Burgess COMMENTARY

Peter Burgess

Safety data collection process as important as what’s collected Take the system beyond paper to process hazard analysis

We cannot improve what we do not measure. That does not, however, mean that Figure 1. Elements of a performance management system. Metrics need to be appropriate for the organization’s maturity level and rich in signal, collected through a robust data collection process, and monitored and reviewed regularly. Appropriate actions need to result from the review process, be it recognition for positive results or corrective actions. Either way, these actions need to be tracked to completion. Courtesy: DuPont Sustainable Solutionswe can improve just because we measure. To be useful, what we choose to measure needs to point us to where improvements need to be made and what actions need to be taken. These measurements should help assure us of the safe operations of our units when they are safe, and alert us to take action when it is necessary to do so. Not having the right measures in place may give an organization a false sense of safety and, even worse, drive complacency.

Many organizations use common measures such as total recordable injuries, injury frequency rates, etc., to satisfy themselves that they are managing safety satisfactorily. Unfortunately, for operations with equipment-based and process-based hazards, these measures are definitely insufficient. In fact, several catastrophic incidents have resulted from reliance on metrics that are unsuitable as process safety performance indicators. In response to this, there has been a rise in the number of references and discussions on the topic of appropriate metrics for Performance Indicator.

Despite all these available resources, however, some organizations still feel inadequate with their process safety performance monitoring. Could it be that this feeling of inadequacy has more to do with how metrics are collected and used? Is the data collection process as important as metric selection in ensuring the effectiveness of performance management?

Common challenges

An organization looking to implement a performance management program is often faced with several common challenges. Many of these challenges stem from the failure in executing the necessary elements of a performance management program (see Figure 1), such as the improper selection of metrics, having a nonrobust data collection process, having an ineffective or non-existent data monitoring and review process, and, most importantly, the failure to initiate actions based on the data review process and to manage them to timely completion. IFS

Specific to data collection, there are other functional and cultural issues that make it especially difficult to implement. Table 1 summarizes the commonly observed challenges in establishing a data collection process.

Ultimately, the organization needs to be confident in the data that it is monitoring and using for the basis of its actions and rewards. Not addressing these common issues will result in an ineffective program, giving the organization a false sense of safety or, conversely, an unnecessary sense of paranoia. The personnel tasked to collect the data may also feel frustrated and confused, and may begin questioning the value of the process. Courtesy: DuPont Sustainable Solutions

Data management systems

Many organizations are now turning to IT tools to help them with data collection, consolidation, and presentation for their performance management program. This is primarily driven by resource constraints. Some organizations build their own data collection templates and dashboards using Microsoft Excel or a similar program, and some use commercial software. Research by the Aberdeen Group highlighted that best-in-class companies tend to invest in an integrated safety system that connects with plant automation data directly. Such a system allows for real-time performance monitoring and provides visibility to plant operation for diagnostic purposes. Having the right non-paper-based IT data management system in place can certainly help with the robustness of a performance management program.

For example, a data management system that works by managing process safety workflows can make performance management more effective and robust. The system works by collecting information as events occur, such as when the need to initiate a change arises, an emergency drill or process hazards analysis (PHA) is conducted, or when a personnel transfer occurs. A user will input the appropriate record into the system indicating that an event has occurred, or is about to occur.

The system will then guide the organization through the workflow as defined and configured into the system. It will trigger appropriate notifications, requests for necessary approvals, or requests for other additional information as necessary, until all requirements to close the record are completed. It essentially functions to electronically guide the organization through its process safety standard.

Figure 2. DuPont PSM Model. Courtesy: DuPont Sustainable Solutions

The idea of collecting data as it is available or required in the workflow is a change from how a performance management program is traditionally done. With this, data collection is now part of the management of the specific PSM element. No longer do personnel feel that they are collecting data for the sake of reporting. The use of the software will provide them with clarity and understanding of the relevance of the data being collected in the overall workflow. Users will have an appreciation that the data being collected is consistent with the requirements of the organization's PSM standard. As a bonus, the system will also force certain operational discipline. Overdue items are highlighted immediately and reminders automatically triggered. Shortcuts or incomplete information on certain forms can be configured so that they are not possible when using the software.

Collecting data as part of workflow management using an IT system will also allow the organization easier access to more information. Information that is not typically reported for performance management purposes is now available electronically. For example, the leadership team can now easily find out if the number of near misses is disproportionally higher with overtime workers, in which case a review of the organization's fatigue management may be warranted. This allows for a more effective diagnosis of trends, facilitated by data queries or built-in search capabilities.

Having data available on demand for management review and diagnosis is powerful. Data no longer needs to be collected specifically for the purpose of performance management review at some regular frequency. Instead, data can now be produced anytime. The system will reflect all changes in real time in the prebuilt dashboards. The idea of being able to produce and report data as needed versus having to collect data reflects a shift from how performance management is traditionally implemented. With this, functional and cultural issues highlighted above can be alleviated, if not eliminated altogether.

Effective models

So, what kind of data management system is most effective? DuPont represents the 14 PSM elements in the spokes-and-wheel model, as shown in Figure 2. An ideal data management system should be able to manage all 14 PSM elements in an integrated manner to get the full benefit of the system. For example, an incident that triggers a record into the incident management module should be able to be linked to the resulting MOC record, if one of the recommendations from the investigation is to execute a process change.

Subsequently, the MOC record should be able to be linked to a PHA record, which may be triggered by the request for the change, and so on. All action items should also be managed centrally, enabling users to have a complete view of all their obligations. This is certainly more preferable than having multiple systems managing the various PSM elements, which will require resources to download and consolidate data from the multiple source points.

In fact, having multiple systems will defeat the benefit of having an IT-based data management system. Also, because workflows may change over time, the chosen system needs to be configurable to give the organization sufficient flexibility to accommodate future changes. But the system is just a tool to facilitate the overall objective. Supporting business processes, operational discipline, leadership support, and the right organizational culture are still important and needed for an effective performance management program. Also, the system will work only as well as the workflow that is configured into it. An effective system will have a practical, yet rigorous element management workflow configured into it.

Having a robust performance management system is critical for the safe operation of an organization. The choice of metrics is important, but equally important are the implementation of an effective data collection and rigorous review process, and the subsequent means to initiate and track actions. Many organizations struggle with these elements. They are usually faced with some common functional and cultural issues when implementing such a system.

Having a non-paper-based data management system can help. It allows for data to be collected as part of workflow management and be produced on demand. The concept of producing versus collecting data for performance management is a paradigm shift that may help organizations with their journey of continuous improvement.

Alfonsius Ariawan is global solutions architect with DuPont Sustainable Solutions (DSS).