Katrina Wilcox Hagberg, Catherine Vasilakis-Scaramozza, Rebecca Persson, David Neasham, George Kafatos, Susan Jick
Purpose: To evaluate the new Clinical Practice Research Datalink (CPRD) Aurum database, we estimated ‘correctness’ (ie accuracy, validity) and ‘completeness’ (ie presence, missingness) of malignant breast cancer diagnoses recorded in CPRD Aurum compared to external linked data sources: Hospital Episode Statistics (HES) Admitted Patient Care (APC), HES Outpatient (OP), and Cancer Registry (CR), and to the previously validated CPRD GOLD. Methods: Linkage-eligible, female patients with incident malignant breast cancer diagnosis recorded in at least one study data source were selected. Correctness was the proportion of malignant breast cancer cases recorded in CPRD Aurum or GOLD who also had a diagnosis recorded in HES APC/OP (2004– 2019) or CR (2004– 2016). Completeness was estimated by identifying all malignant breast cancer diagnoses in HES APC/OP or CR and calculating the proportion with a concordant diagnosis in CPRD Aurum or GOLD. Results: Compared to HES APC/OP, there were 85,659 and 31,452 eligible patients in CPRD Aurum and GOLD, respectively. Correctness estimates were high (CPRD Aurum 83.5%, GOLD 81.7%). Compared to CR, there were 70,190 and 29,597 eligible patients in CPRD Aurum and GOLD, respectively: correctness was 89.1% for CPRD Aurum and 88.2% for GOLD. Completeness estimates for CPRD Aurum and GOLD were high (> 90%). Diagnoses were recorded in CPRD Aurum within − 7 to 74 days of those in the linked sources. Reasons for discordant diagnostic coding included presence of treatment or other clinical codes only, diagnosis coded after end of follow-up, non-malignant breast cancer in linked data, and administrative codes in lieu of diagnostic codes. Conclusion: These results indicate that correctness and completeness of malignant breast cancer diagnoses in CPRD Aurum were high and similar to CPRD GOLD. This provides confidence in use of CPRD Aurum for research purposes. Where complete case capture is important, researchers should consider linkage to HES APC or CR.
Keywords: CPRD Aurum, CPRD GOLD, breast cancer, validation
{"title":"Correctness and Completeness of Breast Cancer Diagnoses Recorded in UK CPRD Aurum and CPRD GOLD Databases: Comparison to Hospital Episode Statistics and Cancer Registry (Companion Paper 2)","authors":"Katrina Wilcox Hagberg, Catherine Vasilakis-Scaramozza, Rebecca Persson, David Neasham, George Kafatos, Susan Jick","doi":"10.2147/clep.s434829","DOIUrl":"https://doi.org/10.2147/clep.s434829","url":null,"abstract":"<strong>Purpose:</strong> To evaluate the new Clinical Practice Research Datalink (CPRD) Aurum database, we estimated ‘correctness’ (ie accuracy, validity) and ‘completeness’ (ie presence, missingness) of malignant breast cancer diagnoses recorded in CPRD Aurum compared to external linked data sources: Hospital Episode Statistics (HES) Admitted Patient Care (APC), HES Outpatient (OP), and Cancer Registry (CR), and to the previously validated CPRD GOLD.<br/><strong>Methods:</strong> Linkage-eligible, female patients with incident malignant breast cancer diagnosis recorded in at least one study data source were selected. Correctness was the proportion of malignant breast cancer cases recorded in CPRD Aurum or GOLD who also had a diagnosis recorded in HES APC/OP (2004– 2019) or CR (2004– 2016). Completeness was estimated by identifying all malignant breast cancer diagnoses in HES APC/OP or CR and calculating the proportion with a concordant diagnosis in CPRD Aurum or GOLD.