A prospective study comparing patient-reported outcomes in Crohn's disease

Background Patient reported outcomes are important in Crohn's disease. In this prospective cohort, we investigated the performance of the Bristol Stool Form Scale (BSFS) and a visual analog scale (VAS) for abdominal pain as outcome measures in Crohn's disease. Methods Patients with active Crohn's disease starting glucocorticoids or anti-tumor necrosis factor were included. Before treatment and 10 weeks later we collected: clinical activity [Harvey Bradshaw Index (HBI) and Crohn's-Disease-Activity-Index (CDAI)], serum C-reactive protein (CRP) and fecal calprotectin, and BSFS (1-7) and a 100-mm VAS based on a 7-day diary. Clinical response was defined as a reduction by at least 3 and at least 100 of HBI and CDAI, respectively. Fecal calprotectin-response and CRP-response were defined as reduction of at least 50%. Results Thirty-eight patients completed follow-up. At baseline, BSFS-parameters correlated more strongly with clinical activity (range:r(s): 0.31-0.74) than with CRP (r(s): -0.01 to 0.16) and fecal calprotectin (r(s): 0.14-0.26). VAS scores correlated very weakly to moderately with clinical activity (r(s): 0.18-0.45), and weakly to moderately with CRP (r(s): 0.24-0.34) and fecal calprotectin (r(s): 0.35-0.43). Changes in VAS scores correlated moderately to strongly (r(s): 0.55-0.71) with changes in clinical activity, and weakly with changes in CRP and fecal calprotectin (r(s): 0.21-0.35). Changes in BSFS parameters correlated weakly to moderately (r(s): 0.23-0.53) with changes in clinical activity, and very weakly to weakly (r(s): 0.01-0.35) with changes in CRP and fecal calprotectin. Responsiveness of VAS and BSFS was moderate to high (Guyatt's statistic 0.41-2.17) and highly dependent on the definition of response. Conclusions The BSFS and a VAS appear to be responsive with moderate-to-strong construct validity to monitor patients with Crohn's disease.