Geography Reference
In-Depth Information
Discussion and Conclusions
This paper presented the usage of (empirical) Bayesian models for the
smoothing of disease rates—5-year prevalence and relative risk. These rates
were based not only on the recorded number of cases of infection by
campylobacteriosis in the Czech Republic, but also on the population density.
The case study showed that smoothing is more suitable for expressing local
differences in prevalence than in the case of relative risk.
Choropleth maps pointed out areas that might possibly create spatial
clusters based on the similarity of disease occurrence. These clusters were
later identified and described by Moran
s I enhanced by empirical Bayesian
rate with permutations. North-eastern Moravia proved to be most affected by
campylobacteriosis.
One has to realize that empirical Bayesian procedures tend to shift values
to the mean risk—global or local by incorporating information between areas.
The risks in areas with more information (e.g., urban areas) are usually less
smoothed than in areas that exhibit higher sampling variation (typically those
with a low number of cases), and thus produce more stable estimates of the
pattern of underlying disease risk (Richardson et al.
2004
). However,
although raw risks can produce “noisy” maps that are difficult to interpret,
oversmoothed maps may produce a homogeneous risk surface, masking the
actual risk distribution (Beale et al.
2008
). It is important to mention that all
the analyses presented in this paper are heavily dependent on the scale. We
chose the scale of municipal districts but results on other scales could show
differences.
'
Acknowledgments The authors gratefully acknowledge the support by the Operational Program
Education for Competitiveness—European Social Fund (project CZ.1.07/2.3.00/20.0170 of the
Ministry of Education, Youth and Sports of the Czech Republic).
References
Anselin L (1995) Local indicators of spatial association—LISA. Geogr Anal 27:93-115
Anselin L (2003) GeoDaTM 0.9 User
'
s Guide
Assuncao R, Reis E (1999) A new proposal to adjust Moran
'
s I for population density. Stat Med
2162:2147-2162
Bailey TC (2001) Spatial statistical methods in health. Cadernos Sa´de P´blica 17(5):1083-1098
Bailey TC, Gatrell AC (1995) Interactive spatial data analysis. Longman Scientific & Technical,
Essex
Beale L, Abellan JJ, Hodgson S, Jarup L (2008) Methodologic issues and approaches to spatial
epidemiology. Environ Health Perspect 116:1105-1110
Bell BS, Hoskins RE, Pickle LW, Wartenberg D (2006) Current practices in spatial analysis of
cancer data: mapping health statistics to inform policymakers and the public. Int J Health
Geogr 5:49
