基于Meta分析的超前和滞后指标相关性研究

doi:10.16265/j.cnki.issn1003-3033.2026.07.1479

摘要/Abstract

摘要：

为增强事故主动预防水平、优化安全管理资源配置,利用Meta分析法定量研究超前和滞后指标相关性。首先,基于24 Model事故致因理论构建“安全文化-安全管理体系-安全能力”的超前和滞后指标分类框架,归类检索到19项超前指标和3项滞后指标;然后,检索中英文数据库,筛选出25项用于Meta分析的文献,提取相关性系数并进行预处理;最后,使用Meta分析计算超前和滞后指标相关性效应值,同时,通过亚组分析检验相关性结果在不同行业(高危/一般)和文化背景(中国/外国)的表现,并根据结果总结安全启示。结果表明:安全心理(相关性系数r=0.617)和安全参与(r=0.571)与事故频率强相关,检查维护(r=0.54)和安全程序(r=0.514)与事故损失强相关;在高危行业中,安全承诺对事故损失的影响更大(r=0.37);在国外的文化背景下,安全条件对事故频率的影响更为突出(r=0.314)。

关键词: Meta分析, 超前指标, 滞后指标, 安全管理, 24Model

Abstract:

In order to enhance the level of active accident prevention and optimize the allocation of safety management resources, the correlation between leading and lagging indicators was quantitatively studied by using Meta-analysis method. Firstly, based on the 24 Model accident cause theory, the leading and lagging indicators classification framework of "safety culture-safety management system-safety capability" was constructed. And 19 leading indicators and 3 lagging indicators were classified. Secondly, the Chinese and English databases were searched to screen out 25 articles that met the criteria for meta-analysis, and extracted and preprocessed the correlation coefficients. Finally, the correlation effect size of the leading and lagging indicators was calculated using Meta-analysis. The performance of the correlation results in different industries (high-risk/general). And cultural background (Chinese/foreign) was tested by subgroup analysis, and safety implications were summarized based on the results. Results shows: safety psychology(r=0.617) and safety participation(r=0.571) are strongly correlated with accident frequency. And inspection and maintenance(r=0.54) and safety procedures(r=0.514) are strongly correlated with accident loss. In high-risk industries, safety commitment(r=0.37) can more impact on accident loss. In a foreign cultural context, the influence of safety conditions(r=0.314) on the frequency of accidents is more prominent.

Key words: Meta-analysis, leading indicator, lagging indicator, safety management, 24 Model

中图分类号:

X928.03事故预防与预测

姜伟, 张卓野, 马晟翔, 李沐珊, 徐源. 基于Meta分析的超前和滞后指标相关性研究[J]. 中国安全科学学报, 2026, 36(7): 12-19.

Jiang Wei, Zhang Zhuoye, Ma Shengxiang, Li Mushan, Xu Yuan. A study on correlation between safety leading and lagging indicators based on Meta-analysis[J]. China Safety Science Journal, 2026, 36(7): 12-19.

