Investigation On Response Ability For Public Health Emergency Among Medical Staff In A Grassroots Unit In Henan Province

Objectives1. To investigate the basic cognitive status of public health emergencies among medical staff from grassroots units, evaluate their response capacity, and explore the main influential factors.2. To explore the establishment of data mining forecasting model of response capacity for public health emergencies among staff, and to be compared with the traditional logistic regression model, finding the best prediction model of response ability, providing new ideas to strengthen the emergency management of human resources.Methods1. Scene Investigation:In this study, a multi-stage stratified cluster sampling method sampled randomly in Henan Province Military District, garrison in Xinxiang and Kaifeng, three grassroots army garrison as the primary sampling unit, and then extracting the medical institutions subordinated to grassroots units at all levels of health service as secondary sampling units; Questionnaire survey was conducted on medical staff of each department from the extraction unites.2. The establishment of database and screening variables:Using SPSS17.0software to establish the emergency capability information database of medical staff, using the single factor Logistic regression to filter variable index from emergency capability related factors, selecting the significant variable index to build the model, screening criteria for P<0.05.3. Building models:Using these screening variables to build non-conditional logistic regression prediction model of response capacity by SPSS software, establishing decision tree and neural network predictive models by statistical software Clementine12.0, using receiver operating curve to evaluate the prediction performance of models. Results1. The number of junior college and below degree medical staff of grassroots army medical insurance system was336(55.5%) in Henan, the number of primary and below the title medical personnel was463(76.4%), the educational and title structure were both olive-shaped. There were differences in primary health institutions and the system of hospital on personnel configuration, grassroots medical staffs were young, unmarried males and education level and professional title level were relatively low.2. The public health emergencies awareness among medical staff from grassroots units was74.9%, the passing rate of emergency response capacity was38.4%.3. The mainly influential factors of medical staff response capacity from the grassroots medical institutions combined with neural network, decision tree and the traditional logistic regression model included:the sense of crisis, infectious disease degree of care, training, education, age, job etc.4. The neural network, decision tree and the traditional logistic regression model forecast accuracy rate were all more than75%. Neural network forecasting accuracy rate was87.12%, the highest among the three. The AUC of neural network model was0.869(0.796,0.938), the AUC of decision tree model was0.797(0.711,0.883), the AUC of logistic regression mode! was0.751(0.656.0.845).Conclusions1. The allocation of human resources of dualistic medical security system from grassroots medical institutions in Henan is irrational, there needing to to improve the structure quality among medical staff from two institutions.2. The cognitive level of knowledge related public health emergency events among medical staff is lower than local medical personnel from medical institution, the response capacity for public health emergency of medical personnel needs to be improved; there need to carry out targeted interventions, so as to improve the ability of the medical staff to cope with public health emergency.3. The common important influential factors on the response ability of public health emergency among medical staff through neural network model, decision tree model and logistic regression model analysis include crisis consciousness, infectious disease concern and participate in training.4. The effectiveness of three building prediction models of emergency response capacity among grassroots medical staff is passable; the neural network model prediction effectiveness is the highest. The models can be used to assess the ability to cope with the auxiliary medical personnel, providing scientific basis for strengthening human resources management.