| Objective: To explore the method of filling the missing data of the scale in the stroke cohort study,to make the data approximate the real situation to the greatest extent,and to improve the data quality;to evaluate the rehabilitation effect of the comprehensive Chinese medicine program on the various functions of stroke patients to improve the health of patients.Methods: Randomly select the data of 400 patients with sequelae of stroke in a multi-center,prospective cohort study conducted in 5 tertiary A hospitals in Shaanxi Province from January 2016 to December 2020,and use R language 4.0.1 to analyze the amount of NIHSS Table,ADL scale,FMAS scale,HAMD scale and HAMA scale data missing features,namely missing proportion,missing pattern,missing mechanism;select the optimal multiple filling method MCMC,use SAS 9.4 to fit the missing data;pass the standard Compare the data results of multiple imputation method and deletion method with 95% CI to evaluate the fitting effect of multiple imputation method;use paired t test to analyze the NIHSS,ADL,FMAS,HAMD and HAMA scales before and after treatment in the two cohorts The two-sample independent t-test was used to analyze the difference in scale scores between the two cohorts to evaluate the rehabilitation effect of the comprehensive Chinese medicine program on the different functions of patients with stroke sequelae.Result:(1)Exploring the method of filling the missing data of stroke-related scales The results of the study showed that the percentage of missing data increased over time,and the missing percentage of the control cohort was significantly higher than that of the observation cohort.The NIHSS scale had the lowest percentage of missing data(0%),and the HAMD scale The highest percentage of missing data(13.5%);NIHSS,ADL,FMAS,HAMD,and HAMA scale data are all missing patterns,in which NIHSS scale is MCAR mechanism,and other scales are MAR mechanism;based on the above data missing characteristics,it is judged MCMC method is the optimal multiple filling method for stroke cohort scale data.The MCMC method produces a smaller standard error and a narrower 95% confidence interval width,and can more effectively use other information to predict missing data and improve data utilization.(2)The curative effect of traditional Chinese medicine intervention in stroke disease showed that: the comprehensive intervention of traditional Chinese medicine and western medicine before and after intervention,NIHSS scale,ADL scale,HAMD scale and HAMA scale scores all showed statistical differences(P<0.05),The scores of sensory function and balance function in the FMAS scale,but the scores of each scale did not show statistical differences between the two groups;the observation cohort was better than the control cohort in improving motor function and joint mobility;in the FMAS scale,the joints The improvement effect of the two groups of cohorts on the pain subscale score was not obvious.Conclusion:(1)Applying multiple filling methods to solve the problem of missing scale data can effectively improve the data utilization of stroke cohort studies,make the statistical results approximate the real measurement situation to the greatest extent,enhance the integrity of the research data,and improve the data quality;(2)Verify the integration of traditional Chinese medicine The program has the advantages of improving patients’ motor function and joint mobility in the treatment of stroke sequelae,and finally formed a comprehensive TCM treatment program for stroke sequelae with precise curative effect and outstanding advantages.Compared with the basic treatment plan of western medicine,the comprehensive treatment plan of traditional Chinese medicine has not yet shown more obvious effects in improving patients’ neurological function,ability of daily living and psychological function.This may be related to the small amount of data included in the analysis and short observation time in this paper. |
PDF Full Text Request