Industry and Occupation in the Electronic Health Record: An Investigation of the National Institute for Occupational Safety and Health Industry and Occupation Computerized Coding System.
Industry and Occupation in the Electronic Health Record: An Investigation of the National Institute for Occupational Safety and Health Industry and Occupation Computerized Coding System.
复制标题
DOI：
10.2196/medinform.4839
复制
发表时间：
2016-02-15
影响因子：
3.2
通讯作者：
Forst L
中科院分区：
文献类型：
--
作者：
Schmitz M;Forst L
在电子健康记录(EHR)中包括有关患者的工作、行业和职业的信息，可以促进职业健康监测、更好的健康结果、预防活动和确定工人赔偿案件。美国国家职业安全与健康研究所(NIOSH)根据1990年人口普查局的代码，开发了一种针对“行业”和“职业”的自动编码系统；其有效性需要结合推动将这些变量强制添加到电子健康记录中进行评估。这项研究的目的是评估NIOSH的工业和职业计算机编码系统(NIOCCS)在应用于根据《平价医疗法案》进行的社区调查中收集的数据时的编码间可靠性；确定使用NIOCCS自动编码的记录的比例。标准职业分类(SOC)代码由几个联邦机构在数据库中使用，这些数据库捕获人口统计、就业和健康信息，以协调这些数据源中与工作活动相关的变量。有359份行业和职业回复是由两名调查人员手工编码的，他们对每一种代码都达成了共识。同样的变量使用NIOCCS在高标准和中等标准水平上进行自动编码。对于SOC代码的前2位，手工编码者之间以及手工编码者共识代码与NIOCCS高置信度代码之间的一致性的Kappa为0.84。对于4位数字，NIOCCS编码与调查者编码的范围从kappa=0.56到0.70。在这项研究中，NIOCCS能够实现输入变量的31%-36%的生产率(即自动编码)，以及49%-58%的“中等置信度”水平。自动编码(生产)率比NIOSH报告的要低一些。手动编码和自动编码数据之间的一致性在2位数级别上是“实质性的”，但在4位数级别上只是“一般”到“良好”。这项工作是实地调查人员执行NIOCCS的基准。进一步的实地测试将阐明NIOCCS在分配代码的能力和编码准确性方面的有效性，并将在推动将这些职业变量纳入EHR时阐明其价值。
Inclusion of information about a patient’s work, industry, and occupation, in the electronic health record (EHR) could facilitate occupational health surveillance, better health outcomes, prevention activities, and identification of workers’ compensation cases. The US National Institute for Occupational Safety and Health (NIOSH) has developed an autocoding system for “industry” and “occupation” based on 1990 Bureau of Census codes; its effectiveness requires evaluation in conjunction with promoting the mandatory addition of these variables to the EHR. The objective of the study was to evaluate the intercoder reliability of NIOSH’s Industry and Occupation Computerized Coding System (NIOCCS) when applied to data collected in a community survey conducted under the Affordable Care Act; to determine the proportion of records that are autocoded using NIOCCS. Standard Occupational Classification (SOC) codes are used by several federal agencies in databases that capture demographic, employment, and health information to harmonize variables related to work activities among these data sources. There are 359 industry and occupation responses that were hand coded by 2 investigators, who came to a consensus on every code. The same variables were autocoded using NIOCCS at the high and moderate criteria level. Kappa was .84 for agreement between hand coders and between the hand coder consensus code versus NIOCCS high confidence level codes for the first 2 digits of the SOC code. For 4 digits, NIOCCS coding versus investigator coding ranged from kappa=.56 to .70. In this study, NIOCCS was able to achieve production rates (ie, to autocode) 31%-36% of entered variables at the “high confidence” level and 49%-58% at the “medium confidence” level. Autocoding (production) rates are somewhat lower than those reported by NIOSH. Agreement between manually coded and autocoded data are “substantial” at the 2-digit level, but only “fair” to “good” at the 4-digit level. This work serves as a baseline for performance of NIOCCS by investigators in the field. Further field testing will clarify NIOCCS effectiveness in terms of ability to assign codes and coding accuracy and will clarify its value as inclusion of these occupational variables in the EHR is promoted.