Abstract
Background: Patient wellbeing is presently the central focus in the shift towards holistic IBD care. There is however, a lack of real world data (RWD) to identify clear targetable areas as framed by patient-reported outcomes (PROs), and thereby the necessary scientific platform to study and improve wellbeing.
Aim: To analyse contemporary RWD that includes all wellbeing dimensions involving GI symptoms, social, psychological and emotional health as reported by patients; and to test scientific method using clinical metadata to build future interventional studies to investigate IBD-associated fatigue.
Methods: We utilised CUCQ32-PROs in 2 current IBD biomarker studies (GI-DAMPs/MUSIC; www.musicstudy.uk with 162 prospectively collected clinical data-points and prospectively recorded endoscopic mucosal healing) (Cohort 1) and via a freely available online wellbeing survey (Cohort 2) that includes non-IBD healthy controls1. CUCQ32 encompasses aspects of fatigue/anxiety/sexual/emotional health; with a score range from 0-272 (higher score = worse QoL).
Results: Overall, we captured ~2500 CUCQ32-PROs (between 2020-23 in UK); 800 and 1700 (112 [6.6%] international) responses from Cohorts 1 and 2 respectively (Table 1). Disease activity: CUCQ32 scores were higher in Cohort 1 and 2 in active IBD (116 vs. 33 and 151 vs. 70; both p<0.001). Even in remission, CUCQ32 is higher than in non-IBD controls, 70 vs. 27 (p<0.0001). CUCQ32 and IBD disease activity markers: In Cohort 1, CUCQ32 significantly correlated with SCCAI, HBI, calprotectin and CRP (r = 0.82, 0.78, 0.25 and 0.16) but not endoscopic mucosal healing. Fatigue: Fatigue scores (no of days tired out 14) correlated well with overall CUCQ32 scores (r = 0.73; p<0.001). We partitioned our dataset to Fatiguehigh vs. Fatiguenormal (defined as '˜tired' ≥10/14 days). CUCQ32 higher 101 vs. 32 respectively (p<0.001). Notably in Cohort 1 and 2, 185/236 and 338/606 (64.8% and 55.8%) in remission and 18/40 (45%) with complete endoscopic mucosal healing are Fatiguehigh. Prediction of Fatigue: Multiple logistic regression modelling using clinical parameters alone was poorly predictive of Fatiguehigh (Area under the Curve, AUC =0.62). We further applied Random Forest machine learning approach and generated AUC 0.62 to predict Fatigue high; relatively poor in contrast with AUCs of 0.84, to predict active IBD. (All data presented as medians).
Conclusions: Fatigue remains a major issue even in clinical remission and mucosal healing. Current high dimensional approaches with machine learning are poorly predictive of fatigue. Our ongoing work will incorporate genetic, microbiome and metabolic data to utilise this initial platform to understand the factors that drive IBD-associated fatigue.
1Well-being, Gut Health and Science: Patient-reported outcomes in Inflammatory Bowel Disease (https://edin.ac/3sTGpZT)
Aim: To analyse contemporary RWD that includes all wellbeing dimensions involving GI symptoms, social, psychological and emotional health as reported by patients; and to test scientific method using clinical metadata to build future interventional studies to investigate IBD-associated fatigue.
Methods: We utilised CUCQ32-PROs in 2 current IBD biomarker studies (GI-DAMPs/MUSIC; www.musicstudy.uk with 162 prospectively collected clinical data-points and prospectively recorded endoscopic mucosal healing) (Cohort 1) and via a freely available online wellbeing survey (Cohort 2) that includes non-IBD healthy controls1. CUCQ32 encompasses aspects of fatigue/anxiety/sexual/emotional health; with a score range from 0-272 (higher score = worse QoL).
Results: Overall, we captured ~2500 CUCQ32-PROs (between 2020-23 in UK); 800 and 1700 (112 [6.6%] international) responses from Cohorts 1 and 2 respectively (Table 1). Disease activity: CUCQ32 scores were higher in Cohort 1 and 2 in active IBD (116 vs. 33 and 151 vs. 70; both p<0.001). Even in remission, CUCQ32 is higher than in non-IBD controls, 70 vs. 27 (p<0.0001). CUCQ32 and IBD disease activity markers: In Cohort 1, CUCQ32 significantly correlated with SCCAI, HBI, calprotectin and CRP (r = 0.82, 0.78, 0.25 and 0.16) but not endoscopic mucosal healing. Fatigue: Fatigue scores (no of days tired out 14) correlated well with overall CUCQ32 scores (r = 0.73; p<0.001). We partitioned our dataset to Fatiguehigh vs. Fatiguenormal (defined as '˜tired' ≥10/14 days). CUCQ32 higher 101 vs. 32 respectively (p<0.001). Notably in Cohort 1 and 2, 185/236 and 338/606 (64.8% and 55.8%) in remission and 18/40 (45%) with complete endoscopic mucosal healing are Fatiguehigh. Prediction of Fatigue: Multiple logistic regression modelling using clinical parameters alone was poorly predictive of Fatiguehigh (Area under the Curve, AUC =0.62). We further applied Random Forest machine learning approach and generated AUC 0.62 to predict Fatigue high; relatively poor in contrast with AUCs of 0.84, to predict active IBD. (All data presented as medians).
Conclusions: Fatigue remains a major issue even in clinical remission and mucosal healing. Current high dimensional approaches with machine learning are poorly predictive of fatigue. Our ongoing work will incorporate genetic, microbiome and metabolic data to utilise this initial platform to understand the factors that drive IBD-associated fatigue.
1Well-being, Gut Health and Science: Patient-reported outcomes in Inflammatory Bowel Disease (https://edin.ac/3sTGpZT)
| Original language | English |
|---|---|
| Publication status | Published - 21 May 2024 |
| Event | Digestive Diseases Week 2024 - Washington DC, United States Duration: 18 May 2024 → 21 May 2024 https://eposters.ddw.org/ddw/#!*menu=6*browseby=3*sortby=2*ce_id=2673 (Link to Conference Website) |
Conference
| Conference | Digestive Diseases Week 2024 |
|---|---|
| Country/Territory | United States |
| City | Washington DC |
| Period | 18/05/24 → 21/05/24 |
| Internet address |
|