Capture-Recapture Estimates of Workplace Injury Rates

描述(申请人提供)：这项研究将使用两个最广泛和最全面的非致命性职业伤害和疾病报告来源来评估报告的完整性和影响差异报告的因素。这些来源是国家工人补偿数据和劳工统计局每年进行的职业伤害和疾病调查。它还将改进对8个州的年度职业伤害和疾病总发生率的估计：俄勒冈州、华盛顿州、威斯康星州、新墨西哥州、佛罗里达州、德克萨斯州、明尼苏达州和西弗吉尼亚州。此外，它还将开发和评估匹配来自多个来源的职业伤害和疾病数据的方法。最后，它将应用捕获-重新捕获分析，捕获比率因个人、雇主和受伤特征而异。这将使我们能够确定这些特征的报告率有何不同。在1998-2000年间，这项研究将通过匹配8个州的个人伤害和疾病来比较这两个来源报告的伤害。它将制定方法来匹配这两个来源的数据，并估计错误匹配的记录的比例。然后，它将对这些州的每个州进行捕获-重新捕获分析，以制定改进的非致命伤害发生率的衡量标准。这项研究将使用Logistic回归来解释雇主、伤害和工人特征的差异捕获。这也将使我们能够确定与少报有关的因素。然后，我们将比较这两个数据系统似乎在多大程度上记录了所研究的州的职业伤害和疾病。这项研究将改进每个研究州涉及休假天数的伤害的发病率和漏报情况的估计，并向州工人补偿系统报告这些伤害。对于8个州中的4个州，估计将涵盖所有涉及至少一天休假的伤害。对于所有8个州，这项研究将为至少8天休假的伤害产生可比的估计。如果根据协变量，各州的捕获率是一致的，那么从研究州估计的捕获率将应用于研究人口之外的州，以提供更好的全国伤害和疾病发生率估计。

DESCRIPTION (provided by applicant): This study will use the two most widespread and comprehensive sources of nonfatal occupational injury and illness reporting to evaluate the completeness of reporting and factors that affect differential reporting. These sources are state workers' compensation data and the Survey of Occupational Injuries and Illnesses conducted annually by the Bureau of Labor Statistics. It also will provide improved estimates of the overall annual incidence of occupational injuries and illnesses for 8 states: Oregon, Washington, Wisconsin, New Mexico, Florida, Texas, Minnesota, and West Virginia. In addition, it will develop and evaluate methods for matching occupational injury and illness data from multiple sources. Finally, it will apply capture-recapture analysis with capture rates that vary by individual, employer, and injury characteristics. This will allow us to determine how reporting rates differ by these characteristics. For 1998-2000, the study will compare reported injuries from these two sources by matching individual injuries and illnesses in 8 states. It will develop methods for matching data from the two sources and for estimating the proportion of records falsely matched. It then will use capture-recapture analysis for each of these states to develop improved measures of nonfatal injury incidence. The study will use logistic regression to account for differential capture by employer, injury, and worker characteristics. This also will enable us to identify factors associated with underreporting. Then, we will compare the degree to which the two data systems appear to capture occupational injuries and illnesses in the states studied. The study will provide improved estimates of incidence and underreporting in each study state for injuries that involve days away from work and which are reported to state workers' compensation systems. For 4 of the 8 states, the estimates will cover all injuries involving at least one day off work. For all 8 states, the study will generate comparable estimates for injuries involving at least 8 days off work. If, contingent on covariates, capture is uniform across states, capture rates estimated from the studied states will be applied to states outside the study population to provide improved national estimates of the incidence of injuries and illnesses.