Development of worksystem safety capability index (WSCI) J. Maiti * Department of Industrial Engineering and Management, Indian Institute of Technology Kharagpur, Kharagpur 721 302, West Bengal, India article info Article history: Received 2 October 2009 Received in revised form 18 April 2010 Accepted 14 May 2010 Keywords: Worksystem safety Work injuries Safety capability index Control chart Loss function and process approach abstract This study proposes a new way of analyzing and evaluating worksystem safety based on probability mod- els, control charts, loss functions and safety capability index. The step-by-step procedures for develop- ment of worksystem safety capability index (WSCI) are presented. The development starts with the adoption of process approach to model safety. The key feature of this development is the effective assim- ilation of quality engineering concepts into safety study. A case study was done in an underground coal mining situation to operationalize the development. Time between occurrences (TBO) of injuries and number of injuries (NOI) per month are considered as safety performance variables. The application shows encouraging results for safety improvement through adoption of control charts, expected loss and WSCI. The development opportunity, application potential and future scope of research are discussed. Ó 2010 Elsevier Ltd. All rights reserved. 1. Introduction Measurement of safety performance of any worksystem is of paramount importance for its effective control and improvement. An effective measurement scheme not only provides opportunity to evaluate the safety status of the system, but also enables man- agement to pinpoint the causes of poor safety performance and ef- forts that ensure improvement. Traditionally, worksystem safety performance is measured in terms of accident/injury rate, an out- come measure that is necessarily caused by other measures such as unsafe conditions or unsafe acts which are the measures of causes. Though the second type of measure is a better indicator of safety status, it is not as popular as the first type. The other mea- sure which can be thought of is measurement of efforts (e.g., effect of training). However, due to emergence to risk assessment meth- odology (see Arunraj and Maiti (2007) and Tixier et al. (2002) for comprehensive review; Arendt (1990) for risk assessment process), measurement of safety performance is done in terms of risk which is the probability of occurrence of an undesired event times its con- sequences (Nieuwhof, 1985). Following Tixier et al., 2002, risk assessment methods can be classified into (i) type of methods as qualitative (e.g., risk matrix, Dey et al., 2004), quantitative (e.g., quantitative risk analysis, Krishna et al., 2003), and semi-quantita- tive (e.g., semi-quantitative fault tree, Hauptamanns, 2004) and (ii) nature of analysis as deterministic (e.g., predictive risk index, Chen and Yang, 2004), probabilistic (e.g., petri nets, Nivolianitou et al., 2004) and combined (e.g., safety analysis, Khan and Abbasi, 1998). Risk, as an aggregate measure, takes care of occurrence of haz- ards and its severity in one index. If computed risk is not accept- able based on certain criterion (e.g., benchmark risk value), risk control is practiced to reduce the risk to acceptable level. The most commonly used criterion for benchmarking risk is ALARP, i.e., as low as reasonably possible (Melchers, 2001). The other issue dis- cussed in risk assessment literature is decision making for risk con- trol. Two important decisions that need to be made are selection of appropriate risk assessment techniques and implementing appro- priate risk control measures (Arunraj and Maiti, 2009; Kjellen et al., 2009). Tixier et al.’s review (2002) could help in the first is- sue. For the second issue, risk based maintenance is a relatively new development (Arunraj and Maiti, 2010; Khan and Haddara, 2003, 2004). Risk assessment and control is collectively termed as risk management. Although risk management is a well thought of and established procedure but it has certain limitations. First, it is a time consuming process and requires experts’ knowledge to conduct it. As a result, risk assessment cannot be done everyday. Secondly, risk assessment has to be done in the planning stage to get its full benefit (an offline measure). Day to day deterioration of the worksystem elements due to operations over time (an online measure) is not explicitly considered in risk calculation. As a result, risk assessment is not as effective in low risk systems as compared to high risk systems (e.g., chemical, nuclear, and geotechnical risk in mining). For day to day deterioration of worksystem elements (entropy risk; Mol, 2003), the traditional risk assessment cannot be a pre- ferred approach. Kjellen (1984) proposed deviation concept in occupational accident control where deviation is said to occur when a system characteristic attains a value beyond norms. Alternatively, 0925-7535/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ssci.2010.05.010 * Tel.: +91 3222 283750; fax.: +91 3222 282272/255303. E-mail address: jmaiti09@gmail.com Safety Science 48 (2010) 1369–1379 Contents lists available at ScienceDirect Safety Science journal homepage: www.elsevier.com/locate/ssci