<br/><strong>Results:</strong> Compared to HES APC/OP, there were 85,659 and 31,452 eligible patients in CPRD Aurum and GOLD, respectively. Correctness estimates were high (CPRD Aurum 83.5%, GOLD 81.7%). Compared to CR, there were 70,190 and 29,597 eligible patients in CPRD Aurum and GOLD, respectively: correctness was 89.1% for CPRD Aurum and 88.2% for GOLD. Completeness estimates for CPRD Aurum and GOLD were high (> 90%). Diagnoses were recorded in CPRD Aurum within − 7 to 74 days of those in the linked sources. Reasons for discordant diagnostic coding included presence of treatment or other clinical codes only, diagnosis coded after end of follow-up, non-malignant breast cancer in linked data, and administrative codes in lieu of diagnostic codes.<br/><strong>Conclusion:</strong> These results indicate that correctness and completeness of malignant breast cancer diagnoses in CPRD Aurum were high and similar to CPRD GOLD. This provides confidence in use of CPRD Aurum for research purposes. Where complete case capture is important, researchers should consider linkage to HES APC or CR.<br/><br/><strong>Keywords:</strong> CPRD Aurum, CPRD GOLD, breast cancer, validation","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138686870","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Letter for the article The Epidemiology of Bile Acid Diarrhea in Denmark
为《丹麦胆汁酸腹泻的流行病学》一文撰写的信件
{"title":"The Epidemiology of Bile Acid Diarrhea in Denmark [Letter]","authors":"Elanda Fikri","doi":"10.2147/clep.s454145","DOIUrl":"https://doi.org/10.2147/clep.s454145","url":null,"abstract":"Letter for the article The Epidemiology of Bile Acid Diarrhea in Denmark","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138686875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Letter for the article The Epidemiology of Bile Acid Diarrhea in Denmark
为《丹麦胆汁酸腹泻的流行病学》一文撰写的信件
{"title":"The Epidemiology of Bile Acid Diarrhea in Denmark [Letter]","authors":"Demes Nurmayanti, Alkausyari Aziz, Khayan Khayan","doi":"10.2147/clep.s453992","DOIUrl":"https://doi.org/10.2147/clep.s453992","url":null,"abstract":"Letter for the article The Epidemiology of Bile Acid Diarrhea in Denmark","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138686936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Martin L Kårhus, Anne-Marie Ellegaard, Matilde Winther-Jensen, Susanne Hansen, Filip K Knop, Line L Kårhus
Objective: Bile acid diarrhea (BAD) is a socially debilitating disease with frequent bowel movements, urgency, and fecal incontinence as the main symptoms. It is caused by excessive bile acid levels in the colon and is most commonly treated with bile acid sequestrants. It is estimated that 1– 2% of the population suffers from the disease, but only a fraction of these are properly diagnosed with the gold standard ⁷⁵selenium-homotaurocholic acid (SeHCAT) test. Here, we use nationwide registries to describe the demographic characteristics of individuals suffering from BAD in Denmark. Methods: Since the International Classification of Diseases diagnosis code for BAD was not used until 2021, we identified the BAD population by referral to SeHCAT testing followed by a prescription of a bile acid sequestrant (colestyramine, colestipol or colesevelam) within 365 days. The study period was from 2003 to 2021. Results: During the study period, a total of 5264 individuals with BAD were identified with large differences between the five regions in Denmark. The number of prescriptions of colestyramine and colesevelam, the number of SeHCAT tests, and the number of individuals diagnosed with BAD increased during the study period. The BAD population had more co-morbidities and more health care contacts as well as lower levels of education and income compared with age- and sex-matched controls from the general population. Conclusion: Using the Danish registries, we identified a BAD population, which seems to be inferior in health care and socio-economic parameters compared with the Danish general population.
