Development of Machine Learning Model to Predict the 5-Year Risk of Starting Biologic Agents in Patients with Inflammatory Bowel Disease (IBD): K-CDM Network Study

Background: The incidence and global burden of inflammatory bowel disease (IBD) have steadily increased in the past few decades. Improved methods to stratify risk and predict disease-related outcomes are required for IBD. Aim: The aim of this study was to develop and validate a machine learning (ML) model to predict the 5-year risk of starting biologic agents in IBD patients. Method: We applied an ML method to the database of the Korean common data model (K-CDM) network, a data sharing consortium of tertiary centers in Korea, to develop a model to predict the 5-year risk of starting biologic agents in IBD patients. The records analyzed were those of patients diagnosed with IBD between January 2006 and June 2017 at Gil Medical Center (GMC; n = 1299) or present in the K-CDM network (n = 3286). The ML algorithm was developed to predict 5- year risk of starting biologic agents in IBD patients using data from GMC and externally validated with the K-CDM network database. Result: The ML model for prediction of IBD-related outcomes at 5 years after diagnosis yielded an area under the curve (AUC) of 0.86 (95% CI: 0.82–0.92), in an internal validation study carried out at GMC. The model performed consistently across a range of other datasets, including that of the K-CDM network (AUC = 0.81; 95% CI: 0.80–0.85), in an external validation study. Conclusion: The ML-based prediction model can be used to identify IBD-related outcomes in patients at risk, enabling physicians to perform close follow-up based on the patient’s risk level, estimated through the ML algorithm.

Download Full-text

Development of machine learning model to predict the risk of 5- year disease related outcomes in patients with inflammatory bowel disease (Preprint)

10.2196/preprints.17304 ◽

2019 ◽

Author(s):

Youn I Choi ◽

Sung Jin Park ◽

Yoon Jae Kim ◽

Kwang Gi Kim ◽

Dong Kyun Park ◽

...

Keyword(s):

Machine Learning ◽

Inflammatory Bowel Disease ◽

Validation Study ◽

Bowel Disease ◽

Medical Center ◽

External Validation ◽

Biologic Agents ◽

Risk Level ◽

Common Data Model ◽

Inflammatory Bowel

BACKGROUND The incidence and global burden of inflammatory bowel disease (IBD) have steadily increased in the past few decades. Improved methods to stratify risk and predict disease-related outcomes are required for IBD. OBJECTIVE The aim of this study was to develop and validate a machine learning (ML) model to predict the 5-year risk of starting biologic agents in IBD patients. METHODS We applied an ML method to the database of the Korean common data model (K-CDM) network, a data sharing consortium of tertiary centers in Korea, to develop a model to predict the 5-year risk of starting biologic agents in IBD patients. The records analyzed were those of patients diagnosed with IBD between January 2006 and June 2017 at Gil Medical Center (GMC; n = 1,299) or present in the K-CDM network (n = 3,286). The ML algorithm was developed using data from GMC and externally validated with the K-CDM network database. RESULTS The ML model for prediction of IBD-related outcomes at 5 years after diagnosis yielded an area under the curve (AUC) of 0.86 (95% CI: 0.82–0.92), in an internal validation study carried out at GMC. The model performed consistently across a range of other datasets, including that of the K-CDM network (AUC = 0.81; 95% CI: 0.80–0.85), in an external validation study. CONCLUSIONS The ML-based prediction model can be used to identify IBD-related outcomes in patients at risk, enabling physicians to perform close follow-up based on the patient’s risk level, estimated through the ML algorithm.

Download Full-text

Effect of Age on the Initiation of Biologic Agent Therapy in Patients With Inflammatory Bowel Disease: Korean Common Data Model Cohort Study (Preprint)

10.2196/preprints.15124 ◽

2019 ◽

Author(s):

Youn I Choi ◽

Yoon Jae Kim ◽

Jun-Won Chung ◽

Kyoung Oh Kim ◽

Hakki Kim ◽

...

