156 Some mainland Scottish localities have higher incidence rates than orkney. Issue 12 (14th November 2019)

• Record Type: Journal Article
• Authors: Kearns, Patrick; Connick, Peter; MacDonald, James; O'Riordan, Jonathon; MacDougall, Niall; Colville, Shuna; Paton, Martin; O'Neill, Martin; Ascherio, Alberto; Chandran, Siddharthan
• Is Part Of: Journal of neurology, neurosurgery and psychiatry. Volume 90:Issue 12(2019)
• Page Start: e43
• Page End: e43
• Publication Date: 2019-11-14
• Subjects: Neurology -- Periodicals; Nervous system -- Surgery -- Periodicals; Psychiatry -- Periodicals; 616.8
• DOI: 10.1136/jnnp-2019-ABN-2.143
• Languages: English
• ISSNs: 0022-3050

Abstract

Abstract : There is wide regional variability in incidence of multiple sclerosis (MS) in Scotland. Extreme high-incidence is well known in Orkney, but most of the Scottish population (98%) lives on the mainland where local patterns are understudied. Well-designed small-areal units of geography and a national prospective incidence register allow us to examine the scale at which incident cases of MS cluster. Using the Scottish Multiple Sclerosis Register (1/1/2010 to 31/12/2017) we created multivariable linear regression models of small area incidence rates adjusted for population structure and latitude on the mainland. To examine the spatial distribution and independence of adjusted-rates, we calculated Moran's I test for global autocorrelation on the residuals. We then used Getis-Ord Gi* and the isotonic, iterative SatScan statistic to identify local areas with significant positive spatial autocorrelation adjusting for multiple testing and spatial dependency. We can with confidence assert that incidence rates are not spatially random after controlling for latitude. Several localities on the mainland are identified which, over this period, had incidence rates (and populations) higher than Orkney. These may be worthy of monitoring, may generate aetiological hypotheses or be suitable for trials of putative-preventative therapies. Unexplained local variation of incidence suggests missing environmental risk factors.