Keywords: Bile acid diarrhea, epidemiology, nationwide registries
目的:胆汁酸腹泻(BAD)是一种使人衰弱的社会性疾病,主要症状为排便频繁、里急后重和大便失禁。它是由结肠中胆汁酸水平过高引起的,最常见的治疗方法是使用胆汁酸螯合剂。据估计,1%- 2% 的人口患有此病,但其中只有一小部分人能够通过黄金标准⁷⁵硒-高陶胆酸(SeHCAT)检测得到正确诊断。在此,我们利用全国性的登记资料来描述丹麦 BAD 患者的人口统计学特征:由于 BAD 的国际疾病分类诊断代码要到 2021 年才开始使用,因此我们通过转诊进行 SeHCAT 检测,并在 365 天内开具胆汁酸螯合剂(可乐定、可乐定或可乐司维兰)处方来确定 BAD 患者。研究时间为 2003 年至 2021 年:在研究期间,共发现了 5264 名胆汁淤积症患者,丹麦的五个地区之间存在很大差异。在研究期间,可乐定和可乐定的处方数量、SeHCAT 检测次数以及被诊断为 BAD 的人数都有所增加。与年龄和性别相匹配的普通人群对照组相比,BAD人群的合并疾病更多,接触的医疗服务更多,教育程度和收入水平也更低:通过丹麦登记,我们发现了一个胆汁酸腹泻人群,与丹麦普通人群相比,该人群的医疗保健和社会经济参数似乎较低:胆汁酸腹泻、流行病学、全国性登记
{"title":"The Epidemiology of Bile Acid Diarrhea in Denmark","authors":"Martin L Kårhus, Anne-Marie Ellegaard, Matilde Winther-Jensen, Susanne Hansen, Filip K Knop, Line L Kårhus","doi":"10.2147/clep.s442054","DOIUrl":"https://doi.org/10.2147/clep.s442054","url":null,"abstract":"<strong>Objective:</strong> Bile acid diarrhea (BAD) is a socially debilitating disease with frequent bowel movements, urgency, and fecal incontinence as the main symptoms. It is caused by excessive bile acid levels in the colon and is most commonly treated with bile acid sequestrants. It is estimated that 1– 2% of the population suffers from the disease, but only a fraction of these are properly diagnosed with the gold standard ⁷⁵selenium-homotaurocholic acid (SeHCAT) test. Here, we use nationwide registries to describe the demographic characteristics of individuals suffering from BAD in Denmark.<br/><strong>Methods:</strong> Since the International Classification of Diseases diagnosis code for BAD was not used until 2021, we identified the BAD population by referral to SeHCAT testing followed by a prescription of a bile acid sequestrant (colestyramine, colestipol or colesevelam) within 365 days. The study period was from 2003 to 2021.<br/><strong>Results:</strong> During the study period, a total of 5264 individuals with BAD were identified with large differences between the five regions in Denmark. The number of prescriptions of colestyramine and colesevelam, the number of SeHCAT tests, and the number of individuals diagnosed with BAD increased during the study period. The BAD population had more co-morbidities and more health care contacts as well as lower levels of education and income compared with age- and sex-matched controls from the general population.<br/><strong>Conclusion:</strong> Using the Danish registries, we identified a BAD population, which seems to be inferior in health care and socio-economic parameters compared with the Danish general population.<br/><br/><strong>Keywords:</strong> Bile acid diarrhea, epidemiology, nationwide registries<br/>","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138547014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Letter for the article Screening History and 7-Year Survival in 32,099 Colorectal Cancer Patients: A Population-Based Cohort Study
致信 32,099 名结直肠癌患者的筛查史和 7 年生存率:基于人群的队列研究
{"title":"Screening History and 7-Year Survival in 32,099 Colorectal Cancer Patients: A Population-Based Cohort Study [Letter]","authors":"Slamet Wardoyo, Taufik Anwar","doi":"10.2147/clep.s451619","DOIUrl":"https://doi.org/10.2147/clep.s451619","url":null,"abstract":"Letter for the article Screening History and 7-Year Survival in 32,099 Colorectal Cancer Patients: A Population-Based Cohort Study","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138547011","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Purpose: The association between body mass index (BMI) and all-cause mortality may vary among hypertensive patients of different ages. This study aimed to investigate the age-dependent association between BMI and all-cause mortality among patients with hypertension. Patients and Methods: A total of 212,394 participants