Background Cluster detection is an essential tool in the public health domain with the goal of detecting anomalous clusters of disease cases. We performed a spatial-time cluster analysis of the Johns Hopkins Lupus cohort with the goal of identifying potential spatial-time clusters of SLE organ specific disease activity.

Methods 1844 patients who fulfil 4 of the 11 American College of Rheumatology classification criteria for SLE and who had recorded home addresses were included in the analysis. Cluster detection analysis in both space and time of disease activity expressed as Physician Global Estimate (PGA) was performed. The area utilised in this analysis was a 350 kilometre radial buffer around the Johns Hopkins Lupus Centre, and included all of Maryland, Delaware, and District of Columbia, as well as parts of Pennsylvania, New Jersey, Virginia, and West Virginia. This area was considered due to the high and consistent density of study participants. The data ranged from 1987 to 2017, with the spring, summer, fall, and winter seasons serving as time units for the temporal based analyses.

Results CNS, renal, and joint flares have both seasonal patterns as well as large-scale multi-year trends. CNS flares clustered between Annapolis, MD and Frederick, MD between 1987 and 2000, renal flares clustered in central Maryland and northern Virginia between 2002 and 2006 and and a joint flare cluster included Delaware, Delaware Bay area, and Chesapeake Bay area between 2003 and 2014. Maps were generated highlighting the study area, flares, and identified clusters from all analyses. The space-time effects of environmental and demographic variables on the identified clusters will be considered in subsequent analysis.

Conclusions We describe the first space-time clusters of lupus organ-specific disease activity strongly supporting the role of environmental factors as drivers of lupus activity.