Major accidents continue to happen in the process industry and often have serious consequences. There are questions on how these accidents happen and how we can monitor safety in a process plant to prevent these accidents from happening. The possibility of measuring the safety level in a process plant is very compelling and different approaches using key performing indicators (leading and lagging indicators with some metrics) have been proposed. The industry has evolved towards a complex sociotechnical environment where the traditional accident causation models used that try to explain how and why accidents happen are no longer valid (or at least are not the fittest option) to analyse these complex systems. New accident models have been developed lately trying to cope with the complexity of new industrial systems. These (relatively) new systemic models have their roots in systems theory. In this work, a novel methodology is presented to measure safety level in a process plant using a systemic approach. The systemic model used in this work is Systems Theoretic Accident Model and Processes (STAMP). The measurement of the safety level is done according to the status of safety checkpoints which is a new concept introduced in this work. A safety checkpoint is a requirement defined to prevent a loss scenario to happen or to satisfy safety measures associated to the loss scenario. The outcome of the analysis obtained with this methodology can be presented in different ways to help management to make informed risk-based decisions and take the actions needed to prevent accidents.