Disease-specific incident glomerulonephritis displays geographic clustering in under-serviced rural areas of British Columbia, Canada

There is little known about geographic variability in the incidence of glomerular disease and its potential implications for care delivery. To evaluate this, we performed a population-level cohort study using a provincial renal pathology database (2000-2012) to capture all incident cases of glomerulonephritis in British Columbia, Canada. This included 401 patients with membranous nephropathy (MN), 824 patients with IgA nephropathy (IgAN), 385 patients with focal segmental glomerulosclerosis (FSGS), 397 patients with lupus nephritis (LN) and 399 patients with ANCA-related glomerulonephritis (ANCA-GN). Geographic clusters were identified using Bayesian spatial models to estimate the incidence of each disease in 74 regions compared to the mean incidence in the entire province (incidence rate ratio, [IRR]), adjusted for region-level age, sex and race. The proportion of overall variability in incidence attributed to inter-regional differences varied by disease: 18% in MN, 81% in IgAN, 18% in FSGS, 59% in ANCA-GN, and 89% in LN. Except for LN, clustering was not explained by demographics. All IgAN and LN clusters were in urban regions close to nephrology centers, whereas ANCA-GN, MN and FSGS clustered mainly in rural regions. All ANCA-GN clusters were rural with median population density 1.2 persons/km 2 and driving distances of 10-676 km to the nearest nephrology center. Thus, we found significant geographic clustering in the incidence of different glomerular diseases. MN, FSGS and ANCA-GN clustered in sparsely populated regions with limited access to care, underscoring the importance of regional variability in glomerular diseases to inform health services delivery.