A safety and health management system that exists is not the same as one that works. QGN 09, the Queensland guidance note on reviewing the effectiveness of the SHMS, is built around that distinction. It covers obligations, acceptable level of risk, workforce involvement, lead and lag indicators, causal analysis, audit findings, contractor performance, and the chronic exposures that never appear in an incident statistic.
This guide covers what an effectiveness review actually has to establish, why the usual metrics answer the wrong question, and how the review connects to the management structure that RS22 requires.
The question the review has to answer
The obligation under the Queensland legislation is to achieve an acceptable level of risk. QGN 09 breaks that into its two established components.
- Risk within acceptable limits. The residual risk sits below the threshold the operation has defined as tolerable.
- Risk as low as reasonably achievable. Beyond being tolerable, everything reasonably practicable has been done.
Both must hold. A site sitting comfortably inside its tolerability threshold while declining a control it could reasonably implement has not met the obligation, and this is the reasoning set out at length in acceptable level of risk and what you can prove.
The effectiveness review is therefore not asking whether the system is documented, audited and current. It is asking whether the risk at this mine is actually being held at an acceptable level by the system as it operates, which is a much harder question and one that document review cannot answer.
Lead and lag indicators, and why lag alone misleads
QGN 09 treats system performance through both lead and lag indicators. The distinction is well known and routinely applied badly.
Lag indicators count what has already happened: injuries, high potential incidents, property damage. They are objective, comparable, and describe a period that is over. Their fatal weakness in a mining context is statistical. Serious events are rare enough that a year without one is entirely compatible with a system that is failing, and everyone knows a site that ran a record year immediately before a fatality.
Lead indicators measure whether the things that prevent events are actually happening: inspections completed on time, critical control verifications performed and passed, actions closed within their due dates, hazards reported and resolved, training current.
The failure mode of lead indicators is different and just as damaging. They are easy to satisfy without effect. A hundred per cent inspection completion rate where the inspections are cursory measures compliance with a schedule, not control effectiveness. The useful lead indicators are the ones that can return bad news: the proportion of critical control verifications that found a defect, the age profile of open actions, the ratio of hazards found by supervisors to those found during audits.
A review that reports only green lead indicators and a clean lag record has usually measured its own reporting rather than its system.
Causal analysis and repairing defences
QGN 09 addresses causal analysis in terms of repairing defences, which is a deliberate framing. When something goes wrong, the question is which defences did not hold and why, not who did the wrong thing.
The practical test of whether a site does this properly is what its corrective actions look like. If the actions arising from investigations are overwhelmingly retraining, toolbox talks and procedure amendments, the site is repairing people rather than defences. Those actions are cheap, fast to close, and almost never change the outcome, which is why the same event recurs with a different name attached.
An effectiveness review should look across a year of investigations at the distribution of control types in the corrective actions. That single view tells you more about the health of the system than any incident rate, and the underlying pattern is explored in why critical controls fail.
Audit and inspection findings as data
Audits and inspections produce findings, findings become actions, and actions get closed. Most sites manage that pipeline adequately and stop there.
The effectiveness question is different. Are the same findings recurring? Are they recurring in the same areas, or under the same supervisors, or in the same part of the system? Is the same non conformance being closed repeatedly, which means the closure is not working? Are findings from external audits materially different from the ones the site generates itself, which means internal assurance is not looking where it should?
Answering those questions requires findings to be structured and located rather than held in a stack of reports. A site that can only read its audit history one document at a time cannot see repetition, and repetition is the signal.
Contractors, and the boundary that is not a boundary
QGN 09 deals with contractor safety performance as part of system effectiveness, not as a separate reporting line. On most open cut operations contractors perform a significant share of the highest risk work: drill and blast, earthworks, maintenance shutdowns, electrical work.
A review that assesses only the operator’s own workforce is assessing a minority of the exposure. Worse, contractor performance data that arrives in a different format, on a different cycle, and is reported separately, cannot be combined with site data to see anything. The principle of one system rather than three is covered in contractor management under one system.
Chronic exposures
QGN 09 explicitly addresses chronic exposures causing incapacity, and this is where most effectiveness reviews are weakest.
Dust, noise, diesel particulate, whole body vibration and hazardous chemicals injure people on a timescale of years to decades. They produce no incident, no investigation and no lag indicator until long after the exposure that caused them. A system judged on incident performance will rate itself highly while accumulating a liability that surfaces after the people involved have left the industry.
Reviewing effectiveness for chronic exposure means looking at exposure monitoring results and trends, at control verification for engineering controls, and at whether monitoring results actually change anything. The dust side of that is covered in RS20 and dust control in surface coal mines.
Workforce involvement
The guidance note treats workforce involvement as part of effectiveness, and there is a diagnostic hidden in it. If hazard reports are declining, that is either a safer site or a workforce that has stopped reporting. Those two look identical in the data and have opposite meanings.
The way to tell them apart is to look at what happens after a report. Time to first response, proportion actioned, and whether the person who reported it ever finds out what happened. Reporting dies where it produces no visible result, and once it dies the system loses its most sensitive early indicator.
Where sites come unstuck
- The review audits compliance instead of effectiveness. Every document current, every audit closed, and no evidence about whether risk is actually controlled.
- Lead indicators that cannot fail. Metrics measuring whether activity occurred rather than what it found.
- The review is annual and disconnected. A once yearly exercise producing a report nobody uses, rather than a continuous reading of the system.
- Actions are counted, not weighted. Ninety closed toolbox talks outrank one unclosed engineering control in the completion statistics.
- Contractors excluded. The majority of high risk work assessed separately or not at all.
- Chronic exposure ignored because nothing has happened yet, which is exactly what chronic means.
Getting data that can answer the question
Almost every failure above is a data structure problem wearing a safety costume. The questions QGN 09 asks are comparative and longitudinal: is this recurring, is it getting worse, where is it concentrated, what changed after we acted. Those questions cannot be answered from documents, only from records that are structured, located and durable.
HazView produces that shape of data as a by product of the work. Hazards, inspections, controls and actions are recorded against the place and the asset they concern, with the person, the time and photographs attached, and kept permanently. That makes the review questions answerable directly: which controls fail verification most often, where do findings repeat, how long do actions actually take to close, whether hazard reporting is holding up in a given crew or quietly stopping.
Effectiveness review also needs data the platform does not hold: exposure monitoring, health surveillance, contractor performance from a contractor’s own system, corporate incident data. The REST API is how those get into the same view rather than into a separate slide. Feeding data in and out, so the analysis happens on one dataset instead of on four exports reconciled by hand, is the whole difference between a review that finds something and a review that reports something.
The judgement stays with people. Deciding whether risk is at an acceptable level is the site senior executive’s call and not a calculation. What better data changes is whether that judgement is made on evidence or on impression.
See how HazView works for managers and SSEs, or book a demo.
Related guides
- Recognised Standards for Queensland Open Cut Coal Mines
- Emergency Preparedness for Small Mines and Quarries
- RS22: Management Structure for the Safety and Health Management System
- The Mine Record: What Queensland Law Actually Requires
- All QLD Open Cut Coal guides
Frequently asked questions
What is QGN 09?
The Queensland guidance note on reviewing the effectiveness of the safety and health management system. It is guidance rather than law, and compliance with it is not by itself sufficient to comply with the legislation.
How often should an SHMS be reviewed?
Formal review on a defined cycle, plus review triggered by change: new plans, significant incidents, restructures, new fleets or major contractor engagements. An annual review with no trigger based reviews will always be looking at a system that has since moved.
What is the difference between a lead and a lag indicator?
Lag indicators count outcomes that have already occurred. Lead indicators measure whether the activities that prevent those outcomes are happening. Both are needed, and lead indicators are only useful if they are capable of returning bad news.
Does a clean incident record mean the system is effective?
No. Serious events are statistically rare, so a period without one is entirely consistent with a failing system. It is evidence of an outcome, not evidence of control.
Should contractors be included in the review?
Yes. QGN 09 treats contractor safety performance as part of system effectiveness, and on most open cut operations contractors perform a large share of the highest risk work. Excluding them assesses a minority of the exposure.
How does this relate to RS22?
RS22 asks whether the management structure for developing and implementing the system is defined and competent. QGN 09 asks whether the system that structure runs is actually working. They are the two halves of the same question and are best reviewed together.