In the primary analysis, patients were grouped according to their PSFS as F0–F2 versus F3/F4. When categorizing patients as having early versus advanced fibrosis, we referred to 80–90% thresholds for each of the eight CRFS variables as defined in Figure 1B. Proportions within each of the four groups were reported as: (i) aligned, CRFS of no/early fibrosis and PSFS both F0–F2; (ii) aligned, CRFS of advanced fibrosis and PSFS both F3/4; (iii) physician underestimated, PSFS F0–F2 but CRFS of advanced fibrosis; or (iv) physician overestimated, PSFS F3/4 but CRFS of no/early fibrosis.
The secondary objective was to identify factors associated with alignment or misalignment between PSFS and CRFS. Alignment was assessed according to treating physician specialty (gastroenterologist, hepatologist, diabetologist), liver biopsy performed (yes, no), and key patient clinical and demographic characteristics, including age (>65 years, ≤65 years), key comorbidities (hypertension, metabolic conditions [any of metabolic syndrome, T2DM, insulin resistance or hyperglycemia]), and ethnicity (White/Caucasian, Asian, Hispanic/Latino, Afro-Caribbean, other). Statistically significant differences were identified using univariate tests (t, chi-squared, Fisher’s exact and log rank tests). P-values of <0.05 were considered statistically significant; summary statistics were used to provide the size and direction of the difference.
Three sensitivity analyses were undertaken to test the assessment of alignment of fibrosis score: