Methods for knowledge extraction from a clinical database on liver diseases

There are many kinds of Chinese traditional patent medicine used in clinical practice and many adverse events have been reported by clinical professionals. Chinese patent medicine's safety problems are the most concerned by patients and physicians. At present, many researchers have studied re-evaluation methods about post marketing Chinese medicine safety inside and outside China. However, it is rare that using data from hospital information system (HIS) to re-evaluating post marketing Chinese traditional patent medicine safety problems. HIS database in real world is a good resource with rich information to research medicine safety. This study planed to analyze HIS data selected from ten top general hospitals in Beijing, formed a large HIS database in real world with a capacity of 1 000 000 cases in total after a series of data cleaning and integrating procedures. This study could be a new project that using information to evaluate traditional Chinese medicine safety based on HIS database. A clear protocol has been completed as for the first step for the whole study. The protocol is as follows. First of all, separate each of the Chinese traditional patent medicines existing in the total HIS database as a single database. Secondly, select some related laboratory tests indexes as the safety evaluating outcomes, such as routine blood, routine urine, feces routine, conventional coagulation, liver function, kidney function and other tests. Thirdly, use the data mining method to analyze those selected safety outcomes which had abnormal change before and after using Chinese patent medicines. Finally, judge the relationship between those abnormal changing and Chinese patent medicine. We hope this method could imply useful information to Chinese medicine researchers interested in safety evaluation of traditional Chinese medicine.

In the field of cancer, the ICD-10 coding convention is based on the site of a neoplasm in the body and usually ignores the morphology, thus the same code may be assigned to tumors of different morphologic types in an organ. Nowadays, all general (provincial) and center hospitals in Thailand are equipped with the hospital information system (HIS) database. This study aimed to find the characteristics and magnitude of agreement represented by the positive predictive value (PPV) of provisional cancer diagnoses in the HIS database in Pattani Hospital in Thailand in comparison with the final cancer diagnosis of the ICD-10 codes generated from a well established cancer registry in Songklanagarind Hospital, the medical school hospital of Prince of Songkla University. Data on cancer patients residing in Pattani province who visited Pattani Hospital from January 2007 to May 2011 were obtained from the HIS database. The ICD-10 codes of the HIS computer database of Pattani Hospital were compared against the ICD-10 codes of the same person recorded in the hospital-based cancer registry of Songklanagarind Hospital. The degree of agreement or positive predictive value (PPV) was calculated for each sex and for both sexes combined. A total of 313 cases (15.9%) could be matched in the two databases. Some 222 cases, 109 males and 113 females, fulfilled the criteria of referral from Pattani to Songklanagarind Hospitals. Of 109 male cancer cases, 76 had the same ICD-10 codes in both hospitals, thus, the PPV was 69.7% (95%CI: 60.2-78.2%). Agreement in 76 out of 113 females gave a PPV of 67.3% (95%CI: 57.8-75.8%). The two percentages were found non-significant with Fisher's exact p-value of 0.773. The PPV for combined cases of both sexes was 68.5% (95%CI: 61.9-74.5%). Changes in final diagnosis in the referral system are common, thus the summary statistics of a hospital without full investigation facilities must be used with care, as the statistics are biased towards simple diseases able to be investigated by available facilities. A systematic feedback of patient information from a tertiary to a referring hospital should be considered to increase the accuracy of statistics and to improve the comprehensive care of cancer patients.

Delayed patient recruitment is a common problem in clinical studies. Hospital information systems (HIS) contain data items relevant for inclusion or exclusion criteria of these studies. We developed and assessed a system to support patient recruitment using HIS data. We developed a workflow integrated in our HIS to notify study physicians about potential trial subjects. Automatic HIS database queries based on inclusion and exclusion criteria for each clinical study are performed regularly and generate e-mail notifications via a communication server. Study physicians can verify eligibility with a specific HIS study module. The system performance was assessed with a survey addressing utility, usability, stability, change in recruitment rate, and estimated time savings. During 10 months of operation, 1328 notifications were generated and 329 enrollments (25%) were documented for seven studies. Precision of alerts depends on availability of appropriate HIS items. Utility and usability were assessed as good, and stability as excellent. Users reported an increased patient recruitment rate for three studies. Three studies reported an estimated time saving of 10 min per recruited patient. The main perceived benefit was systematic identification of potentially eligible patients without time-consuming patient screening procedures in the different parts of the hospital. Notifications about potentially eligible patients depend on HIS data quality regarding inclusion/exclusion criteria, in particular, completeness, timeliness, and validity. Routine HIS data can support patient recruitment for clinical studies by means of an automated notification workflow and efficient access to clinical data.

