Construction of Influenza Early Warning Model Based on Combinatorial Judgment Classifier: A Case Study of Seasonal Influenza in Hong Kong.

OBJECTIVE: The annual influenza epidemic is a heavy burden on the health care system, and has increasingly become a major public health problem in some areas, such as Hong Kong (China). Therefore, based on a variety of machine learning methods, and considering the seasonal influenza in Hong Kong, the study aims to establish a Combinatorial Judgment Classifier (CJC) model to classify the epidemic trend and improve the accuracy of influenza epidemic early warning. METHODS: The characteristic variables were selected using the single-factor statistical method to establish the influencing factor system of an influenza outbreak. On this basis, the CJC model was proposed to provide an early warning for an influenza outbreak. The characteristic variables in the final model included atmospheric pressure, absolute maximum temperature, mean temperature, absolute minimum temperature, mean dew point temperature, the number of positive detections of seasonal influenza viruses, the positive percentage among all respiratory specimens, and the admission rates in public hospitals with a principal diagnosis of influenza. RESULTS: The accuracy of the CJC model for the influenza outbreak trend reached 96.47%, the sensitivity and specificity change rates of this model were lower than those of other models. Hence, the CJC model has a more stable prediction performance. In the present study, the epidemic situation and meteorological data of Hong Kong in recent years were used as the research objects for the construction of the model index system, and a lag correlation was found between the influencing factors and influenza outbreak. However, some potential risk factors, such as geographical nature and human factors, were not incorporated, which ideally affected the prediction performance to some extent. CONCLUSION: In general, the CJC model exhibits a statistically better performance, when compared to some classical early warning algorithms, such as Support Vector Machine, Discriminant Analysis, and Ensemble Classfiers, which improves the performance of the early warning of seasonal influenza.

Patients' preferences for health insurance coverage of new technologies for treating chronic diseases in China: a discrete choice experiment.

OBJECTIVES: Our study aimed to inform insurance decision-making in China by investigating patients' preferences for insurance coverage of new technologies for treating chronic diseases. DESIGN: We identified six attributes of new medical technologies for treating chronic diseases and used Bayesian-efficient design to generate choice sets for a discrete choice experiment (DCE). After conducting the DCE, we analysed the data by mixed logit regression to examine patient-reported preferences for each attribute. SETTING: The DCE was conducted with patients in six tertiary hospitals from four cities in Jiangsu province. PARTICIPANTS: Patients aged 18 years or older with a history of diabetes or hypertension and taking medications regularly for more than 1 year were recruited (n=408). RESULTS: The technology attributes regarding expected gains in health outcomes from the treatment, high likelihood of effective treatment and low incidence of serious adverse events were significant, positive predictors of choice by the study patients (p<0.01). The out-of-pocket cost was a significant, negative attribute for the entire study sample (ß = -0.258, p<0.01) and for the patients with Urban-Rural Residents Basic Medical Insurance (URRBMI) (ß = -0.511, p<0.01), but not for all the patients with Urban Employees Basic Medical Insurance (UEBMI) (ß = -0.071, p>0.05). The severity of target disease was valued by patients with lower EQ-5D-5L index value as well as URRBMI enrollees. CONCLUSIONS: Patients highly valued the health benefits and risks of new technologies, which were closely linked to their feelings of disease and perceptions of health-related quality of life. However, there existed heterogeneity in preferences between URRBMI and UEBMI patients. Further efforts should be made to reduce the gap between insurance schemes and make safe and cost-effective new technologies as a priority for health insurance reimbursement.

Clinic expert information extraction based on domain model and block importance model.

To extract expert clinic information from the Deep Web, there are two challenges to face. The first one is to make a judgment on forms. A novel method based on a domain model, which is a tree structure constructed by the attributes of query interfaces is proposed. With this model, query interfaces can be classified to a domain and filled in with domain keywords. Another challenge is to extract information from response Web pages indexed by query interfaces. To filter the noisy information on a Web page, a block importance model is proposed, both content and spatial features are taken into account in this model. The experimental results indicate that the domain model yields a precision 4.89% higher than that of the rule-based method, whereas the block importance model yields an F1 measure 10.5% higher than that of the XPath method.

[Method of detection of negative terminology in Chinese electronic medical record].

The method for detecting the negative terms in Chinese electronic medical record (EMR) is useful in providing evidence for constructing concept index. In this respect, we adopted an improved method which combined maximum matching with mutual information in order to extract terms in EMRs. This method can overcome the influence of overlay ambiguity. In addition, for the determination of negative semantic, we also adopted an improved method which combined rule-based method with word co-occurrence. This new method can reduce the probability of appearance of false positive terms caused by punctuation input errors. The result showed that the negative predictive value is 7.85% higher than the rule-based method.

[Study on Information Extraction of Clinic Expert Information from Hospital Portals].

Clinic expert information provides important references for residents in need of hospital care. Usually, such information is hidden in the deep web and cannot be directly indexed by search engines. To extract clinic expert information from the deep web, the first challenge is to make a judgment on forms. This paper proposes a novel method based on a domain model, which is a tree structure constructed by the attributes of search interfaces. With this model, search interfaces can be classified to a domain and filled in with domain keywords. Another challenge is to extract information from the returned web pages indexed by search interfaces. To filter the noise information on a web page, a block importance model is proposed. The experiment results indicated that the domain model yielded a precision 10.83% higher than that of the rule-based method, whereas the block importance model yielded an F1 measure 10.5% higher than that of the XPath method.