ABSTRACT
Objective To create risk predictive models of healthcare-seeking delay among imported malaria patients in Jiangsu Province based on machine learning algorithms, so as to provide insights into early identification of imported malaria cases in Jiangsu Province. Methods Case investigation, first symptoms and time of initial diagnosis of imported malaria patients in Jiangsu Province in 2019 were captured from Infectious Disease Report Information Management System and Parasitic Disease Prevention and Control Information Management System of Chinese Center for Disease Control and Prevention. The risk predictive models of healthcare-seeking delay among imported malaria patients were created with the back propagation (BP) neural network model, logistic regression model, random forest model and Bayesian model using thirteen factors as independent variables, including occupation, species of malaria parasite, main clinical manifestations, presence of complications, severity of disease, age, duration of residing abroad, frequency of malaria parasite infections abroad, incubation period, level of institution at initial diagnosis, country of origin, number of individuals travelling with patients and way to go abroad, and time of healthcare-seeking delay as a dependent variable. Logistic regression model was visualized using a nomogram, and the nomogram was evaluated using calibration curves. In addition, the efficiency of the four models for prediction of risk of healthcare-seeking delay among imported malaria patients was evaluated using the area under curve (AUC) of receiver operating characteristic curve (ROC). The importance of each characteristic was quantified and attributed by using SHAP to examine the positive and negative effects of the value of each characteristic on the predictive efficiency. Results A total of 244 imported malaria patients were enrolled, including 100 cases (40.98%) with the duration from onset of first symptoms to time of initial diagnosis that exceeded 24 hours. Logistic regression analysis identified a history of malaria parasite infection [odds ratio (OR) = 3.075, 95% confidential interval (CI): (1.597, 5.923)], long incubation period [OR = 1.010, 95% CI: (1.001, 1.018)] and seeking healthcare in provincial or municipal medical facilities [OR = 12.550, 95% CI: (1.158, 135.963)] as risk factors for delay in seeking healthcare among imported malaria cases. BP neural network modeling showed that duration of residing abroad, incubation period and age posed great impacts on delay in healthcare-seek among imported malaria patients. Random forest modeling showed that the top five factors with the greatest impact on healthcare-seeking delay included main clinical manifestations, the way to go abroad, incubation period, duration of residing abroad and age among imported malaria patients, and Bayesian modeling revealed that the top five factors affecting healthcare-seeking delay among imported malaria patients included level of institutions at initial diagnosis, age, country of origin, history of malaria parasite infection and individuals travelling with imported malaria patients. ROC curve analysis showed higher overall performance of the BP neural network model and the logistic regression model for prediction of the risk of healthcare-seeking delay among imported malaria patients (Z = 2.700 to 4.641, all P values < 0.01), with no statistically significant difference in the AUC among four models (Z = 1.209, P > 0.05). The sensitivity (71.00%) and Youden index (43.92%) of the logistic regression model was higher than those of the BP neural network (63.00% and 36.61%, respectively), and the specificity of the BP neural network model (73.61%) was higher than that of the logistic regression model (72.92%). Conclusions Imported malaria cases with long duration of residing abroad, a history of malaria parasite infection, long incubation period, advanced age and seeking healthcare in provincial or municipal medical institutions have a high likelihood of delay in healthcare-seeking in Jiangsu Province. The models created based on the logistic regression and BP neural network show a high efficiency for prediction of the risk of healthcare-seeking among imported malaria patients in Jiangsu Province, which may provide insights into health management of imported malaria patients.
ABSTRACT
Objective@#To construct and confirm a predictive model for the risks of cardiovascular diseases (CVD) with metabolic syndrome (MS) and its factors in Xinjiang Kazakh population.@*Methods@#A total of 2 286 Kazakh individuals were followed for 5 years from 2010 to 2012 as baseline survey. They were recruited in Xinyuan county, Yili city, Xinjiang. CVD cases were identified via medical records of the local hospitals in 2013, 2016 and 2017, respectively. Factor analysis was performed on 706 MS patients at baseline, and main factors, age, and sex were extracted from 18 medical examination indexs to construct a predictive model of CVD risk. After excluding the subjects with CVD at baseline and incomplete data, 2007 were used as internal validation, and 219 Kazakhs in Halabra Township were used as external validation. Logistic regression discriminations were used for internal validation and external validation, as well as to calculate the probability of CVD for each participant and receiver operating characteristic curves.@*Results@#The prevalence of MS in Kazakh was 30.88%. Seven main factors were extracted from the Kazakh MS population, namely obesity factor, blood lipid and blood glucose factor, liver function factor, blood lipid factor, renal metabolic factor, blood pressure factor, and liver enzyme factor. The area under the curve (AUC) for predicting CVD in the internal validation was 0.773 (95%CI 0.754-0.792). In the external validation, the AUC for predicting CVD was 0.858 (95%CI 0.805-0.901).@*Conclusions@#The CVD risk prediction model constructed by 7 main factors extracted from Kazakh MS patients has high validation efficiency and can be used for risk assessment of CVD in Xinjiang Kazakh population.
