The Cerebral Haemorrhage Anatomical RaTing inStrument (CHARTS): Development and assessment of reliability

Abstract Purpose The causes, risk factors and prognosis of spontaneous intracerebral haemorrhage (ICH) are partly determined by anatomical location (specifically, lobar vs. non-lobar (deep and infratentorial) regions). We systematically developed a rating instrument to reliably classify ICH location. Methods We used a two-stage iterative Delphi-style method for instrument development. The resultant Cerebral Haemorrhage Anatomical RaTing inStrument (CHARTS) was validated on CT and MRI scans from a cohort of consecutive patients with acute spontaneous symptomatic ICH by three independent raters. We tested interrater and intrarater reliability using kappa statistics. Results Our validation cohort included 227 patients (58% male; median age: 72.4 (IQR: 67.1–74.6)). The interrater reliability for the main analyses (i.e. including any lobar ICH; all deep and infratentorial anatomical categories (lentiform, caudate thalamus; brainstem; cerebellum); and uncertain location) was excellent (all kappa values > 0.80) both in pair-wise between-rater comparisons and across all raters. The intrarater reliability was substantial to almost perfect (k = 0.83; 95%CI: 0.77–0.88 and k = 0.95; 95%CI: 0.92–0.96 respectively). All kappa statistics remained consistent for individual cerebral lobar regions. Conclusions The CHARTS instrument can be used to reliably and comprehensively map the anatomical location of spontaneous ICH, and may be helpful for studying important questions regarding causes, risk factors, prognosis, and for stratification in clinical trials.