P651 Diagnostic delay, effectiveness and safety outcomes in a real-world cohort of patients with Crohn’s disease: data from administrative databases in Tuscany, Italy

Bertani, L.(1);Ferraro, S.(2);Bartolini, C.(3);Convertino, I.(2);Giometto, S.(2);Cappello, E.(2);Valdiserra, G.(2);Tillati, S.(2);Blandizzi, C.(2);Lucenteforte, E.(2);Gini, R.(3);Tuccori, M.(2,4);Costa, F.(5);

(1)University of Pisa, Translational Research and New Technologies in Medicine and Surgery, Pisa, Italy;(2)University of Pisa, Clinical and Experimental Medicine, Pisa, Italy;(3)Tuscan Regional Healthcar Agency, Ars, Florence, Italy;(4)Pisa University Hospital, Unit of Adverse Drug Reactions Monitoring, Pisa, Italy;(5)Pisa University Hospital, General Surgery and Gastroenterology - IBD Unit, Pisa, Italy

Background

Diagnostic delay (DD) of Crohn’s disease (CD) is poorly investigated in the real-world setting, and could affect significantly the outcome of CD. The aim of the present study was to quantify the possible DD of CD in CD patients in Tuscany region of Italy, evaluating effectiveness and safety outcomes in CD patients with DD.

Methods

We conducted a retrospective cohort study on data extracted from administrative databases of Tuscany. We included patients >18 years old (yo) with a record of ICD-9 diagnosis for CD or disease exemption or a dispensation of oral budesonide from 6/1/2011 to 6/30/2016 (index date, ID) whichever came first. Patients with a look-back period (time before ID) < 5 y and follow-up (FUP) < 3 y were excluded. Patients with at least one Emergency Department (ED) access or hospitalisation for gastrointestinal causes in the lookback period (excluding 6 months before ID) could have a possible DD. DD was classified in short (up to 18 months) and long (up to 60 months). We performed survival analyses (Kaplan-Meier curves) for effectiveness (time free from the first: dispensation of azathioprine, biologic drug and ileocecal resection surgery) and safety outcomes (time free from first: ED access and/or hospitalization for any cause). Hazard ratio (HR) was calculated by using Cox models adjusted for age, gender and number of concomitant drugs in the month before ID. Both outcomes were evaluated for dichotomous (No DD and DD) and categorical variables of DD (No DD, short DD and long DD).

Results

Among 3342 CD patients, 584 (17,5%) had a suspected DD: 212 and 372 patients had short and long DD, respectively, p<0.001. The effectiveness analysis revealed that about the 20% of patients with long DD were treated with a biologic within 3 years, as compared with those with short DD or no DD (10%), p=0.0007. The adjusted (a)HR for biologic drug was 1.30 (Confidence Interval, CI 95%: 1.12-1.56) for patients with long DD and 1.02 (CI 95%:0.82-1.30) for those with short DD. No significant aHRs were found for the other effectiveness outcomes. The safety analysis showed that 75% of patients with DD had first ED access or hospitalization for any causes with the FUP, as compared with those without (about 60%), p<0.001. The aHR showed an increased risk for safety outcomes in patients with DD: 1.71 (CI 95%:1.61-1.91) for the ED accesses, 1.31 (CI 95%:1.19-1.49) for hospitalization, and 1.54 (CI 95%:1.41-1.71) for the composite outcome.

Conclusion

More than 17% of patients had a putative DD (mostly long DD). A risk of treatment with biologic drug, ED accesses and hospitalizations for any cause was found for patients with DD compared with those timely diagnosed.