Suk Yun Kang1, Sang Won Seo2, Joo Yong Kim3. 1. Department of Neurology, Dongtan Sacred Heart Hospital, Hallym University College of Medicine, 7, Keunjaebong-gil, Hwaseong, Gyeonggi-do, 18450, Republic of Korea. sukyunkang@hanmail.net. 2. Business Information Team, Ilsong Educational Foundation, Hallym University Medical Center, Seoul, Republic of Korea. 3. Department of Neurology, Dongtan Sacred Heart Hospital, Hallym University College of Medicine, 7, Keunjaebong-gil, Hwaseong, Gyeonggi-do, 18450, Republic of Korea.
Abstract
BACKGROUND: Postoperative delirium (POD) in older adults is a very serious complication. Due to the complexity of too many risk factors (RFs), an overall assessment of RFs may be needed. The aim of this study was to evaluate comprehensively the RFs of POD regardless of the organ undergoing operation, efficiently incorporating the concept of comprehensive big data using a smart clinical data warehouse (CDW). METHODS: We reviewed the electronic medical data of inpatients aged 65 years or older who underwent major surgery between January 2010 and June 2016 at Hallym University Sacred Heart Hospital. The following six major operation types were selected: cardiac, stomach, colorectal, hip, knee, and spine. Clinical features, laboratory findings, perioperative variables, and medication history were compared between patients without POD and with POD. RESULTS: Six hundred eighty-six of 3634 patients (18.9%) developed POD. In multivariate logistic regression analysis, common, independent RFs of POD were as follows (descending order of odds ratio): operation type ([hip] OR 8.858, 95%CI 3.432-22.863; p = 0.000; [knee] OR 7.492, 95%CI 2.739-20.487; p = 0.000; [spine] OR 6.919, 95%CI 2.687-17.815; p = 0.000; [colorectal] OR 2.037, 95%CI 0.784-5.291; p = 0.144; [stomach] OR 1.500, 95%CI 0.532-4.230; p = 0.443; [cardiac] reference), parkinsonism (OR 2.945, 95%CI 1.564-5.547; p = 0.001), intensive care unit stay (OR 1.675, 95%CI 1.354-2.072; p = 0.000), stroke history (OR 1.591, 95%CI 1.112-2.276; p = 0.011), use of hypnotics and sedatives (OR 1.307, 95%CI 1.072-1.594; p = 0.008), higher creatinine (OR 1.107, 95%CI 1.004-1.219; p = 0.040), lower hematocrit (OR 0.910, 95%CI 0.836-0.991; p = 0.031), older age (OR 1.053, 95%CI 1.037-1.069; p = 0.000), and lower body mass index (OR 0.967, 95%CI 0.942-0.993; p = 0.013). The use of analgesics (OR 0.644, 95%CI 0.467-0.887; p = 0.007) and antihistamines/antiallergics (OR 0.764, 95%CI 0.622-0.937; p = 0.010) were risk-reducing factors. Operation type with the highest odds ratio for POD was orthopedic surgery. CONCLUSIONS: Big data analytics could be applied to evaluate RFs in electronic medical records. We identified common RFs of POD, regardless of operation type. Big data analytics may be helpful for the comprehensive understanding of POD RFs, which can help physicians develop a general plan to prevent POD.
BACKGROUND:Postoperative delirium (POD) in older adults is a very serious complication. Due to the complexity of too many risk factors (RFs), an overall assessment of RFs may be needed. The aim of this study was to evaluate comprehensively the RFs of POD regardless of the organ undergoing operation, efficiently incorporating the concept of comprehensive big data using a smart clinical data warehouse (CDW). METHODS: We reviewed the electronic medical data of inpatients aged 65 years or older who underwent major surgery between January 2010 and June 2016 at Hallym University Sacred Heart Hospital. The following six major operation types were selected: cardiac, stomach, colorectal, hip, knee, and spine. Clinical features, laboratory findings, perioperative variables, and medication history were compared between patients without POD and with POD. RESULTS: Six hundred eighty-six of 3634 patients (18.9%) developed POD. In multivariate logistic regression analysis, common, independent RFs of POD were as follows (descending order of odds ratio): operation type ([hip] OR 8.858, 95%CI 3.432-22.863; p = 0.000; [knee] OR 7.492, 95%CI 2.739-20.487; p = 0.000; [spine] OR 6.919, 95%CI 2.687-17.815; p = 0.000; [colorectal] OR 2.037, 95%CI 0.784-5.291; p = 0.144; [stomach] OR 1.500, 95%CI 0.532-4.230; p = 0.443; [cardiac] reference), parkinsonism (OR 2.945, 95%CI 1.564-5.547; p = 0.001), intensive care unit stay (OR 1.675, 95%CI 1.354-2.072; p = 0.000), stroke history (OR 1.591, 95%CI 1.112-2.276; p = 0.011), use of hypnotics and sedatives (OR 1.307, 95%CI 1.072-1.594; p = 0.008), higher creatinine (OR 1.107, 95%CI 1.004-1.219; p = 0.040), lower hematocrit (OR 0.910, 95%CI 0.836-0.991; p = 0.031), older age (OR 1.053, 95%CI 1.037-1.069; p = 0.000), and lower body mass index (OR 0.967, 95%CI 0.942-0.993; p = 0.013). The use of analgesics (OR 0.644, 95%CI 0.467-0.887; p = 0.007) and antihistamines/antiallergics (OR 0.764, 95%CI 0.622-0.937; p = 0.010) were risk-reducing factors. Operation type with the highest odds ratio for POD was orthopedic surgery. CONCLUSIONS: Big data analytics could be applied to evaluate RFs in electronic medical records. We identified common RFs of POD, regardless of operation type. Big data analytics may be helpful for the comprehensive understanding of POD RFs, which can help physicians develop a general plan to prevent POD.
Entities:
Keywords:
Clinical data warehouse; Delirium; Operation; Parkinsonism; Risk factor; Type of surgery
Authors: Maria Teresa Infante; Matteo Pardini; Maurizio Balestrino; Cinzia Finocchi; Laura Malfatto; Giuseppe Bellelli; Giovanni Luigi Mancardi; Carlo Gandolfo; Carlo Serrati Journal: Neurol Sci Date: 2017-02-01 Impact factor: 3.307
Authors: Aladine A Elsamadicy; Owoicho Adogwa; Michael Ongele; Amanda R Sergesketter; Aaron Tarnasky; David E T Lubkin; Nicolas Drysdale; Joseph Cheng; Carlos A Bagley; Isaac O Karikari Journal: World Neurosurg Date: 2018-01-31 Impact factor: 2.104