with hypertension aged 20–85 years from Minhang Hypertension Standardization Management System in Shanghai of China were included. Follow-up began at the time when individuals were first recorded and ended at death, loss to follow-up, or December 31, 2018, whichever came first. Additive Cox proportional hazards models with thin plate smoothing functions and conventional Cox proportional hazards models were adopted to examine the relationship between BMI, age, and mortality. The joint effect of BMI and age on mortality was assessed using a bivariate response model. Results: We found that the BMI–mortality relationship followed a U-shaped pattern, with a trough at 26–27 kg/m 2 . Compared with normal weight, underweight was associated with a 50% increased risk of premature mortality (hazard ratio 1.50, 95% confidence interval 1.43 to 1.57). Whereas among those aged 45–59 and 60–85 years, overweight was associated with 13% (0.87, 0.80 to 0.94) and 18% (0.82, 0.80 to 0.84) reduction in risk of death, respectively. Bivariate response model indicated a significant interaction between BMI and age (P < 0.05). Among younger and older patients, we found a descending trend for mortality risk, with BMI increasing at different age levels, whereas a reverse J-shaped relation pattern was observed among middle-aged patients. Conclusion: The impact of BMI on all-cause mortality in hypertensive patients varies with age, and moderate weight gain may benefit longevity in middle-aged and older patients.
{"title":"Age-Dependent Association Between Body Mass Index and All-Cause Mortality Among Patients with Hypertension: A Longitudinal Population-Based Cohort Study in China","authors":"Yifang Huang, Jiahuan Peng, Weibing Wang, Xueying Zheng, Guoyou Qin, Huilin Xu","doi":"10.2147/clep.s442162","DOIUrl":"https://doi.org/10.2147/clep.s442162","url":null,"abstract":"Purpose: The association between body mass index (BMI) and all-cause mortality may vary among hypertensive patients of different ages. This study aimed to investigate the age-dependent association between BMI and all-cause mortality among patients with hypertension. Patients and Methods: A total of 212,394 participants with hypertension aged 20–85 years from Minhang Hypertension Standardization Management System in Shanghai of China were included. Follow-up began at the time when individuals were first recorded and ended at death, loss to follow-up, or December 31, 2018, whichever came first. Additive Cox proportional hazards models with thin plate smoothing functions and conventional Cox proportional hazards models were adopted to examine the relationship between BMI, age, and mortality. The joint effect of BMI and age on mortality was assessed using a bivariate response model. Results: We found that the BMI–mortality relationship followed a U-shaped pattern, with a trough at 26–27 kg/m 2 . Compared with normal weight, underweight was associated with a 50% increased risk of premature mortality (hazard ratio 1.50, 95% confidence interval 1.43 to 1.57). Whereas among those aged 45–59 and 60–85 years, overweight was associated with 13% (0.87, 0.80 to 0.94) and 18% (0.82, 0.80 to 0.84) reduction in risk of death, respectively. Bivariate response model indicated a significant interaction between BMI and age (P < 0.05). Among younger and older patients, we found a descending trend for mortality risk, with BMI increasing at different age levels, whereas a reverse J-shaped relation pattern was observed among middle-aged patients. Conclusion: The impact of BMI on all-cause mortality in hypertensive patients varies with age, and moderate weight gain may benefit longevity in middle-aged and older patients.","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138524532","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zheng Wang, Dahai Yu, Uchechukwu Levi Osuagwu, Karen Pickering, John Baker, Richard Cutfield, Yamei Cai, Brandon J Orr-Walker, Gerhard Sundborn, Zhanzheng Zhao, David Simmons