图/表 7

图1

图2

图3

表1

表2

表3

表4

参考文献 21

[1]	Xu Jing, Cheung C, Manu P, et al. Safety leading indicators in construction: a systematic review[J]. Safety Science, 2021, 139: DOI:10.1016/j.ssci.2021.105250.
[2]	Hinze J, Thurman S, Wehle A. Leading indicators of construction safety performance[J]. Safety Science, 2013, 51(1): 23-28. doi: 10.1016/j.ssci.2012.05.016
[3]	14 dead in texas refinery blast[EB/OL]. (2005-03-23). https://www.cbsnews.com/news/14-dead-in-texas-refinery-blast/.
[4]	Toellner J. Improving safety & health performance: identifying & measuring leading indicators[J]. Professional Safety, 2001, 46(9): 42-47.
[5]	Golabchi H, Abellanosa A D, Lefsrud L, et al. A comprehensive systematic review of safety leading indicators in construction[J]. Safety Science, 2024, 172: DOI:10.1016/j.ssci.2024.106433.
[6]	BhagwatK, Delhi V S K, Nanthagopalan P, Construction safety performance measurement using a leading indicator-based jobsite safety inspection method: case study of a building construction project[J]. International Journal of Occupational Safety and Ergonomics, 2022, 28(4): 2645-2656. doi: 10.1080/10803548.2021.2012350
[7]	Dadashi H M, Behnam B, Sebt M H, et al. BIM-based safety leading indicators measurement tool for construction sites[J]. International Journal of Civil Engineering, 2023, 21(2): 265-282. doi: 10.1007/s40999-022-00754-9
[8]	刘艺, 张金锁, 贺琦, 等. 组织支持感对矿工不安全行为影响的链式中介机制[J]. 中国安全科学学报, 2025, 35(7): 22-30. doi: 10.16265/j.cnki.issn1003-3033.2025.07.1766
	Liu Yi, Zhang Jinsuo, He Qi, et al. Chain intermediary mechanism of perceived organizational support on miners' unsafe behaviour[J]. China Safety Science Journal, 2025, 35(7): 22-30.
[9]	Alruqi W M, Hallowell M R. Critical success factors for construction safety: review and meta-analysis of safety leading indicators[J]. Journal of Construction Engineering and Management, 2019, 145(3):DOI:10.1061/(asce)co.1943-7862.0001626.
[10]	Sattari F, Macciotta R, Kurian D, et al. Application of bayesian network and artificial intelligence to reduce accident/incident rates in oil & gas companies[J]. Safety Science, 2021, 133: DOI:10.1016/j.ssci.2020.104981.
[11]	Liu Suxia, Nkrumah E N K, Akoto L S, et al. The state of occupational health and safety management frameworks (OHSMF) and occupational injuries and accidents in the ghanaian oil and gas industry: assessing the mediating role of safety knowledge[J]. BioMed Research International, 2020:DOI:10.1155/2020/6354895.
[12]	Zhang Tianyue, Liu Jiayu, Hong Liang. Application of multidimensional structural equations in the emergency management of coal mine accidents[J]. Processes, 2023, 11(3):DOI:10.3390/pr11030882.
[13]	Ghahramani A, Ebrahimi M, Hajaghazadeh M. Development and psychometric evaluation of an occupational health and safety performance tool for manufacturing companies[J]. Heliyon, 2023, 9(6):DOI:10.1016/j.heliyon.2023.e17343.
[14]	吴超, 王秉. 安全资源学的学科理论研究[J]. 中国安全科学学报, 2024, 34(1): 1-9. doi: 10.16265/j.cnki.issn1003-3033.2024.01.0475
	Wu Chao, Wang Bing. Research on disciplinary theory on saleresourcesology[J]. China Safety Science Joumal, 2024, 34(1):1-9.
[15]	牛茂辉, 李威君, 刘音, 等. 基于24Model的动火作业事故致因文本挖掘[J]. 中国安全科学学报, 2025, 35(3): 151-158. doi: 10.16265/j.cnki.issn1003-3033.2025.03.0757
	Niu Maohui, Li Weijun, Liu Yin, et al. Text mining of causes of hot working accidents based on 24Model[J]. China Safety Science Journal, 2025, 35(3):151-158. doi: 10.16265/j.cnki.issn1003-3033.2025.03.0757
[16]	Gurevitch J, Koricheva J, Nakagawa S, et al. Meta-analysis and the science of research synthesis[J]. Nature, 2018, 555(7695): 175-182. doi: 10.1038/nature25753
[17]	Tong Ruipeng, Yang Xiaoyi, Li Hongwei, et al. Dual process management of coal miners' unsafe behaviour in the Chinese context: evidence from a meta-analysis and inspired by the JD-R model[J]. Resources Policy, 2019, 62: 205-217. doi: 10.1016/j.resourpol.2019.03.019
[18]	Liu Xinxia, Huang Guoxian, Huang Huiqiang, et al. Safety climate, safety behavior, and worker injuries in the Chinese manufacturing industry[J]. Safety Science, 2015, 78: 173-178. doi: 10.1016/j.ssci.2015.04.023
[19]	Fischer R, Milfont T. Standardization in psychological research[J]. International Journal of Psychological Research, 2010, 3(1): 88-96. doi: 10.21500/20112084.852
[20]	Allford L. The auditing of process safety[J]. Journal of Loss Prevention in the Process Industries, 2016, 43: 747-752. doi: 10.1016/j.jlp.2016.07.001
[21]	Zara J, Nordin S M, Isha A S N. Influence of communication determinants on safety commitment in a high-risk workplace: a systematic literature review of four communication dimensions[J]. Frontiers in Public Health, 2023, 11: DOI:10.3389/fpubh.2023.1225995.

