Validity of generic risk factors and the strain index for predicting nontraumatic distal upper extremity morbidity.
Validity of generic risk factors and the strain index for predicting nontraumatic distal upper extremity morbidity.
复制标题
一般危险因素和应变指数用于预测非创伤性远端上肢发病率的有效性。
DOI：
--
复制
发表时间：
2001
期刊：
影响因子：
--
通讯作者：
K. Knox
中科院分区：
文献类型：
--
作者：
J. Moore;N. Rucker;K. Knox
由2名对发病率指标不知情的评估者对56个工作岗位的9个单独的通用危险因素、8个通用危险因素的组合、任何通用危险因素的存在和应变指数进行评估。然后，根据文献中的建议，将工作分配到二分危险分类（问题与安全）。OSHA 200日志用于确定远端上肢（DUE）发病率的历史证据，并将工作分配到二分的发病率分类（阳性与阴性），以无记录病例与一个或多个记录病例作为标准。通过使用2 × 2列联表比较危险和发病率分类，评估关联证据和预测效度措施。5个单独的通用危险因素，3个通用危险因素组合，以及任何通用危险因素的存在与发病率分类无关。与DUE发病率相关的四种单独的通用危险因素和五种通用危险因素组合的比值比估计在3.3-36.0之间。应变指数在所有暴露因子中估计比值比最大，为108.3。根据预测效度模式对暴露方法进行分组。除了一个例外，个别通用危险因素及其组合具有高敏感性低特异性（许多假阳性），低敏感性高特异性（许多假阴性）或低敏感性低特异性。唯一显示出合理预测效度的通用风险因素是手套的使用——其敏感性、特异性、阳性预测值和阴性预测值均为0.75。应变指数的表现优于任何单个或组合的一般风险因素。其敏感性、特异性、阳性预测值、阴性预测值均约为0.90。
Nine individual generic risk factors, eight combinations of generic risk factors, the presence of any generic risk factor, and the Strain Index were evaluated for 56 jobs by 2 evaluators blinded to morbidity measures. Jobs then were assigned to dichotomous hazard classifications (problem versus safe) according to recommendations from the literature. OSHA 200 logs were used to ascertain historical evidence of distal upper extremity (DUE) morbidity, and jobs were assigned to a dichotomous morbidity classification (positive versus negative) using none versus one or more recorded cases as the criterion. Evidence of association and measures of predictive validity were evaluated by comparing hazard and morbidity classifications using 2 x 2 contingency tables. Five individual generic risk factors, three generic risk factor combinations, and the presence of any generic risk factor were not associated with morbidity classification. The odds ratio estimates among the four individual generic risk factors and the five combinations of generic risk factors associated with DUE morbidity varied from 3.3-36.0. The Strain Index had the largest estimated odds ratio of any exposure factor at 108.3. The exposure methods were grouped according to patterns of predictive validity. With one exception, the individual generic risk factors and their combinations had high sensitivity with low specificity (many false-positives), low sensitivity with high specificity (many false-negatives), or low sensitivity with low specificity. The only generic risk factor that demonstrated reasonable predictive validity was the use of gloves--its sensitivity, specificity, positive predictive value, and negative predictive value were equal to 0.75. The Strain Index performed better than any of the individual or combinations of generic risk factors. Its sensitivity, specificity, positive predictive value, and negative predictive value were all approximately 0.90.