Purpose: We aimed to examine socioeconomic inequality (SI) in cause-specific outcomes among adults with impaired glucose tolerance (IGT) and/or Impaired fasting glucose (IFG) in New Zealand (NZ) over 25 years. Patients and Methods: A population-based open cohort was derived from Diabetes Care Support Service in NZ with national databases linkage. Patients aged ≥ 18 years with IGT and/or IFG were enrolled between 01/01/1994 and 31/07/2018 and followed up until death or 31/12/2018. Incident outcomes (all-cause, premature, cardiovascular, and cancer death; cardiovascular, myocardial infarction, stroke, heart failure, and end-stage kidney disease hospitalization) by demographic, anthropometric, socioeconomic status, clinical measurements, enrol-time-periods, and IGT/IFG were evaluated. Adjusted incidence rate ratios, absolute risk difference, and SI measurements (slope and relative index of inequality) were estimated using Age-Period-Cohort models. Results: 29,894 patients (58.5 (SD 14.3) years mean age; 52.2% female) were enrolled with 5.6 (IQR: 4.4– 7.4) years of median follow-up. Mortality rates decreased, whereas hospitalization (except myocardial infarction) rates increased. SI was significant for each outcome. Higher mortality and hospitalization rates and worsened SI were common in men, older, the most deprived, and M&amacrori patients, as well as patients with obesity, current smoking, with both IFG and IGT, and greater metabolic derangement (higher systolic blood pressure, lipids, and HbA1c, and lower level of mean arterial pressure). Conclusion: Enhanced management strategies are necessary for people with IGT and/or IFG to address persisting SI, especially for men, older people, current smokers, NZ European and M&amacrori patients, patients with obesity, or with any abnormal metabolic measurements.
{"title":"Health Inequality in Eight Adverse Outcomes Over a 25-Year Period in a Multi-Ethnic Population in New Zealand Population with Impaired Glucose Tolerance and/or Impaired Fasting Glucose: An Age-Period-Cohort Analysis","authors":"Zheng Wang, Dahai Yu, Uchechukwu Levi Osuagwu, Karen Pickering, John Baker, Richard Cutfield, Yamei Cai, Brandon J Orr-Walker, Gerhard Sundborn, Zhanzheng Zhao, David Simmons","doi":"10.2147/clep.s433714","DOIUrl":"https://doi.org/10.2147/clep.s433714","url":null,"abstract":"<strong>Purpose:</strong> We aimed to examine socioeconomic inequality (SI) in cause-specific outcomes among adults with impaired glucose tolerance (IGT) and/or Impaired fasting glucose (IFG) in New Zealand (NZ) over 25 years.<br/><strong>Patients and Methods:</strong> A population-based open cohort was derived from Diabetes Care Support Service in NZ with national databases linkage. Patients aged ≥ 18 years with IGT and/or IFG were enrolled between 01/01/1994 and 31/07/2018 and followed up until death or 31/12/2018. Incident outcomes (all-cause, premature, cardiovascular, and cancer death; cardiovascular, myocardial infarction, stroke, heart failure, and end-stage kidney disease hospitalization) by demographic, anthropometric, socioeconomic status, clinical measurements, enrol-time-periods, and IGT/IFG were evaluated. Adjusted incidence rate ratios, absolute risk difference, and SI measurements (slope and relative index of inequality) were estimated using Age-Period-Cohort models.<br/><strong>Results:</strong> 29,894 patients (58.5 (SD 14.3) years mean age; 52.2% female) were enrolled with 5.6 (IQR: 4.4– 7.4) years of median follow-up. Mortality rates decreased, whereas hospitalization (except myocardial infarction) rates increased. SI was significant for each outcome. Higher mortality and hospitalization rates and worsened SI were common in men, older, the most deprived, and M&amacrori patients, as well as patients with obesity, current smoking, with both IFG and IGT, and greater metabolic derangement (higher systolic blood pressure, lipids, and HbA1c, and lower level of mean arterial pressure).<br/><strong>Conclusion:</strong> Enhanced management strategies are necessary for people with IGT and/or IFG to address persisting SI, especially for men, older people, current smokers, NZ European and M&amacrori patients, patients with obesity, or with any abnormal metabolic measurements.<br/><br/><strong>Keywords:</strong> impaired glucose tolerance, impaired fasting glucose, New Zealand, M&amacrori, ethnic disparity, socioeconomic inequality, mortality, cardiovascular diseases, heart failure, cancer, end-stage renal disease<br/>","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138524603","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yuchen Gao, Chunrong Wang, Wenhao Dong, Bianfang Li, Jianhui Wang, Jun Li, Yu Tian, Jia Liu, Yuefu Wang
Background: To derive and validate a machine learning (ML) prediction model of acute kidney injury (AKI) that could be used for AKI surveillance and management to improve clinical outcomes. Methods: This retrospective cohort study was conducted in Fuwai Hospital, including patients aged 18 years and above undergoing cardiac surgery admitted between January 1, 2017, and December 31, 2018. Seventy percent of the observations were randomly selected for training and the remaining 30% for testing. The demographics, comorbidities, laboratory examination parameters, and operation details were used to construct a prediction model for AKI by logistic regression and eXtreme gradient boosting (Xgboost). The discrimination of each model was assessed on the test cohort by the area under the receiver operator characteristic (AUROC) curve, while calibration was performed by the calibration plot. Results: A total of 15,880 patients were enrolled in this study, and 4845 (30.5%) had developed AKI. Xgboost model had the higher discriminative ability compared with logistic regression (AUROC, 0.849 [95% CI, 0.837– 0.861] vs 0.803[95% CI 0.790– 0.817], P< 0.001) in the test dataset. The estimated glomerular filtration (eGFR) and creatine on intensive care unit (ICU) arrival are the two most important prediction parameters. A SHAP summary plot was used to illustrate the effects of the top 15 features attributed to the Xgboost model. Conclusion: ML models can provide clinical decision support to determine which patients should focus on perioperative preventive treatment to preemptively reduce acute kidney injury by predicting which patients are not at risk.
背景:推导并验证急性肾损伤(AKI)的机器学习(ML)预测模型,该模型可用于AKI的监测和管理,以改善临床结果。方法:本回顾性队列研究在阜外医院进行,纳入2017年1月1日至2018年12月31日住院的18岁及以上心脏手术患者。70%的观察对象被随机选择用于训练,剩下的30%用于测试。利用人口统计学、合并症、实验室检查参数和手术细节,采用logistic回归和极限梯度增强(Xgboost)方法构建AKI预测模型。通过接收算子特征曲线下面积(AUROC)评估各模型在试验队列中的判别性,并通过校准图进行校准。结果:本研究共纳入15880例患者,其中4845例(30.5%)发生AKI。与logistic回归相比,Xgboost模型具有更高的判别能力(AUROC, 0.849 [95% CI, 0.837 - 0.861] vs 0.803[95% CI 0.790 - 0.817], P<0.001)。重症监护病房(ICU)到达时估计的肾小球滤过(eGFR)和肌酸是两个最重要的预测参数。使用SHAP总结图来说明归因于Xgboost模型的前15个特性的影响。结论:ML模型可以通过预测哪些患者无危险,为确定哪些患者围手术期应重点进行预防性治疗,先发制人地减少急性肾损伤提供临床决策支持。
