(1) Developing and Testing of Diagnostic Algorithms for The Identification of Inflammatory Bowel Diseases Using a Claims Database from Southern Italy

Sunday, August 28, 2022

8:00 AM – 1:45 PM CEST

Location: Hall C

Publication Number: 1046

Presenting Author(s)

Elena Sofia Fiore
– Department of Diagnostics and Public Health, University of Verona, Verona, Padova, Padova, Italy

Co-Author(s)

Ylenia Ingrasciotta
– Messina, Messina, Italy
Valentina Isgrò
– Department of Diagnostics and Public Health, University of Verona, Verona, Italy, Italy
Valentina Ientile
– University of Messina, Italy, Italy
Saveria S Foti
– Data scientist, Academic spin‑off “INSPIRE, Innovative Solutions for Medical Prediction and Big Data Integration in Real World Setting”, Azienda Ospedaliera Universitaria “G. Martino”, Messina, Italy, Sicilia, Italy
Luca L'Abbate
– Department of Diagnostics and Public Health, University of Verona, Verona, Italy, Italy
Andrea Fontana
– Unit of Biostatistics, Fondazione IRCCS Casa Sollievo della Sofferenza, San Giovanni Rotondo, Italy, Italy
Michele Tari
– Pharmacist, Caserta LHU, Italy, Italy
Angela Alibrandi
– Professor, Department of Economics, University of Messina, Messina, Italy, Sicilia, Italy
Gianluca Trifiro, MD, FISPE
– Full Professor of Pharmacology, University of Verona, Verona, Italy

Background: Inflammatory bowel diseases (IBDs) are chronic diseases with increased morbidity and reduced quality of life. Development of newly coding algorithms for IBD identification using claims databases is needed.

Objectives: to develop and test specific coding algorithms to identify Crohn's disease (CD) and ulcerative colitis (UC) in a cohort of users of biologics approved for IBDs, using claims database of an Italian Local Health Unit (LHU).

Methods: Fully anonymized data were extracted from the claims database of Caserta LHU during the years 2015-2018. As gold standard, we considered confirmed diagnoses of CD and UC, as reported by specialists in electronic therapeutic plans (ETPs). Five CD/UC specific algorithms were developed. IBDs-related ICD9-CM codes as primary cause of hospital admissions, exemption from healthcare service co-payment codes and dispensing of biologic/non-biologic drugs only approved for CD/UC were the components of the newly developed IBD coding algorithms. Sensitivity (Se)/specificity (Sp)/positive (PPV)/negative predicted values (NPV) were estimated. Performance of each algorithm was evaluated through the Youden index. Such algorithms were developed to distinguish patients with CD/UC from non-IBD cases in the cohort of users of biologics approved for CD/UC treatment.

Results: Among 1 million inhabitants from Caserta LHU, 627 subjects received at least a biologic drug for which ETP was available. Of these, 134 (21%) biologic users had ≥1 ETP with IBD diagnosis; specifically, 83 (62%) and 51 (38%) patients were identified as confirmed diagnosis of CD and UC, separately. The best algorithm for CD identification (Youden Index (0.953)) included diagnosis code (i.e. ICD-9:555.x) OR specific exemption code (009.555/555) OR [non-specific exemption code (009) AND dispensing of non-biologic drugs with exclusive CD indication (methotrexate/budesonide)] (Se:98.8%; Sp:96.5%; PPV:81.2%). The best algorithm for UC identification (Youden index (0.851)) included diagnosis code (i.e. ICD-9:556.x) OR specific exemption code (009.555/555) OR [non-specific exemption code (009) AND golimumab] OR dispensing of non-biologic drugs with exclusive UC indication (i.e. mesalazine/balsalazide/budesonide) (Se:94.1%; Sp:91.0%; PPV:48.0%). Applying these algorithms to the whole cohort of biologic users with ≥1 ETP, estimated prevalence of CD and UC patients was 20% and 21%, respectively.

Conclusions: In a cohort of biologic users from a population-based claims database from Southern Italy, a newly developed coding algorithm including diagnostic and exemption codes in combination with drug dispensing was able to accurately identify patients receiving biologics for IBDs and further differentiate those treated for CD or UC.