertqvist M, Vanwambeke SO. Shaping zoonosis risk: landscape ecology vs. landscape attractiveness for people, the case of tick- borne encephalitis in Sweden. 2014; 1–10.

24. Garcia-Marti I, Zurita-Milla R, Harms MG, Swart A. Using volunteered observations to map human exposure to ticks. Sci Rep. 2018;8: 15435. doi: 10.1038/s41598-018-33900-2 30337654

25. Senaratne H, Mobasheri A, Ali AL, Capineri C, Muki Haklay M. A review of volunteered geographic information quality assessment methods. Int J Geogr Inf Sci. 2017;31: 139–167. doi: 10.1080/13658816.2016.1189556

26. Mehdipoor H, Zurita-Milla R, Augustijn E-W, van Vliet AJH. Checking the Consistency of Volunteered Phenological Observations While Analysing Their Synchrony. Int J Geo-Information. 2018;7: 22. doi: 10.3390/ijgi7120487

27. Krawczyk B. Learning from imbalanced data: open challenges and future directions. Prog Artif Intell. 2016;5: 221–232. doi: 10.1007/s13748-016-0094-0

28. Lee SK, Jin S. Decision tree approaches for zero-inflated count data. J Appl Stat. 2006;33: 853–865. doi: 10.1080/02664760600743613

29. Breiman L. Random Forests. Mach Learn. 2001;45: 5–32.

30. Braks M, van Wieren S, Takken W, Sprong H. Ecology and prevention of Lyme borreliosis. Wageningen Academic Publishers; 2016. http://dx.doi.org/10.3920/978-90-8686-838-4

31. Burgdorfer W, Barbour A, Hayes S, Benach J, Grunwaldt E, Davis J. Lyme Disease: a tick-borne spirochetosis? Science (80-). 1982;18: 1317–1319. doi: 10.1126/science.7043737 7043737

32. Falco RC, Fish D. Potential for Exposure to Tick Bites in Recreational Parks in a Lyme Disease Endemic Area. 1989;79. doi: 10.2105/AJPH.79.1.12 2909174

33. Magnarelli LA, Denicola A, Stafford KC, Anderson JF. Borrelia burgdorferi in an urban environment: White-tailed deer with infected ticks and antibodies. J Clin Microbiol. 1995;33: 541–544. 7751354

34. Ostfeld R. Lyme Disease: the ecology of a complex system. 1st Editio. New York, New York, USA: Oxford University Press; 2012.

35. Garcia-Martí I, Zurita-Milla R, van Vliet AJH, Takken W. Modelling and mapping tick dynamics using volunteered observations. Int J Health Geogr. 2017;16. doi: 10.1186/s12942-017-0114-8 29137670

36. Ostfeld R, Canham C, Oggenfuss K, Winchcombe R, Keesing F. Climate, deer, rodents, and acorns as determinants of variation in lyme-disease risk. PLoS Biol. 2006;4: e145. doi: 10.1371/journal.pbio.0040145 16669698

37. Randolph SE, Storey K. Impact of Microclimate on Immature Tick-Rodent Host Interactions (Acari: Ixodidae): Implications for Parasite Transmission. 1999; 741–748.

38. Kelly D, Koenig WD, Liebhold AM. An intercontinental comparison of the dynamic behavior of mast seeding communities. Popul Ecol. 2008;50: 329–342. doi: 10.1007/s10144-008-0114-4

39. Buonaccorsi JP, Elkinton J, Koenig W, Duncan RP, Kelly D, Sork V. Measuring mast seeding behavior: relationships among population variation, individual variation and synchrony. J Theor Biol. 2003;224: 107–114. doi: 10.1016/s0022-5193(03)00148-6 12900208

40. Tack W, Madder M, Baeten L, De Frenne P, Verheyen K. The abundance of Ixodes ricinus ticks depends on tree species composition and shrub cover. Parasitology. 2012;139: 1273–81. doi: 10.1017/S0031182012000625 22717041

41. Berger KA, Ginsberg HS, Gonzalez L, Mather TN. Relative humidity and activity patterns of Ixodes scapularis (Acari: Ixodidae). J Med Entomol. 2014;51: 769–76. Available: http://www.ncbi.nlm.nih.gov/pubmed/25118408

42. Estrada-Peña A, de la Fuente J, Latapia T, Ortega C. The Impact of Climate Trends on a Tick Affecting Public Health: A Retrospective Modeling Approach for Hyalomma marginatum (Ixodidae). PLoS One. 2015;10: e0125760. doi: 10.1371/journal.pone.0125760 25955315

43. Swart A, Ibañez-Justicia A, Buijs J, van Wieren SE, Hofmeester TR, Sprong H, et al. Predicting Tick Presence by Environmental Risk Mapping. Front Public Heal. 2014;2: 1–8. doi: 10.3389/fpubh.2014.00238 25505781

44. Bennet L, Halling a, Berglund J. Increased incidence of Lyme borreliosis in southern Sweden following mild winters and during warm, humid summers. Eur J Clin Microbiol Infect Dis. 2006;25: 426–32. doi: 10.1007/s10096-006-0167-2 16810531

45. Zeman P, Benes C. Peri-urbanisation, counter-urbanisation, and an extension of residential exposure to ticks: A clue to the trends in Lyme borreliosis incidence in the Czech Republic? Ticks Tick Borne Dis. 2014;5: 907–916. doi: 10.1016/j.ttbdis.2014.07.006 25113985

