Sustainable Corrective Actions - Is SMART Good Enough?

Industry does not have a good track record of implementing lessons learned. Many investigation reports of significant PSM incidents include root causes such as “Corrective Action Needs Improvement” or “Corrective Action Not Implemented”. Many recommendations developed to address PSM incidents are not effective (they don’t directly address the root cause or are not broad enough to look past the specific incident) or are not sustainable (they are not designed with sufficient integration to existing systems to change the way things work).

The acronym, SMART, has long been used as a measure of quality for recommendations, not only within the context of incident investigations, but in audits or general problem solving methodologies. This presentation explores the applicability of SMART and how a more sustainable method must be employed to ensure recommendations for significant incidents are not only effective, but designed to integrate with existing systems and include validation KPIs to track their implementation.

Methods & Procedures:

The use of the SMART acronym was reviewed in the context of more recent research. While the components of SMART remain important, even recommendations that meet SMART requirements are often subject to failure for reasons mentioned previously, among others.

First, many companies fail to tie identified root causes directly to the recommendation that is supposed to address it. A format was developed, along with investigation quality assurance processes, that require the basis for the root cause (the finding that produced it) to be directly addressed by the recommendation.

Second, recommendations often address only local failures (this department, this rig, this platform, this incident, etc.). The format also requires asking key questions to determine if the root cause is “generic” in nature and should be expanded to address possible wider deficiencies (embedded in the work systems of the organization as a whole).

Third, since a recommendation is simply a form of risk control, a review of the types of risk controls was completed. Then an internal company review was performed to determine the types of systems that were in already place to support similar types of risk controls. From this review a list of risk control assurances was developed with which to integrate and to evaluate each recommendation for sustainability. The likelihood of a recommendation being sustainable could be directly related to the number and level of assurances that were embedded within the recommendation.

Finally, KPIs that show the percentage of recommendations completed on time are often not indicative of the successful implementation of a recommendation, but simply the desire to “close” the recommendation for the sake of achieving a KPI. A system of KPI validation was developed to address this issue that looks at the success of the recommendation over a period of 18 months to assure sustainability.

Results & Conclusions:

While the specific aspects of SMART are still important, even with all the attributes of SMART in place, the model demonstrates how corrective actions/recommendations using only the SMART acronym can still be highly likely to fail. An algorithm allocating higher effectiveness to engineering controls, a lower effectiveness to administrative controls and the lowest effectiveness to PPE controls was developed so that the effectiveness of one corrective action or group of actions could be compared to another action or group of actions. The use of quality assurance processes to review investigation reports and ensure that recommendations address the respective bases for root causes and include risk control assurances has significantly improved the quality and sustainability of recommendations. A project is currently underway that will review the implementation of recommendations over an 18 month period to ensure sustainability and provide a check to the KPI of “Percentage of Recommendations Completed On Time”.

Novel & Additive Information:

The algorithm can also be used to look at effectiveness per resources required for implementation. By dividing the cost of the corrective actions by the effectiveness score, the efficiency ($/Risk Reduction Unit) of the corrective actions can be compared.