指标	检验方法	A₁	A₃	A₄	A₆	B₂	B₃	B₄	B₆	B₇	B₈	B₉	B₁₀	C₁
事故数 D₁	Egger	—	0.078	—	—	—	—	0.530	—	0.460	0.579	0.226	0.701	0.314
	Begg	—	0.497	—	—	—	—	0.497	—	0.497	0.497	0.332	0.602	0.142
	I²/%	—	99.55	—	—	—	—	98.01	—	87.49	76.63	86.08	99.57	99.38
事故损失D₂	Egger	0.377	0.257	0.801	0.718	0.105	0.503	0.063	0.055	0.489	0.084	0.764	0.718	—
	Begg	0.174	0.117	0.601	0.602	0.497	0.117	0.851	0.602	0.501	0.573	0.852	0.602	—
	I²/%	81.23	59.70	98.60	97.77	85.08	57.06	98.81	63.49	78.07	98.64	83.78	78.81	—

滞后指标	超前指标	K	N	r	95%CI(r)	p	I²
事故数 D₁	安全参与A₃	4	3 142	0.571	[0.547,0.594]	^**	99.55
	安全培训B₄	4	10 046	0.319	[-0.069,0.624]	×	98.01
	安全程序B₇	4	9 244	0.42	[0.277,0.544]	^*	87.49
	风险评估B₈	4	9 720	0.432	[0.327,0.525]	^**	76.63
	安全条件B₉	8	4 764	0.279	[0.151,0.398]	^*	86.08
	应急管理B₁₀	3	3 358	0.493	[0.467,0.518]	^**	99.57
	安全心理C₁	4	1 077	0.617	[0.579,0.653]	^**	99.38
事故损失 D₂	安全承诺A₁	4	4 824	0.214	[0.187,0.246]	^*	81.23
	安全参与A₃	3	682	0.07	[-0.197,0.328]	×	69.70
	激励A₄	3	869	0.444	[0.389,0.496]	^**	98.6
	沟通A₆	3	902	0.307	[0.252,0.382]	^*	97.77
	安全会议B₂	4	957	0.387	[0.332,0.44]	^**	85.08
	安全审计B₃	3	700	0.113	[0.039,0.186]	^*	57.06
	安全培训B₄	6	6 665	0.372	[0.113,0.584]	^**	98.81
	检查维护B₆	3	1 481	0.54	[0.503,0.575]	^*	63.49
	安全程序B₇	4	1 011	0.514	[0.216,0.724]	^*	78.07
	风险评估B₈	6	10 797	0.442	[0.047,0.718]	^**	98.64
	安全条件B₉	6	1 891	0.46	[0.254,0.625]	^**	83.78
	应急管理B₁₀	3	1 122	0.46	[0.187,0.668]	^*	78.81

滞后指标	超前指标	剔除文献	原始效应值			剔除后效应值			\|Δr\|
滞后指标	超前指标	剔除文献	r	95%CI	p	r	95%CI	p	\|Δr\|
事故数D₁	应急管理B₁₀	Vinodkumar等	0.493	[0.467, 0.518]	^**	0.226	[0.186, 0.265]	^**	0.267
事故损失D₂	安全程序B₇	Liu Suxia等	0.514	[0.216, 0.724]	^*	0.52	[-0.176, 0.869]	×	0.006
	安全审计B₃	Rajendran	0.133	[0.012, 0.212]	^*	0.161	[-0.734, 0.851]	×	0.048
	检查维护B₆	Tand等	0.54	[0.011, 0.833]	^*	0.585	[-0.646, 0.971]	×	0.045

潜在调节变量		滞后指标	超前指标	K	N	r	95%CI(r)	p值
行业	一般	事故损失D₂	安全培训B₄	2	4 570	0.049	[-0.024,0.12]	×
	高危		安全培训B₄	4	6 665	0.512	[0.476,0.547]	^**
	一般		安全承诺A₁	2	4 570	0.14	[0.092,0.188]	^**
	高危		安全承诺A₁	2	254	0.37	[0.096,0.592]	^**
文化背景	中国	事故损失D₂	安全程序B₇	2	302	0.479	[0.042,0.763]	^*
	外国		安全程序B₇	2	709	0.498	[0.442,0.553]	^**
	中国		风险评估B8	3	1 114	0.464	[0.303,0.599]	^**
	外国		风险评估B8	3	9 683	0.44	[0.065,0.706]	^*
	中国		安全培训B₄	4	5 366	0.396	[0.056,0.653]	^*
	外国		安全培训B₄	2	1 299	0.324	[-0.133,0.667]	×
	中国	事故数D₁	安全条件B₉	2	866	0.335	[-0.059,0.639]	×
	外国		安全条件B₉	4	1 025	0.508	[0.295,0.673]	^**
	中国		安全培训B₄	2	1 148	0.24	[-0.134,0.554]	×
	外国		安全培训B₄	2	8 898	0.395	[0.014,0.676]	^*
	中国		安全条件B₉	3	1 325	0.238	[0.107,0.361]	^**
	外国		安全条件B₉	5	3 439	0.314	[0.19,0.428]	^**