Keyword(s):

Inflammatory Bowel Disease ◽

Bowel Disease ◽

Data Model ◽

Biologic Agents ◽

Age At Diagnosis ◽

Common Data Model ◽

Event Free Survival ◽

Free Survival ◽

Network Database ◽

Inflammatory Bowel

BACKGROUND The Observational Health Data Sciences and Informatics (OHDSI) network is an international collaboration established to apply open-source data analytics to a large network of health databases, including the Korean common data model (K-CDM) network. OBJECTIVE The aim of this study is to analyze the effect that age at diagnosis has on the prognosis of inflammatory bowel disease (IBD) in Korea using a CDM network database. METHODS We retrospectively analyzed the K-CDM network database from 2005 to 2015. We transformed the electronic medical record into the CDM version 5.0 used in OHDSI. A worsened IBD prognosis was defined as the initiation of therapy with biologic agents, including infliximab and adalimumab. To evaluate the effect that age at diagnosis had on the prognosis of IBD, we divided the patients into an early-onset (EO) IBD group (age at diagnosis <40 years) and a late-onset (LO) IBD group (age at diagnosis ≥40 years) with the cutoff value of age at diagnosis as 40 years, which was calculated using the Youden index method. We then used the logrank test and Cox proportional hazards model to analyze the effect that age at diagnosis (EO group vs LO group) had on the prognosis in patients with IBD. RESULTS A total of 3480 patients were enrolled. There was 2017 patients with ulcerative colitis (UC) and 1463 with Crohn’s disease (CD). The median follow up period was 109.5 weeks. The EO UC group was statistically significant and showed less event-free survival (ie, experiences of biologic agents) than the LO UC group (P<.001). In CD, the EO CD group showed less event-free survival (ie, experiences of biologic agents) than the LO CD group. In the Cox proportional hazard analysis, the odds ratio (OR) of the EO UC group on experiences of biologic agents compared with the LO UC group was 2.3 (95% CI 1.3-3.8, P=.002). The OR of the EO CD group on experiences of biologic agents compared with the LO CD group was 5.4 (95% CI 1.9-14.9, P=.001). CONCLUSIONS The EO IBD group showed a worse prognosis than the LO IBD group in Korean patients with IBD. In addition, this study successfully verified the CDM model in gastrointestinal research.

Download Full-text

Effect of Age on the Initiation of Biologic Agent Therapy in Patients With Inflammatory Bowel Disease: Korean Common Data Model Cohort Study

JMIR Medical Informatics ◽

10.2196/15124 ◽

2020 ◽

Vol 8 (4) ◽

pp. e15124

Author(s):

Youn I Choi ◽

Yoon Jae Kim ◽

Jun-Won Chung ◽

Kyoung Oh Kim ◽

Hakki Kim ◽

...

Keyword(s):