ABSTRACT
Objective This study aimed to provide an epidemiological modeling method to evaluate the risk of metabolic syndrome (MS) development in the coming 5 years among 35-74 year-olds from Taiwan.Methods A cohort of 13 973 subjects aged 35-74 years who did not have metabolic syndrome but took the initial testing during 1997-2006 was formed to derive a risk score which tended to predict the incidence of MS.Multivariate logistic regression was used to derive the risk functions and using the ‘check-up center' (Taipei training cohort) as the overall cohort.Rules based on these risk functions were evaluated in the remaining three centers (as testing cohort).Risk functions were produced to detect the MS on a training sample using the multivariate logistic regression models.Started with those variables that could predict the MS through univariate models,we then constructed multivariable logistic regression models in a stepwise manner which eventually could include all the variables.The predictability of the model was evaluated by areas under curve (AUC) the receiver-operating characteristic (ROC) followed by the testification of its diagnostic property on the testing sample.Once the final model was defined,the next step was to establish rules to characterize 4 different degrees of risks based on the cut points of these probabilities,after being transformed into normal distribution by log-transformation.Results At baseline,the range of the proportion of MS was 23.9％ and the incidence of MS in 5-years was 11.7％ in the non-MS cohort.The final multivariable logistic regression model would include ten risk factors as:age,history of diabetes,contractive pressure,fasting blood-glucose,triglyceride,high density lipoprotein cholesterol,low density lipoprotein cholesterol,body mass index and blood uric acid.AUC was 0.827 (95％ CI:0.814-0.839) that could predict the development of MS within the next 5 years.The curve also showed adequate performance in the three tested samples,with the AUC and 95％ CI as 0.813 (0.789-0.837),0.826(0.800-0.852) and 0.794(0.768-0.820),respectively.After labeling the degrees of the four risks,it was showed that over 17.6％ of the incidence probability was in the population under mediate risk while over 59.0％ of them was in the high risk group,respectively.Conclusion Both predictability and reliability of our Metabolic Syndrome Risk Score Model,derived based on Taiwan MJ Longitudinal Health-checkup-based Population Database,were relatively satisfactory in the testing cohort.This model was simple,with practicable predictive variables and feasible form on degrees of risk.This model not only could help individuals to assess the situation of their own risk on MS but could also provide guidance on the group surveillance programs in the community regarding the development of MS.
ABSTRACT
Objective This study aimed to provide an epidemiological modeling in evaluating the risk of developing obesity within 5 years in Taiwan population aged 30-59 years.Methods After excluding 918 individuals who were obesitive at baseline,a cohort of 14 167 non-obesity subjects aged 30-59 years in the initial year during 1998-2006,was formed to derive a Risk Score which could predict the incident obesity (IO).Multivariate logistic regression was used to derive the risk functions,using the check-up center (Taipei training cohort,n=8104) of the overall cohort.Rules based on these risk functions were evaluated in the left three centers (testing cohort,n=6063).Risk functions were produced to detect the IO on a training sample using the multivariate logistic regression models.Starting with variables that could predict the IO through univariate models,we constructed multivariable logistic regression models in a stepwise manner which eventually could include all the variables.We evaluated the predictability of the model by the area under the receiver-operating characteristic (ROC) curve (AUC) and to testify its diagnostic property on the testing sample.Once the final model was defined,the next step was to establish rules to characterize 4different degrees of risk based on the cut points of these probabilities after transforming into normal distribution by log-transformation.Results At baseline,the range of the proportion of normal weight,overweight and obesity were 50.00％ 60.00％,26.47％-31.11％ and 5.76％-7.24％ respectively in tour check-up centers of Taiwan.After excluding 918 obesity individuals at baseline,we ascertained 386 (2.73％,386/14 167) cases having IO and 2.66％-2.91％ of them having centered obesity in the four check-up centers respectivcly.Final multivariable logistic regression model would include five risk lactors:sex,age,history of diabetes,weight deduction ≥4 kg within 3 months and waist circumference.The area under the ROC curve (AUC) was 0.898 (95％CI,0.884-0.912) that could predict the development of obesity within 5 years.The curve also had adequate performance in testing the sample [AUC=0.881 (95％CI,0.862 0.900) ].After labeling the four risk degrees,16.0％ and 2.9％ of the total subjects were in the mediate and high risk populations respectively and were 7.8 and 16.6 times higher,when comparing with the population at risk in general.Conclusion The predictability and reliability of our obesity risk score model,derived based on Taiwan MJ Longitudinal Health-checkup-based Population Database,were relatively satisfactory,with its simple and practicable predictive variables and the risk degree form.This model could help individuals to self assess the situation of risk on obesity and could also guide the community caretakers to monitor the trend of obesity development.