Limits...
Cluster Detection Tests in Spatial Epidemiology: A Global Indicator for Performance Assessment.

Guttmann A, Li X, Feschet F, Gaudart J, Demongeot J, Boire JY, Ouchchane L - PLoS ONE (2015)

Bottom Line: New or improved CDTs are regularly proposed to epidemiologists and must be subjected to performance assessment.In a simulation study, performance is measured for many tests.We evidence the properties of these two indicators and the superiority of the cumulated TC to assess performance.

View Article: PubMed Central - PubMed

Affiliation: Department of Biostatistics, Clermont University Hospital, Clermont-Ferrand, France; UMR CNRS UDA 6284 ISIT, Auvergne University, Clermont-Ferrand, France.

ABSTRACT
In cluster detection of disease, the use of local cluster detection tests (CDTs) is current. These methods aim both at locating likely clusters and testing for their statistical significance. New or improved CDTs are regularly proposed to epidemiologists and must be subjected to performance assessment. Because location accuracy has to be considered, performance assessment goes beyond the raw estimation of type I or II errors. As no consensus exists for performance evaluations, heterogeneous methods are used, and therefore studies are rarely comparable. A global indicator of performance, which assesses both spatial accuracy and usual power, would facilitate the exploration of CDTs behaviour and help between-studies comparisons. The Tanimoto coefficient (TC) is a well-known measure of similarity that can assess location accuracy but only for one detected cluster. In a simulation study, performance is measured for many tests. From the TC, we here propose two statistics, the averaged TC and the cumulated TC, as indicators able to provide a global overview of CDTs performance for both usual power and location accuracy. We evidence the properties of these two indicators and the superiority of the cumulated TC to assess performance. We tested these indicators to conduct a systematic spatial assessment displayed through performance maps.

No MeSH data available.


Related in: MedlinePlus

Size of the at-risk population for each SU in the Auvergne region, as defined by mean number of live births per year between 1999 and 2006 (source: INSEE).Q1:≤ 17; Q2:> 17 and ≤ 35; Q3:> 35 and ≤ 70; Q4:> 70.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4472237&req=5

pone.0130594.g001: Size of the at-risk population for each SU in the Auvergne region, as defined by mean number of live births per year between 1999 and 2006 (source: INSEE).Q1:≤ 17; Q2:> 17 and ≤ 35; Q3:> 35 and ≤ 70; Q4:> 70.

Mentions: We applied these two baseline risks (incidences) of birth defects to the same at-risk population, which size was approximated by mean annual number of live births. (The distribution of the at-risk population is shown in Fig 1.) For each baseline incidence (I = 2.26% of births or I = 0.48%), we defined two cluster collections by applying two relative risks (3 and 6) to the same pattern of location and cluster size. The relative risks were chosen in order to observe all the range of performance. Each cluster collection contains 221 clusters of four SUs (one central SU and its three nearest neighbors in euclidean distances) successively centered on each SU of the region.


Cluster Detection Tests in Spatial Epidemiology: A Global Indicator for Performance Assessment.

Guttmann A, Li X, Feschet F, Gaudart J, Demongeot J, Boire JY, Ouchchane L - PLoS ONE (2015)

Size of the at-risk population for each SU in the Auvergne region, as defined by mean number of live births per year between 1999 and 2006 (source: INSEE).Q1:≤ 17; Q2:> 17 and ≤ 35; Q3:> 35 and ≤ 70; Q4:> 70.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4472237&req=5

pone.0130594.g001: Size of the at-risk population for each SU in the Auvergne region, as defined by mean number of live births per year between 1999 and 2006 (source: INSEE).Q1:≤ 17; Q2:> 17 and ≤ 35; Q3:> 35 and ≤ 70; Q4:> 70.
Mentions: We applied these two baseline risks (incidences) of birth defects to the same at-risk population, which size was approximated by mean annual number of live births. (The distribution of the at-risk population is shown in Fig 1.) For each baseline incidence (I = 2.26% of births or I = 0.48%), we defined two cluster collections by applying two relative risks (3 and 6) to the same pattern of location and cluster size. The relative risks were chosen in order to observe all the range of performance. Each cluster collection contains 221 clusters of four SUs (one central SU and its three nearest neighbors in euclidean distances) successively centered on each SU of the region.

Bottom Line: New or improved CDTs are regularly proposed to epidemiologists and must be subjected to performance assessment.In a simulation study, performance is measured for many tests.We evidence the properties of these two indicators and the superiority of the cumulated TC to assess performance.

View Article: PubMed Central - PubMed

Affiliation: Department of Biostatistics, Clermont University Hospital, Clermont-Ferrand, France; UMR CNRS UDA 6284 ISIT, Auvergne University, Clermont-Ferrand, France.

ABSTRACT
In cluster detection of disease, the use of local cluster detection tests (CDTs) is current. These methods aim both at locating likely clusters and testing for their statistical significance. New or improved CDTs are regularly proposed to epidemiologists and must be subjected to performance assessment. Because location accuracy has to be considered, performance assessment goes beyond the raw estimation of type I or II errors. As no consensus exists for performance evaluations, heterogeneous methods are used, and therefore studies are rarely comparable. A global indicator of performance, which assesses both spatial accuracy and usual power, would facilitate the exploration of CDTs behaviour and help between-studies comparisons. The Tanimoto coefficient (TC) is a well-known measure of similarity that can assess location accuracy but only for one detected cluster. In a simulation study, performance is measured for many tests. From the TC, we here propose two statistics, the averaged TC and the cumulated TC, as indicators able to provide a global overview of CDTs performance for both usual power and location accuracy. We evidence the properties of these two indicators and the superiority of the cumulated TC to assess performance. We tested these indicators to conduct a systematic spatial assessment displayed through performance maps.

No MeSH data available.


Related in: MedlinePlus