Inflammatory Bowel Disease ◽

Bowel Disease ◽

Data Model ◽

Biologic Agents ◽

Age At Diagnosis ◽

Common Data Model ◽

Event Free Survival ◽

Free Survival ◽

Network Database ◽

Inflammatory Bowel

Background The Observational Health Data Sciences and Informatics (OHDSI) network is an international collaboration established to apply open-source data analytics to a large network of health databases, including the Korean common data model (K-CDM) network. Objective The aim of this study is to analyze the effect that age at diagnosis has on the prognosis of inflammatory bowel disease (IBD) in Korea using a CDM network database. Methods We retrospectively analyzed the K-CDM network database from 2005 to 2015. We transformed the electronic medical record into the CDM version 5.0 used in OHDSI. A worsened IBD prognosis was defined as the initiation of therapy with biologic agents, including infliximab and adalimumab. To evaluate the effect that age at diagnosis had on the prognosis of IBD, we divided the patients into an early-onset (EO) IBD group (age at diagnosis <40 years) and a late-onset (LO) IBD group (age at diagnosis ≥40 years) with the cutoff value of age at diagnosis as 40 years, which was calculated using the Youden index method. We then used the logrank test and Cox proportional hazards model to analyze the effect that age at diagnosis (EO group vs LO group) had on the prognosis in patients with IBD. Results A total of 3480 patients were enrolled. There was 2017 patients with ulcerative colitis (UC) and 1463 with Crohn’s disease (CD). The median follow up period was 109.5 weeks. The EO UC group was statistically significant and showed less event-free survival (ie, experiences of biologic agents) than the LO UC group (P<.001). In CD, the EO CD group showed less event-free survival (ie, experiences of biologic agents) than the LO CD group. In the Cox proportional hazard analysis, the odds ratio (OR) of the EO UC group on experiences of biologic agents compared with the LO UC group was 2.3 (95% CI 1.3-3.8, P=.002). The OR of the EO CD group on experiences of biologic agents compared with the LO CD group was 5.4 (95% CI 1.9-14.9, P=.001). Conclusions The EO IBD group showed a worse prognosis than the LO IBD group in Korean patients with IBD. In addition, this study successfully verified the CDM model in gastrointestinal research.

Download Full-text

Clinical applications of artificial intelligence and machine learning‐based methods in inflammatory bowel disease

Journal of Gastroenterology and Hepatology ◽

10.1111/jgh.15405 ◽

2021 ◽

Vol 36 (2) ◽

pp. 279-285

Author(s):

Shirley Cohen‐Mekelburg ◽

Sameer Berry ◽

Ryan W Stidham ◽

Ji Zhu ◽

Akbar K Waljee

Keyword(s):

Artificial Intelligence ◽

Machine Learning ◽

Inflammatory Bowel Disease ◽

Bowel Disease ◽

Clinical Applications ◽

Inflammatory Bowel

Download Full-text

QuantiFERON-TB Gold Test Conversion Is Associated with Active Tuberculosis Development in Inflammatory Bowel Disease Patients Treated with Biological Agents: An Experience of a Medical Center in Taiwan

Gastroenterology Research and Practice ◽

10.1155/2019/7132875 ◽

2019 ◽

Vol 2019 ◽

pp. 1-9 ◽

Cited By ~ 1

Author(s):

Hsiang-Chun Lai ◽

Chia-Hsi Chang ◽

Ken-Sheng Cheng ◽

Tsung-Wei Chen ◽

Yuan-Yao Tsai ◽

...

Keyword(s):

Inflammatory Bowel Disease ◽

Bowel Disease ◽

Biological Treatment ◽

Medical Center ◽

Biological Agents ◽

Baseline Serum ◽

Tertiary Referral Center ◽

Western Countries ◽

Before And After ◽

Inflammatory Bowel

Taiwan has a lower prevalence of inflammatory bowel disease (IBD) and a higher prevalence of tuberculosis (TB) infection than Western countries. The aim of this study was to investigate the prevalence of latent TB (LTB) and active TB infection in IBD patients treated with biological agents. From January 2000 to September 2018, we retrospectively collected data from IBD patients treated with biological agents at a tertiary referral center. Patients underwent a QuantiFERON-TB Gold test (QFT) to screen for TB infection before and after biological treatment courses. The diagnostic age, sex, body mass index, hepatitis B virus infection, biochemistry profile, treatment regimens, and the results of the QFT were analyzed. Overall, 130 IBD patients who received biological treatment were enrolled. The results of the QFT before biological treatment were determined in 120 patients (92%); of these, 10 were positive (8%), 110 were negative (85%), and 10 were indeterminate (9%). Six patients demonstrated seroconversion after biological treatment, as determined by the QFT. Three patients (2.4%) developed active pulmonary TB after biological treatment. In subgroup analysis, the positive QFT patients had a trend of lower baseline serum C-reactive protein and erythrocyte sedimentation rate levels than the negative QFT group. The present study demonstrates that the prevalence of LTB before and after biological treatment is higher in Taiwan than in most Western countries and similar to other Asian countries. Therefore, screening and monitoring of TB infection are necessary for IBD patients before and during biological treatments in Taiwan.

Download Full-text