脊髓损伤神经学分类国际标准（ISNCSCI）计算机算法的开发与验证
Development and validation of a computerized algorithm for International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI).
作者信息
Walden K, Bélanger L M, Biering-Sørensen F, Burns S P, Echeverria E, Kirshblum S, Marino R J, Noonan V K, Park S E, Reeves R K, Waring W, Dvorak M F
机构信息
Rick Hansen Institute, Vancouver, British Columbia, Canada.
Vancouver Spine Program, Vancouver General Hospital, Vancouver, British Columbia, Canada.
出版信息
Spinal Cord. 2016 Mar;54(3):197-203. doi: 10.1038/sc.2015.137. Epub 2015 Sep 1.
STUDY DESIGN
Validation study.
OBJECTIVES
To describe the development and validation of a computerized application of the international standards for neurological classification of spinal cord injury (ISNCSCI).
SETTING
Data from acute and rehabilitation care.
METHODS
The Rick Hansen Institute-ISNCSCI Algorithm (RHI-ISNCSCI Algorithm) was developed based on the 2011 version of the ISNCSCI and the 2013 version of the worksheet. International experts developed the design and logic with a focus on usability and features to standardize the correct classification of challenging cases. A five-phased process was used to develop and validate the algorithm. Discrepancies between the clinician-derived and algorithm-calculated results were reconciled.
RESULTS
Phase one of the validation used 48 cases to develop the logic. Phase three used these and 15 additional cases for further logic development to classify cases with 'Not testable' values. For logic testing in phases two and four, 351 and 1998 cases from the Rick Hansen SCI Registry (RHSCIR), respectively, were used. Of 23 and 286 discrepant cases identified in phases two and four, 2 and 6 cases resulted in changes to the algorithm. Cross-validation of the algorithm in phase five using 108 new RHSCIR cases did not identify the need for any further changes, as all discrepancies were due to clinician errors. The web-based application and the algorithm code are freely available at www.isncscialgorithm.com.
CONCLUSION
The RHI-ISNCSCI Algorithm provides a standardized method to accurately derive the level and severity of SCI from the raw data of the ISNCSCI examination. The web interface assists in maximizing usability while minimizing the impact of human error in classifying SCI.
SPONSORSHIP
This study is sponsored by the Rick Hansen Institute and supported by funding from Health Canada and Western Economic Diversification Canada.
研究设计
验证性研究。
目的
描述脊髓损伤神经学分类国际标准（ISNCSCI）计算机应用程序的开发与验证。
背景
急性和康复护理数据。
方法
基于2011版ISNCSCI和2013版工作表开发了里克·汉森研究所 - ISNCSCI算法（RHI - ISNCSCI算法）。国际专家制定了设计和逻辑，重点关注可用性和功能，以规范对具有挑战性病例的正确分类。采用五阶段流程来开发和验证该算法。对临床医生得出的结果与算法计算结果之间的差异进行了核对。
结果
验证的第一阶段使用48个病例来开发逻辑。第三阶段使用这些病例以及另外15个病例进行进一步的逻辑开发，以对具有“不可测试”值的病例进行分类。在第二和第四阶段进行逻辑测试时，分别使用了来自里克·汉森脊髓损伤登记处（RHSCIR）的351个和并1998个病例。在第二和第四阶段识别出的23个和286个差异病例中，有2个和6个病例导致了算法的更改。在第五阶段使用108个新的RHSCIR病例对该算法进行交叉验证时，未发现需要进一步更改的情况，因为所有差异均是由于临床医生的错误所致。基于网络的应用程序和算法代码可在www.isncscialgorithm.com上免费获取。
结论
RHI - ISNCSCI算法提供了一种标准化方法，可从ISNCSCI检查的原始数据中准确得出脊髓损伤的水平和严重程度。网络界面有助于最大限度地提高可用性，同时在脊髓损伤分类中最小化人为错误的影响。
资助
本研究由里克·汉森研究所赞助，并得到加拿大卫生部和加拿大西部经济多样化部的资金支持。
Zotero 插件