Digital archive of theses discussed at the University of Pisa


Thesis etd-10142021-152949

Thesis type
Tesi di laurea magistrale LM5
Thesis title
Impact of diagnostic delay on drug utilization and use of healthcare facilities in patients with ulcerative colitis: review of literature and real-world data from an original observational study conducted on the administrative healthcare database of Tuscany.
Course of study
relatore Prof.ssa Martelli, Alma
relatore Dott. Tuccori, Marco
correlatore Dott.ssa Ferraro, Sara
  • diagnostic delay
  • diagnosis
  • inflammatory bowel diseases
  • ulcerative colitis
  • surgery
  • complications
  • treatment
  • Crohn’s disease
Graduation session start date
Release date
Diagnostic delay (DD) in ulcerative colitis (UC) and its clinical consequences are poorly investigated in the real-world setting. In this thesis, we evaluated the impact of DD on drug utilization and the use of healthcare facilities in patients with UC. We performed a literature review to evaluate available evidence on DD in UC by using PubMed free search engine (MEDLINE) according to specific inclusion criteria. The included studies were analysed, and we tabulated both their main methodologic characteristics and results. Moreover, we developed a research protocol to perform an original retrospective observational study aimed at quantifying the DD in UC patients and to evaluate its impact of effectiveness and safety outcomes. We provided real-world evidence on data extracted from the administrative healthcare database of Tuscany. We included adult patients with a record of UC diagnosis or co-payment exemption code for UC occurred between June 1st, 2011 and December 31st, 2018. The date of the earliest of these events was the index date (ID). Patients with a look-back period (time before ID) < 5 years and follow-up < 2 years were excluded. DD was defined by a record of emergency department (ED) access or hospitalization for gastrointestinal causes in the 7-60 months (7-18 months: short DD; 19-60 months: long DD) preceding ID. Patients with an ED access or hospitalization for gastrointestinal causes recorded in the 6 months before the ID or with none of these records were considered timely diagnosed (TD). We performed survival analyses (Kaplan-Meier curves) for effectiveness outcomes (time free from the first: dispensation of azathioprine, advanced therapy and surgery) and safety outcomes (time free from the first: ED access and/or hospitalization for any cause) over a 2-years follow-up period. Adjusted hazard ratio (aHR) was calculated by using Cox models adjusted for age, gender and number of concomitant drugs at cohort entry. Both outcomes were evaluated for dichotomous (TD and DD) and categorical (TD, short DD and long DD) variables of DD. After the literature review, we included 29 articles for the evaluation of DD in UC. The results showed that most studies were based on interviews or have been performed by using disease registries. We observed that the definition of DD was different between studies, and that frequency and duration (2-96 months) of DD reported in the results were heterogeneous. The majority of studies included less than 1,000 UC patients. Few studies analysed impact of DD on treatment and surgery. However, they highlighted that those patients with DD were more prone to disease complication compared with TD patients. Our original observational study included 3,804 patients and 313 (8.3%) patients had a DD: 113 and 200 patients had short and long DD, respectively. No significant aHR were found for the effectiveness outcomes. The safety analysis revealed that patients with DD have an increased risk for safety outcomes compared with TD patients: aHR: 1.5 (CI 95%: 1.3-1.7) for access to ED or hospitalizations, aHR: 1.4 (CI 95%:1.2-1.7) for hospitalization and 1.8 (CI 95%:1.5-2.0) for ED accesses.
This study demonstrated that about 8.3% of patients showed a DD (mostly long DD) and an increased ED access and hospitalization for any cause was found for patients with DD compared with TD patients. The methodology of our protocol was poorly comparable with that of other study designs. Overall, we observed that our UC population showed a lower frequency of DD than patients included in the other studies, while the mean time of DD (27.5 months) was often higher than that observed in other analyses. The comparison between our results and those of a previous study performed with the same methodology but conducted on Crohn’s disease (CD), showed a lower frequency of DD in UC patients compared to CD patients (17.5%). We also found that patients with UC did not have a significant risk of being treated with advanced therapy, differently from what observed for CD patients. Instead, the risk of access to ED or hospitalization for any cause was found for both inflammatory bowel disease populations.