Prediction Of Nursing Workload Based On Longitudinal Data

[Objective]To measure the nursing workload of patients in gastroenterology department in different hospitalization days,analyze the influencing factors of nursing workload in different hospitalization days,construct a scientific and objective nursing workload prediction index system,establish a workload prediction model,realize the dynamic prediction of nursing workload in gastroenterology department,and provide data support for guiding clinical prospective fine management and reasonable allocation of nursing manpower.[Method](1)Based on the review of relevant domestic and foreign literature,understand the current research status of influencing factors and existing problems of the prediction of nursing workload in China,analyze the causes of the problems deeply,and determine the theoretical basis and ideas of this study.(2)Using the method of on-the-spot investigation and the "wechat program for nursing information collection and management" developed by the research group,the daily direct nursing hours and influencing factors of 133 inpatients in the Department of Gastroenterology of two top three traditional Chinese medicine hospitals in Jiangsu Province and Yunnan Province were measured longitudinally..(3)Multiple stepwise regression method was used to analyze the influencing factors of direct nursing hours in patients with different hospitalization days,and the predictive indicators of nursing workload were screened.Principal component analysis was used to further simplify the data structure,and a multi-dimensional nursing workload prediction index system was constructed.(4)Based on the longitudinal data of nursing workload in gastroenterology department(direct nursing hours and prediction index system of patients during hospitalization),the workload prediction model was explored based on workload time series,which was influenced by time,season and other external factors;the sensitive prediction indexes of nursing workload were screened by correlation analysis method,and entered into the multiple regression prediction model to construct the workload prediction model based on internal factors.[Result](1)Distribution of the nursing workload of the gastroenterology department in different hospitalization days:This research included 133 participants in the gastroenterology department of two tertiary hospitals of TCM in the course of their hospitalization for a total of 1139 days and the average days(8.56±2.53)of direct nursing hours.The results revealed that with the extension of hospitalization time,the overall tendency of direct nursing hours presented an inverted-U-shaped relationship,and there were statistical differences indirect nursing hours on different hospitalization days(P<0.05).The average daily direct nursing hours for 8-10 days after admission were the longest at 61.54 min,while the average daily direct nursing hours for discharge were the shortest at 31.22 min.(2)Influencing factors of the nursing workload of the gastroenterology department in different hospitalization days:Multivariate stepwise regression analysis showed that self-care ability had a negative predictive effect on the nursing workload on the admission day,and explaining 40.3%of it;the difference between 55.7%and 78.4%was explained by factors such as hospitalization for 2-10 days,the nursing unit,self-care ability,the severity of illnesses,occupations,incomes,prior history,chronic diseases,etc;while self-care ability,the nursing unit and income on discharge day can only explain 26.3%of the difference.(3)The construction of nursing workload prediction index system in digestive medicine:The method of combining qualitative and quantitative to construct the nursing workload prediction index system,finally,it determines six primary indicators including body disease,social demography factor,organizational environment factor,disease type factor,nutrition factor and time factor,and 14 secondary indicators in total.(4)Prediction model of nursing workload in the gastroenterology department:The stationary R2 value of the time series model prediction result is 0.654,which has good accuracy and fit,and can predict the short-term trend.Based on multiple linear regression prediction,the model’s goodness of fit is 0.615,and average daily direct care hours per patient(seconds)=9490.183+1.958*age-83.271*body mass index-35.764*self-care ability+59.146*severity of disease+294.030*type of disease+64.144*hospitalization days.[Conclusion](1)This research investigated nursing workload data longitudinally and captures the trend of the nursing workload with changes in influencing factors dynamically.It has great value for evaluating the nursing workflow of different hospitalization days and improving the accuracy and guidance of prediction.(2)Workload prediction indicators are the basis for a comprehensive and accurate measurement of workload.The index system construction method combines subjective and objective,covering comprehensive content,strong operability,and has greater practical guiding significance.It is for clinical nursing workload measurement and later modeling.Forecast research provides a certain reference and reference.(3)In this paper,time series model and multiple linear regression model were used to dynamically predict nursing workload.The time series model considers time,season and other external factors,and can be used to predict nursing workload periodically,such as the nursing workload of a department in a week.Multiple linear regression model was used to get the calculation formula of nursing workload prediction for different types of patients.The indicators of the formula cover the key content of workload,which can achieve accurate prediction of daily nursing workload,and has strong operability and clinical practical value.