Application of the Potthoff-Whittinghill technique to detecting temporal clustering of diagnoses of type 1 diabetes at ages 0–14 years in Northumberland, Newcastle upon Tyne and North Tyneside during 1990–2007 inclusive.

Notes:aPW is the one-step estimate of β, the extra-Poisson variation, calculated as S/i(0) in the notation of Muirhead [22].bSE is the standard error of PW in the absence of extra-Poisson variation, calculated as 1/√i(0) in the notation of Muirhead [22].cp-values have been calculated using 10000 simulations of PW, assuming Poisson variation. All p-values are one-sided.dThe analysis between years was based on all 526 cases, whereas the other analyses excluded the 58 cases with a diagnosis date of 1st July.