The primary objective of this study is to investigate various applications of artificial intelligence (AI) and statistical methodologies for analyzing and managing peritoneal metastases (PM) caused by gastrointestinal cancers. Relevant keywords and search criteria were comprehensively researched on PubMed and Google Scholar to identify articles and reviews related to the topic. The AI approaches considered were conventional machine learning (ML) and deep learning (DL) models, and the relevant statistical approaches included biostatistics and logistic models. The systematic literature review yielded nearly 30 articles meeting the predefined criteria. Analyses of these studies showed that AI methodologies consistently outperformed traditional statistical approaches. In the AI approaches, DL consistently produced the most precise results, while classical ML demonstrated varied performance but maintained high predictive accuracy. The sample size was the recurring factor that increased the accuracy of the predictions for models of the same type. AI and statistical approaches can detect PM developing among patients with gastrointestinal cancers. Therefore, if clinicians integrated these approaches into diagnostics and prognostics, they could better analyze and manage PM, enhancing clinical decision-making and patients' outcomes. Collaboration across multiple institutions would also help in standardizing methods for data collection and allowing consistent results.

Comprehensive electronic patient records have been established in major hospitals where data analyses in clinical databases are of interest for retrospective studies. In addition, there are a variety of other features which may support research and teaching, as is illustrated by some of the functions implemented within the Kisim system at Zurich University Hospital. Search tools for diagnosis, procedures or DRGs are useful in identifying and recruiting patients for student courses and clinical case presentations. Electronic boards improve the organisation and individual preparation of meetings and provide long-term documentation. Flagging the records of patients participating in prospective clinical trials furnishes linked information for the team treating these patients. Clinical data warehouses allow selection, analysis, presentation and export of data, and also data mining to extract previously unknown, potentially valuable information. Computer-assisted structured acquisition of data and scores enhances quantitative documentation over time. Decision support and alerting become increasingly important and need to be considered in clinical studies; we are evaluating an algorithm for automatic alerts to prevent venous thromboembolism in hospitalised patients. In conclusion, implementation of various functions in clinical information systems may provide significant support for both teaching and research.

Establishment of safety evidence for Xingxue® Shuxuening injection

Towards integration of clinical decision support in commercial hospital information systems using distributed, reusable software and knowledge components

Liver transplantation for alcoholic hepatitis in the United States: Excellent outcomes with profound temporal and geographic variation in frequency.

Prevention of Missing Data in Clinical Research Studies

It is not enough merely to manage medical information. It is difficult to justify the cost of hospital information systems (HIS) or intensive care unit (ICU) patient data management systems (PDMS) on this basis alone. The real benefit of an integrated HIS or PDMS is in decision support. We recently went to the bedside of a critically ill patient and counted the current information categories (not repeated measures) that were reviewed for physician decision-making. The total was in excess of 236! Dr. Eddy summarized it best: ‘It is simply unrealistic to think that individuals can synthesize in their head scores of pieces of evidence, accurately estimate the outcomes of different options, and accurately judge the desirability of those outcomes for patients… All confirm what would be expected from common sense: The complexity of modern medicine exceeds the inherent limitations of the unaided human mind” [1]. Although there are a variety of HIS and ICU PDMS systems available, there are few that provide ICU decision support.

