Shah Rajesh, Kelley John, Amundsen Tyson, Coggins Kenneth, Edwards Audrene, Johnson Christopher M
Division of Gastroenterology, Baylor Scott and White Medical Center - Austin, Austin, Texas.
Department of Medicine, Baylor Scott and White Medical Center - Temple, Temple, Texas.
Proc (Bayl Univ Med Cent). 2022 Dec 21;36(2):165-170. doi: 10.1080/08998280.2022.2156025. eCollection 2023.
There is a growing recognition that social determinants of health (SDOH) influence outcomes in patients with chronic diseases. This study aimed to investigate the influence of SDOH on outcomes in patients with inflammatory bowel disease (IBD). We conducted a retrospective cohort study of adult patients with IBD from 1996 to 2019. Patients were identified using ICD-10 codes for ulcerative colitis and Crohn's disease, and chart review was performed to validate the diagnosis and extract clinical information. SDOH factors including food security, financial resources, and transportation were self-reported by the patient. Random forest models were trained and tested in R to predict either IBD-related hospitalization or surgery. A total of 175 patients were studied, and the majority reported no financial resource, food security, or transportation concerns. For the model using clinical predictors, the sensitivity was 0.68 and specificity was 0.77 with an area under the receiver operating characteristic curve (AUROC) of 0.77. The model's performance did not significantly improve with the addition of SDOH information (AUROC of 0.78); however, model performance did vary by phenotype (AUROC of 0.86 for patients with Crohn's disease and AUROC of 0.68 for patients with ulcerative colitis). Further research is needed to understand the role of SDOH factors and IBD-related outcomes.
人们越来越认识到健康的社会决定因素(SDOH)会影响慢性病患者的治疗结果。本研究旨在调查SDOH对炎症性肠病(IBD)患者治疗结果的影响。我们对1996年至2019年的成年IBD患者进行了一项回顾性队列研究。使用国际疾病分类第十版(ICD - 10)中溃疡性结肠炎和克罗恩病的编码来识别患者,并通过病历审查来验证诊断并提取临床信息。患者自行报告包括食品安全、财务资源和交通等SDOH因素。在R语言中训练并测试随机森林模型,以预测IBD相关的住院或手术情况。总共研究了175名患者,大多数患者表示没有财务资源、食品安全或交通方面的担忧。对于使用临床预测指标的模型,敏感性为0.68,特异性为0.77,受试者工作特征曲线下面积(AUROC)为0.77。加入SDOH信息后,模型的性能没有显著改善(AUROC为0.78);然而,模型性能因疾病表型而异(克罗恩病患者的AUROC为0.86,溃疡性结肠炎患者的AUROC为0.68)。需要进一步研究以了解SDOH因素与IBD相关治疗结果的作用。