The Study Of Applicaion Of PRO For Effictiveness Assessment Of Acupuncture For Cervical Spondylosis Based On Artificial Neural Network

ObjectiveThe study makes an attemtpt to test the applicability of Patient-reported outcomes (PRO) for the Cervial Spondylosis through the reliability and validity, as well as responsiveness analysis of SF-36. And then to use an assessment model based on artificial neural network (ANN) to assessment the effictiveness of the acupuncture treatment for the Cervial Spondylosis (CS), and try to build up a clinical outcome assessment model.MethodsThe research, it is a randomized controlled experiment. The patients of Cervial Spondylosis are divided into acupuncture group and placebo acupuncture group.. After treatments, we apply SF-36health questionnaire and visual analogue scale VAS as an index of patient-reported outcomes. At the same time the outcomes of SF-36are regarded as the main indexes, And as well as the outcomes of the VAS are regarded as the rference indexes. The collection of PRO data in clinial trials is used to analyze the reliability and validity and responsiveness of the health questionnaire with SPSS19.0. After that we adhibit an algorithm based on ANN. It is made up of three layers and it includs one called feedforward neural network and another called back-propagation algorithm. The algorithm is used of evaluating the effictiveness of acupunture treatment of cervical spondylosis, on account of the outcomes of SF-36. Therefore, we can further explore that whether the ANN it is suitable to appraise the effictiveness of acupuncture or not, for the sake of set up a clincal outcome assessment model.ResultsThe result of the reliability analysis shows that the Cronbach’s a for the overall rliabilify of SF-36is0.825with the standardizd items of0.872. The valif ity analysis reveals there are eight common factors in the SF-36qustionnaire, in addition their rate of cumulative make a contribution to76.31%. We can make a conclusion from the outcomes of response analysis, SF-36score in the termination of remedy, follow-up of one month and3months after the remedy are different from those before the remedy except the mental healdh (P<0.05). The outcomes of the reliability, valifity and responsiveness analysis tells that PRO has a good applicability to judge the effictiveness of the acupuncture to the curing of cervical spondylosis. In the further rsearch, we adhibit an algorithm which consists of three layer with back-propagation to extracte the features from the SF-36questionnaire. The precision of the ANN based model reaches96.75%, which shows good learning capacity in appraising the effictiveness of the acupuncture therapy.ConelusionThe outcomes of the research show that the PRO assessment is suitable to judge the effictiveness of acupuncture for healing the patients with CS in a complex context. We bring the algorithm which is called ANN-BP, after the confirmation of this test we find out it is feasible means for judging the effictiveness of the acupuncture therapy.