{"title":"An Explainable Machine Learning Model to Predict Acute Kidney Injury After Cardiac Surgery: A Retrospective Cohort Study","authors":"Yuchen Gao, Chunrong Wang, Wenhao Dong, Bianfang Li, Jianhui Wang, Jun Li, Yu Tian, Jia Liu, Yuefu Wang","doi":"10.2147/clep.s404580","DOIUrl":"https://doi.org/10.2147/clep.s404580","url":null,"abstract":"<strong>Background:</strong> To derive and validate a machine learning (ML) prediction model of acute kidney injury (AKI) that could be used for AKI surveillance and management to improve clinical outcomes.<br/><strong>Methods:</strong> This retrospective cohort study was conducted in Fuwai Hospital, including patients aged 18 years and above undergoing cardiac surgery admitted between January 1, 2017, and December 31, 2018. Seventy percent of the observations were randomly selected for training and the remaining 30% for testing. The demographics, comorbidities, laboratory examination parameters, and operation details were used to construct a prediction model for AKI by logistic regression and eXtreme gradient boosting (Xgboost). The discrimination of each model was assessed on the test cohort by the area under the receiver operator characteristic (AUROC) curve, while calibration was performed by the calibration plot.<br/><strong>Results:</strong> A total of 15,880 patients were enrolled in this study, and 4845 (30.5%) had developed AKI. Xgboost model had the higher discriminative ability compared with logistic regression (AUROC, 0.849 [95% CI, 0.837– 0.861] vs 0.803[95% CI 0.790– 0.817], <em>P</em>< 0.001) in the test dataset. The estimated glomerular filtration (eGFR) and creatine on intensive care unit (ICU) arrival are the two most important prediction parameters. A SHAP summary plot was used to illustrate the effects of the top 15 features attributed to the Xgboost model.<br/><strong>Conclusion:</strong> ML models can provide clinical decision support to determine which patients should focus on perioperative preventive treatment to preemptively reduce acute kidney injury by predicting which patients are not at risk.<br/><br/>","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138524565","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lars Frost, Albert Joensen, Ulla Dam-Schmidt, I. Qvist, Margit Brinck, Axel Brandes, Ulla Davidsen, Ole Pedersen, Dorte Damgaard, Inge Mølgaard, Robert Bedsted, Anders Damgaard Møller Schlünsen, Miriam Grijota Chousa, Julie Andersen, Asger Pedersen, S. P. Johnsen, N. Vinter
{"title":"The Danish Atrial Fibrillation Registry: A Multidisciplinary National Pragmatic Initiative for Monitoring and Supporting Quality of Care Based on Data Retrieved from Administrative Registries","authors":"Lars Frost, Albert Joensen, Ulla Dam-Schmidt, I. Qvist, Margit Brinck, Axel Brandes, Ulla Davidsen, Ole Pedersen, Dorte Damgaard, Inge Mølgaard, Robert Bedsted, Anders Damgaard Møller Schlünsen, Miriam Grijota Chousa, Julie Andersen, Asger Pedersen, S. P. Johnsen, N. Vinter","doi":"10.2147/clep.s443473","DOIUrl":"https://doi.org/10.2147/clep.s443473","url":null,"abstract":"","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139023429","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-11-29eCollection Date: 2023-01-01DOI: 10.2147/CLEP.S433983
Pernille Tilma Tonnesen, Kevin Kris Warnakula Olesen, Pernille Gro Thrane, Christine Gyldenkerne, Christian Daugaard Peters, Niels Henrik Buus, Michael Maeng
Purpose: To investigate the interplay between chronic kidney disease (CKD) and coronary artery disease (CAD) on the incidence of cardiovascular events in patients with suspected chronic coronary syndrome (CCS).
Patients and methods: Patients with suspected CCS who underwent first-time coronary angiography in Western Denmark between 2003 and 2016 were included in this cohort study. Moreover, an age- and sex-matched general population cohort was established. Patients were stratified according to estimated glomerular filtration rate (eGFR). Presence of CAD was defined as ≥1 obstructive stenosis or non-obstructive diffuse disease. Major adverse cardiovascular events (MACE) were defined as a composite of myocardial infarction, ischemic stroke, and cardiac death.
