Article type
Abstract
Background:Rehabilitation Nutrition intervention appears to be important in order to increase intervention effect or to prevent iatrogenic sarcopenia We have tried creating a systematic review (SR) for Japanese rehabilitation nutrition guidelines. The problem is to select from a large amount of literature searched which can be a very burdensome for clinicians. Creating a guideline is often done by clinicians who are not familiar with preparing guidelines and SRs. Training is needed.
Objectives:We attempted to resolve these problems by using Web Soft which incorporates artificial intelligence, educating using social network system and reducing the labour.
Methods: We received the results of a comprehensive literature search necessary for artificial intelligence software called Rayyan. In order to reduce the bias and eligibility criteria for the literature, a protocol was created in accordance with the process of the Cochrane Handbook before registration selection by educating in SNS and registered in prospero. The the primary and secondary screening used FaceBook messenger. For the risk of bias evaluation, we used a website with artificial intelligence called Robot Reviewer that evaluates the quality of the paper merely by attaching PDF of RCT to the web.
Results: Rayyan automatically sorted each document by incorporating the results of exhaustive literature search required for SR. Not only essential databases for SR such as MEDLINE, but also corresponded to Japanese such as Ichushi-Web. Systematic reviewers could use the literature selection smoothly with Google translationin the Rayyan website.Selection of 365 literature on cancer rehabilitation was previously unavoidable even if it took more than a month, but this time it was completed in one week. Using the group in the FaceBook, primary and secondary screening was possible on smartphones as well. Robot Reviewer evaluated risk of bias in 10 seconds
Conclusions:Using artificial intelligence software and FaceBook as a platform can greatly reduce the labour and time needed to develop guidelines.
Objectives:We attempted to resolve these problems by using Web Soft which incorporates artificial intelligence, educating using social network system and reducing the labour.
Methods: We received the results of a comprehensive literature search necessary for artificial intelligence software called Rayyan. In order to reduce the bias and eligibility criteria for the literature, a protocol was created in accordance with the process of the Cochrane Handbook before registration selection by educating in SNS and registered in prospero. The the primary and secondary screening used FaceBook messenger. For the risk of bias evaluation, we used a website with artificial intelligence called Robot Reviewer that evaluates the quality of the paper merely by attaching PDF of RCT to the web.
Results: Rayyan automatically sorted each document by incorporating the results of exhaustive literature search required for SR. Not only essential databases for SR such as MEDLINE, but also corresponded to Japanese such as Ichushi-Web. Systematic reviewers could use the literature selection smoothly with Google translationin the Rayyan website.Selection of 365 literature on cancer rehabilitation was previously unavoidable even if it took more than a month, but this time it was completed in one week. Using the group in the FaceBook, primary and secondary screening was possible on smartphones as well. Robot Reviewer evaluated risk of bias in 10 seconds
Conclusions:Using artificial intelligence software and FaceBook as a platform can greatly reduce the labour and time needed to develop guidelines.