Title
Research on PRO scale of acupuncture for cervical spondylosis with multidimensional item response theory
Abstract
Aim of this study is to explore the applicability of patient reported outcome (PRO) scale to the efficacy evaluation of acupuncture for cervical spondylosis. Multidimensional three-parameter logistic model and multidimensional graded response model were established by programming autonomously with SF-36 scale. Markov chain Monte Carlo algorithm and expectation maximization algorithm were adopted to realize the parameter estimation in the models respectively. The applicability of PRO scale such as SF-36 scale to cervical spondylosis was evaluated by the indexes such as item discriminations, difficulties and conjecture degrees from the estimated information. In experimental studies, two multidimensional item response models were established and both have sound values of discriminations, difficulties and conjecture degrees. What's more, a clinical assessment system of acupuncture for cervical spondylosis was established by revising the original PRO scale based on MIRT. In conclusion, patients' ability level on the effect of acupuncture for cervical spondylosis can be reflected well by the revised SF-36 scale. Furthermore, multidimensional item response theory plays an important role in the research of PRO scale of acupuncture for cervical spondylosis.
Year
DOI
Venue
2016
10.1109/ICMLC.2016.7872999
2016 International Conference on Machine Learning and Cybernetics (ICMLC)
Keywords
Field
DocType
Multidimensional Item Response Theory,Cervical Spondylosis,Patient reported outcome(PRO),Scale,Markov chain Monte Carlo algorithm,Expectation maximization algorithm
Response model,Computer science,Expectation–maximization algorithm,Markov chain monte carlo algorithm,Cervical spondylosis,Acupuncture,Artificial intelligence,Physical medicine and rehabilitation,Item response theory,Logistic regression,Machine learning,Patient-reported outcome
Conference
Volume
ISBN
Citations 
2
978-1-5090-0391-4
0
PageRank 
References 
Authors
0.34
0
5
Name
Order
Citations
PageRank
Xiu-Juan Chen100.34
Hong-Lai Zhang222.76
Feng Lu300.34
Hang Wei401.01
Qinqun Chen502.03