46. Padgett KA, Bonilla DL. Novel exposure sites for nymphal Ixodes pacificus within picnic areas. Ticks Tick Borne Dis. 2011;2: 191–195. doi: 10.1016/j.ttbdis.2011.07.002 22108011

47. Hahn MB, Bjork JKH, Neitzel DF, Dorr FM, Whitemarsh T, Boegler KA, et al. Evaluating acarological risk for exposure to Ixodes scapularis and Ixodes scapularis-borne pathogens in recreational and residential settings in Washington County, Minnesota. Ticks Tick Borne Dis. 2017; 0–1. doi: 10.1016/j.ttbdis.2017.11.010 29195857

48. Zeman P, Benes C, Markvart K. Increasing Residential Proximity of Lyme Borreliosis Cases to High-Risk Habitats: A Retrospective Study in Central Bohemia, the Czech Republic, 1987–2010. Ecohealth. 2015;12: 519–522. doi: 10.1007/s10393-015-1016-5 25698296

49. Li S, Colson V, Lejeune P, Vanwambeke SO. On the distance travelled for woodland leisure via different transport modes in Wallonia, south Belgium. Urban For Urban Green. 2016;15: 123–132. doi: 10.1016/j.ufug.2015.12.007

50. Lambin EF, Tran A, Vanwambeke SO, Linard C, Soti V. Pathogenic landscapes: interactions between land, people, disease vectors, and their animal hosts. Int J Health Geogr. 2010;9: 54. doi: 10.1186/1476-072X-9-54 20979609

51. Nielsen AB, Heyman E, Richnau G. Liked, disliked and unseen forest attributes: Relation to modes of viewing and cognitive constructs. J Environ Manage. 2012;113: 456–466. doi: 10.1016/j.jenvman.2012.10.014 23122619

52. Hadiji F, Molina A, Natarajan S, Kersting K. Poisson Dependency Networks: Gradient Boosted Models for Multivariate Count Data. Mach Learn. 2015;100: 477–507. doi: 10.1007/s10994-015-5506-z

53. Louppe G, Wehenkel L, Sutera A, Geurts P. Understanding variable importances in forests of randomized trees. Neural Inf Process Syst. 2013; 1–9. NIPS2013_4928

54. Breiman L. Bagging predictors. Mach Learn. 1996;24: 123–140. doi: 10.1007/BF00058655

55. Ho TK. The random subspace method for constructing decision forests. IEEE Trans Pattern Anal Mach Intell. 1998;20: 832–844. doi: 10.1109/34.709601

56. Japkowicz N, Stephen S. The class imbalance problem: a systematic study. Intell Data Anal. 2002; 426–449.

57. He H, Garcia EA. Learning from imbalanced data. IEEE Trans Knowl Data Eng. 2009;21: 1263–1284.

58. Lambert D. Zero-inflated poisson regression, with an application to defects in manufacturing. Technometrics. 1992;34: 1–14.

59. Greene WH. Accounting for Excess Zeros and Sample Selection in Poisson and Negative Binomial Regression Models. Biol Philos. 1994;9: 265–265.

60. Stefanoff P, Rubikowska B, Bratkowski J, Ustrnul Z, Vanwambeke SO, Rosinska M. A predictive model has identified tick-borne encephalitis high-risk areas in regions where no caseswere reported previously, Poland, 1999–2012. Int J Environ Res Public Health. 2018;15: 1–17. doi: 10.3390/ijerph15040677 29617333

61. Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al. Scikit-learn: Machine Learning in Python. J Mach Learn Res. 2011;12: 2825–2830. Available: http://scikit-learn.sourceforge.net.

62. Oliphant TE. Guide to NumPy. 2006. p. 371. http://www.numpy.org/

63. Seabold S, Perktold J. Statsmodels: Econometric and Statistical Modeling with Python. PROC 9th PYTHON Sci CONF. 2010; 57.

64. GDAL Development Team. GDAL Geospatial Data Abstraction Library: Version 2.1.0, Open Source Geospatial Foundation. Open Source Geospatial Foundation; 2018. http://gdal.osgeo.org/

65. Met Office UK. Cartopy: a cartographic python library with a matplotlib interface. Exeter, Devon; 2010. http://scitools.org.uk/cartopy

66. Hunter JD. Matplotlib: A 2D graphics environment. Computing in Science and Engineering. IEEE COMPUTER SOC; 2007. pp. 90–95.

67. Oliphant TE. Python for scientific computing. Comput Sci Eng. 2007; 10–20. doi: 10.1109/MCSE.2007.58

68. Crommentuijn LEM, Farjon JMJ, den Dekker C, van der Wulp N. Belevingswaardenmonitor Nota Ruimte 2006: Nulmeting landschap en groen in en om de stad. Bilthoven; 2007.

69. Roos-Klein Lankhorst J, de Vries S, Buijs AE, Bloemmen MHI, Schuiling C. BelevingsGIS versie 2: waardering van het Nederlandse landschap door de bevolking op kaart. Wageningen; 2005.

70. Heylen D, Lasters R, Adriaensen F, Fonville M, Sprong H, Matthysen E. Ticks and tick-borne diseases in the city: Role of landscape connectivity and green space characteristics in a metropolitan area. Sci Total Environ. 2019. doi: 10.1016/j.scitotenv.2019.03.235 30921726

Modelling tick bite risk by combining random forests and count data regression models

Souhrn

Klíčová slova:

Zdroje

PLOS One

Aktuální možnosti diagnostiky a léčby litiáz