Integrated clinical and laboratory databases

Objective To judge the prognosis of adult recipients after first liver transplantation by stepwise discriminant analysis,and screen out the main influencing factors.Methods The clinical data of 221 patients who received liver transplantation at the Nanfang Hospital of Southern Medical University were retrospectively analyzed.A total of 218 patients who met the criteria were divided into the training samples (188 patients admitted from August 2004 to June 2010) and checking samples (30 patients admitted from July 2010 to February 2011),and then all patients were re-divided into dead group (survival time ≤ 90 days,34 patients) and surviving group (survival time ＞ 90 days,184 patients).Factors which had significant difference after the univariate analysis was further analyzed by the stepwise discriminant analysis method.All data were analyzed by the t test,rank sum test,chi-square test or Fisher exact probability test.Results The ages of the recipients in the dead group and the living group were (54 ± 11)years and (51 ± 11)years,respectively,with no significant difference between the 2 groups (t =-1.681,P ＞ 0.05).The preoperative levels of hemoglobin in the dead group and the living group were 106.7 g/L and 119.2 g/L,respectively,with a significant difference between the 2 groups (t =2.809,P ＜ 0.05).There were significant differences in the levels of serum creatinine,urea nitrogen,albumin,total bilirubin,indirect bilirubin,Na+,prothrombin time,activated partial thromboplastin time,international normalized ratio,fibrinogen,prothrombin activity,platelet,nutrition risk index,model for end-stage liver disease score,number of patients with preoperative hepatic encephalopathy (HE),preoperative hepatorenal syndrome (HRS),preoperative digestive tract bleeding,preoperative infection,preoperative diabetes,Child-Turcotte-Pugh score,cardiac function classification and anesthesia risk rating operation time,anhepatic time,volume of intraoperative blood transfusion,volume of peritoneal effusion ; intraoperative urine output,between the 2 groups (Z =-2.277,-2.595,-3.290,-3.486,-2.562,-2.577,-3.670,-3.882,-3.625,-3.557,-3.837,-1.974,-3.693,-3.815,x2 =19.632,9.756,28.143,Z =-4.175,-3.905,-4.865,-3.564,-5.822,P ＜ 0.05).Preoperative HE,preoperative HRS,duration of operation,intraoperative blood transfusion and intraoperative volume of urine were the independent influencing factors of early prognosis after liver transplantation.The standardized partial regression coefficients were 0.146,0.188,0.257,0.181,-0.340,89.9％ (169/188) of the training samples and 90.0％ (27/30) of the checking samples were correctly classified.Conclusion Based on factors including HRS,HE,intraoperative blood transfusion,intraoperative volume of urine and duration of operation,the early prognosis can be judged in adult recipients after first liver transplantation. Key words: Liver transplantation; Early prognosis; Stepwise discriminant analysis; Hepatorenal syndrome ; Hepatic encephalopathy

A hospital information system (HIS) in a large hospital has been introduced to store not only the patient records, but also all the logs of laboratory or radiological examinations and clinical actions. Thus, through data mining, these stored data can be used not only for decision support but also for hospital management. In this paper, we applied several exploratory data analysis techniques and generalized linear model to data extracted from HIS, which can be viewed as a first step to data mining for HIS. The results show several interesting results for hospital management, which suggests that the reuse of stored data will give a powerful tool to support a long-period management of a university hospital.

Most of the gene expression data analysis algorithms require the entire gene expression matrix without any missing values. Hence, it is necessary to devise methods which would impute missing data values accurately. There exist a number of imputation algorithms to estimate those missing values. This work starts with a microarray dataset containing multiple missing values. We first apply the modified version of the fuzzy theory based existing method LRFDVImpute to impute multiple missing values of time series gene expression data and then validate the result of imputation by genetic algorithm (GA) based gene ranking methodology along with some regular statistical validation techniques, like RMSE method. Gene ranking, as far as our knowledge, has not been used yet to validate the result of missing value estimation. Firstly, the proposed method has been tested on the very popular Spellman dataset and results show that error margins have been drastically reduced compared to some previous works, which indirectly validates the statistical significance of the proposed method. Then it has been applied on four other 2-class benchmark datasets, like Colorectal Cancer tumours dataset (GDS4382), Breast Cancer dataset (GSE349-350), Prostate Cancer dataset, and DLBCL-FL (Leukaemia) for both missing value estimation and ranking the genes, and the results show that the proposed method can reach 100% classification accuracy with very few dominant genes, which indirectly validates the biological significance of the proposed method.

The objective of this review is to describe the implementation of human factors principles for the design of alerts in clinical information systems. First, we conduct a review of alarm systems to identify human factors principles that are employed in the design and implementation of alerts. Second, we review the medical informatics literature to provide examples of the implementation of human factors principles in current clinical information systems using alerts to provide medication decision support. Last, we suggest actionable recommendations for delivering effective clinical decision support using alerts. A review of studies from the medical informatics literature suggests that many basic human factors principles are not followed, possibly contributing to the lack of acceptance of alerts in clinical information systems. We evaluate the limitations of current alerting philosophies and provide recommendations for improving acceptance of alerts by incorporating human factors principles in their design.