Dissertation > Excellent graduate degree dissertation topics show

The Application of Generalize Additive Models in Medical Cost Control

Author: KeHui
Tutor: MaJun
School: Tianjin Medical University
Course: Epidemiology and Biostatistics,
Keywords: Cataract Hospitalization costs Influencing factors Generalized additive models Multiple Linear Regression
CLC: R197.1
Type: Master's thesis
Year: 2011
Downloads: 90
Quote: 1
Read: Download Dissertation

Abstract

Objective: To understand the 2003-2007 Tianjin enrolled in the basic composition of cataract patients in the hospital costs and compare the different characteristics of patient fees; analysis of hospital costs and the impact of factors using generalized additive models (GAM), and The results of multiple linear regression analysis method compared to explore the feasibility of the application of generalized additive models influencing factors on hospital costs, its future in the medical field method reference for further reform and improve the medical insurance system reference data, and also provide guidance for medical service units in the future work to reduce the burden on patients to improve their own efficiency. Methods: January 1, 2003 January 1 - 2007 Tianjin SilverSea database system in the diagnosis of cataract (ICD-10 coding H25 H26.9) all discharged cases 5030 cases of its inpatient costs Description and patients with different characteristics to be compared. First adopted: SPSS16.0 fitting multiple linear regression model of cataract hospitalization costs affect factors were analyzed, and then using the generalized additive model fitting the S-PLUS6.2 with SAS9.2, multiple linear regression analysis method The results were compared. Results: Cataract ancient hospital costs and the cost of materials the highest proportion, accounting for 44.07% of the total costs, followed by surgery fees, inspection fees, drug charges, more than 90% of the total hospital costs of the four cost. Gray relational analysis found that the cost of materials related degree, followed by surgery. Comparison of cataract patients with or without a second diagnosis, a second diagnosis of patients in both times the total cost of the drug fees, inspection fees, treatment fees and bed charges and no second diagnosis of patients compared the difference was statistically significant (P lt; 0.05), while no significant difference in surgery and material costs. Between the different levels of hospital inpatients, times the average total cost, drug fees, inspection fees, medical fees, medical fees, bed charges, the cost of materials and other costs differences were statistically significant (P lt; 0.05). Pairwise comparisons can be learned, times the average total cost of tertiary hospitals gt; the two hospitals gt; a hospital. The results show that using multiple linear regression model analysis of influencing factors of 5030 cases of cataract patients in hospital costs, six factors affecting the cost of inpatient days (β = 0.726), the tertiary hospitals VS a hospital (β = 0.181), the insured annual 2005VS2003 (β = 0.055), the insured annual 2006VS2003 (β = 0.051), the insured annual 2007VS2003 (β = 0.056), supplies ratio (β = 0.427). The analysis results show that the influencing factors of 5030 cases of cataract patients in hospital costs using generalized additive models cost influential factors in addition to the multi-linear model of six factors, as well as the personnel category, the age of these two factors. And the the smooth figure shows generalized additive model analysis of structure, age, length of stay, supplies proportion of these variables and outcome variables (amount) is not a linear relationship. Difference test further model fitting of the two methods and found that the two models are different, but smaller the GAM fitting model SBC and AIC than the multiple linear regression model can explain our data using the the GAM fitting than the multiple linear regression fit better. Conclusion: The study by generalized additive model analysis of the Tianjin 5030 cases insured influencing factors for cataract patients in hospital costs, not only filter out the factors that affect them, the relevant government authorities to formulate relevant policies of health insurance and medical service units how to strengthen its management in future work to improve economic efficiency and reduce the burden on patients to achieve a win-win situation and provide a valuable reference. In addition, the methodology also provides a reference, multiple linear regression data does not meet the normality and homogeneity of variance due to variable between the outcome variable does not satisfy the linear premise and analyzed using multiple linear model, The results will deviate from actual results, and even lead to unreliable results. Generalized additive models precisely to solve such problems. In this study, data analysis using generalized additive models found differences with the general linear model screening variables influencing factors, not only more than the multiple linear model introduced age variables, and the use of the smoothing parameter map can provide richer information, prompted in can try this method to solve a similar problem in the medical field in the future.

Related Dissertations

CLC: > Medicine, health > Preventive Medicine,Health > Health care organizations and career ( Health Management ) > Health care systems and institutions > Health care system
© 2012 www.DissertationTopic.Net Mobile