Monitoring, audit and management review at a plating plant

Under 9.1.1 the organisation must determine what needs monitoring, the methods, the criteria, when to monitor, and must ensure monitoring and measuring equipment is calibrated or verified and maintained. Atlas fails on several counts:

• The anemometer. Data from an instrument with no valid calibration cannot support conclusions about LEV performance. Either recalibrate and re-baseline, or treat the readings as indicative only - but the decision and its rationale must be recorded. Verification of monitoring equipment is a requirement, not good practice.
• LEV thorough examination. Statutory examination by a competent person is a legal requirement in most jurisdictions, and its absence is both a 6.1.3 and a 9.1.2 failure. Monthly flow checks do not substitute for it.
• Health surveillance coverage. 60% attendance means the organisation cannot demonstrate that it is detecting work-related ill health. Day-shift-only clinics are a design flaw; the legal duty usually attaches to the exposed population, not to the convenient population.
• Compliance evaluation (9.1.2). There is no evidence of systematic evaluation of compliance against the legal register, with frequency, method and records.
• Data not used. 212 near misses with two confirmed dermatitis cases is a signal. Frequency, location, task, shift and body-part analysis would show whether the pattern sits in tank loading, splash during maintenance, or PPE availability.

The general rule: monitoring that never changes a decision is theatre. Each measure should have a defined owner, a review frequency, and an action trigger.

The programme fails 9.2.2 in three ways. Coverage: auditing the same four departments annually while never auditing the night shift, contractors or the wastewater plant means significant risks and significant processes are outside assurance. Basis: scheduling should be risk-based, taking account of the importance of processes, changes affecting the organisation, and the results of previous audits - not simply rotating through familiar areas. Objectivity: the HSE coordinator auditing her own area breaches the requirement to select competent auditors and ensure objectivity and impartiality; nobody should audit their own work.

Rebuilding it:

1. Build a risk-based audit schedule from the significant risks and legal obligations, with the plating lines, night shift, contractor control and wastewater plant on a shorter cycle than low-risk administrative areas.
2. Define auditor competence: OH&S knowledge, audit technique, familiarity with the specific processes, and independence from the area audited. Use trained cross-functional auditors or a qualified external auditor where internal independence is impossible.
3. Use the standard's audit criteria: conformity with ISO 45001 and with the organisation's own arrangements, plus effectiveness. Auditing documents alone will not find an unexamined LEV system.
4. Include worker interviews and shift observations, and audit contractors as contractors under 8.1.4.
5. Report results to top management and to the workers concerned, grade findings, and track corrective actions to verified closure under 10.2.

Audit results are an input to 9.3, so a weak programme weakens management review as well.

Clause 9.3.2 requires defined inputs, and most of them are absent. Missing at Atlas:

• status of actions from previous reviews;
• changes in external and internal issues relevant to the OH&S system;
• fulfilment of legal requirements and other requirements - the central question given the LEV and surveillance failures;
• OH&S risks and opportunities and other issues, including the dermatitis cases and what they say about exposure control;
• extent to which objectives have been met;
• OH&S performance information: incident trends, monitoring and measurement results, conformity with legal requirements, audit results, and consultation and participation of workers;
• adequacy of resources - calibration, clinic hours, auditor competence all require decisions;
• relevant communication with interested parties, including the regulator and workers;
• opportunities for continual improvement.

A single page of minutes suggests a discussion, not a review. The practical difference a proper review makes is that it forces the linkage: dermatitis cases plus uncalibrated monitoring plus 60% surveillance coverage is a resource decision and a control decision, not three unrelated facts.

Outputs under 9.3.3 must then be recorded: decisions on improvement opportunities, needed changes to the system, resource requirements, and any actions, including integration with other business processes. Those outputs become the 10.1 and 10.2 agenda and the next review's first input, which is what turns the cycle into improvement rather than annual ritual.