Results: A total of 42,611 patients were included with a median follow-up of 7.3 years. Patients without and with CAD had MACE rates per 100 person-years that were 0.52 and 1.67 for eGFR ≥90 mL/min/1.73 m2, 0.68 and 2.09 for eGFR 60-89 mL/min/1.73 m2, 1.27 and 3.85 for eGFR 30-59 mL/min/1.73 m2, and 2.27 and 6.92 for eGFR <30 mL/min/1.73 m2. Comparing to eGFR ≥90 mL/min/1.73 m2, the adjusted incidence rate ratios for MACE were 1.29 (1.10-1.51) for eGFR 60-89 mL/min/1.73 m2, 1.86 (1.49-2.33) for eGFR 30-59 mL/min/1.73 m2, and 3.57 (1.92-6.67) for eGFR <30 mL/min/1.73 m2 in patients without CAD, and 1.11 (1.03-1.20), 1.71 (1.55-1.90), and 2.46 (1.96-3.09) in patients with CAD. The inverse relationship between kidney function and risk of MACE was confirmed when comparing patients with and without CAD to matched individuals in the general population.
Conclusion: Absence of CAD is a strong negative predictor of major adverse cardiovascular events in patients with CKD.
{"title":"Absence of Coronary Artery Disease is a Strong Negative Predictor of Major Adverse Cardiovascular Events in Patients with Chronic Kidney Disease.","authors":"Pernille Tilma Tonnesen, Kevin Kris Warnakula Olesen, Pernille Gro Thrane, Christine Gyldenkerne, Christian Daugaard Peters, Niels Henrik Buus, Michael Maeng","doi":"10.2147/CLEP.S433983","DOIUrl":"10.2147/CLEP.S433983","url":null,"abstract":"<p><strong>Purpose: </strong>To investigate the interplay between chronic kidney disease (CKD) and coronary artery disease (CAD) on the incidence of cardiovascular events in patients with suspected chronic coronary syndrome (CCS).</p><p><strong>Patients and methods: </strong>Patients with suspected CCS who underwent first-time coronary angiography in Western Denmark between 2003 and 2016 were included in this cohort study. Moreover, an age- and sex-matched general population cohort was established. Patients were stratified according to estimated glomerular filtration rate (eGFR). Presence of CAD was defined as ≥1 obstructive stenosis or non-obstructive diffuse disease. Major adverse cardiovascular events (MACE) were defined as a composite of myocardial infarction, ischemic stroke, and cardiac death.</p><p><strong>Results: </strong>A total of 42,611 patients were included with a median follow-up of 7.3 years. Patients without and with CAD had MACE rates per 100 person-years that were 0.52 and 1.67 for eGFR ≥90 mL/min/1.73 m<sup>2</sup>, 0.68 and 2.09 for eGFR 60-89 mL/min/1.73 m<sup>2</sup>, 1.27 and 3.85 for eGFR 30-59 mL/min/1.73 m<sup>2</sup>, and 2.27 and 6.92 for eGFR <30 mL/min/1.73 m<sup>2</sup>. Comparing to eGFR ≥90 mL/min/1.73 m<sup>2</sup>, the adjusted incidence rate ratios for MACE were 1.29 (1.10-1.51) for eGFR 60-89 mL/min/1.73 m<sup>2</sup>, 1.86 (1.49-2.33) for eGFR 30-59 mL/min/1.73 m<sup>2</sup>, and 3.57 (1.92-6.67) for eGFR <30 mL/min/1.73 m<sup>2</sup> in patients without CAD, and 1.11 (1.03-1.20), 1.71 (1.55-1.90), and 2.46 (1.96-3.09) in patients with CAD. The inverse relationship between kidney function and risk of MACE was confirmed when comparing patients with and without CAD to matched individuals in the general population.</p><p><strong>Conclusion: </strong>Absence of CAD is a strong negative predictor of major adverse cardiovascular events in patients with CKD.</p>","PeriodicalId":10362,"journal":{"name":"Clinical Epidemiology","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2023-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10693773/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138482094","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}