• hospitalization
• biologic drug
• administrative database
• diagnosis
• Crohn’s disease
• diagnostic delay

Diagnostic delay in Crohn's disease is mainly documented by registry-based studies or patient interviews. Given the impact of diagnostic delay on drug-utilization, quality of life and health costs, studying this phenomenon in large populations is of considerable interest. Observing the diagnostic delay on administrative databases can lead to limitations deriving from the lack of possibility to include outpatients or to identify the causes of the delay. However, it is worthy of attention the evaluation of patients transiting through the hospital setting with possible evident symptoms but who may not have been diagnosed before.
The purpose of this study was to identify possible misdiagnoses in subjects detected in the hospital setting to assess their impact on the drug-utilization and use of health system facilities.
This thesis describes a retrospective cohort study conducted on the administrative databases of Tuscany. We included patients with a first event of: 1) diagnosis of Crohn's disease (recorded in the emergency department or in the hospital discharge records) or 2) Crohn's disease exemption from co-payment code or 3) oral budesonide dispensation, between 01/06/2011 and 30/06/2016. The date of the first, or the only, occurrence of one of these events represented the index date. Patients’ follow-up was three years. Subjects were divided into two comparison groups: patients with at least one access to the emergency department or hospitalization for gastrointestinal causes (proxy of disease symptoms) in the five years prior to the index date and patients without. We calculated the time free from the use of biological drugs and from hospitalization and/or emergency department access for all causes through a survival analysis (Kaplan-Meier). The hazard ratio was calculated using Cox models.
We identified 3342 patients with Crohn's disease. Among them, 584 (17.5%) had previous hospital access due to gastrointestinal causes (220 patients in the 7-18 months prior to the index date and 364 in the 19-60 months prior to the index date). A statistically significant difference was observed for both time free from the use of biological drugs (adjusted HR: 2.17 95% CI: 1.75-2.71) and time free from access to the emergency department or hospitalization for all causes (adjusted HR: 1.59 95% CI:1.44-1.75) in patients with previous access to hospital due to gastrointestinal causes compared to patients without.
The risk of early use of biological drugs and access to the emergency room or hospitalization in the three years following diagnosis was higher for patients with previous access to the hospital for gastrointestinal causes than those without.