SIF Precursor Review: Build It in 14 Days

The Campbell Institute's 2018 SIF prevention research made one problem hard to ignore: recordable injury rates can improve while serious injury and fatality exposure remains alive in the work. This guide shows EHS managers how to build a 14-day SIF precursor review that converts weak signals, control gaps, and field evidence into decisions before the monthly dashboard turns green.

Why a SIF precursor review is different from a safety dashboard

A SIF precursor review is a short-cycle routine that asks where serious injury or fatality exposure is forming, even when no recordable injury has occurred. A dashboard usually explains what already entered the database, while a precursor review tests whether high-consequence work is drifting away from its controls.

As Andreza Araujo argues in her Portuguese title Muito Além do Zero, translated as Far Beyond Zero, lagging indicators look in the rearview mirror because they show the consequence, not the cause. That distinction matters because a clean month can mean capability, luck, or underreporting, and only one of those deserves executive confidence.

Use the review for high-risk activities such as energy isolation, confined space, lifting, work at height, mobile equipment, hot work, and chemical exposure. If your operation already tracks fatal-risk blind spots in the safety dashboard, this routine becomes the weekly evidence engine behind that view.

Step 1: Define the review boundary

The review boundary tells the team exactly which work enters the 14-day scan. Start with 5 to 8 high-consequence work families, since a wider scope usually produces noise before it produces action.

Across 25+ years leading EHS in multinational operations, Andreza Araujo identifies that safety teams often fail because they try to review every minor event with the same intensity. A SIF review has a different purpose, because it protects life where the credible worst outcome is severe, not where the paperwork is easiest to collect.

Choose the boundary with operations, maintenance, EHS, and the area owner in the same room. Write one sentence for each included work family, name the critical control attached to it, and exclude low-consequence items that belong in ordinary trend review.

Step 2: Build the precursor dictionary

A precursor dictionary converts vague concern into repeatable classification. The dictionary should define the event type, exposure activity, failed or degraded control, potential severity, and review owner in no more than 1 page.

What most safety blogs do not mention is that a precursor label without a control reference becomes a new way to count anxiety. The useful question is not whether something felt risky, but which barrier was missing, weak, bypassed, late, or unverifiable at the moment risk escaped its ordinary boundary.

Use 4 fields at minimum: activity, control, deviation, and potential outcome. For example, energized maintenance with lockout verification skipped is not just a near miss, because it is a high-consequence exposure whose control failed before harm appeared.

Step 3: Pull the first 30 signals

The first dataset should be deliberately small because the goal is calibration, not a permanent database design. Pull 30 signals from incident reports, near misses, stop-work records, field inspections, permit audits, maintenance backlog, and supervisor notes.

In A Ilusão da Conformidade, translated as The Illusion of Compliance, Andreza Araujo warns that good indicators do not guarantee good practices. This is exactly where a precursor review earns its place, because it compares the official record with what supervisors and technicians can prove in the field.

30 signals is enough for the first calibration round if the sample includes at least 3 work fronts and 2 shifts. Avoid building the review from injury records alone, since the most valuable precursor is often the one that never became an injury.

Step 4: Score potential severity before actual outcome

Potential severity asks what could reasonably have happened if timing, distance, energy, or recovery had changed. The actual outcome stays visible, but it does not control the score.

This step is where many dashboards hide fatal risk because a dropped load that missed a worker by 2 meters can be treated as minor, while the control failure was severe. Andreza Araujo's work on zero-accident thinking is useful here because the absence of injury should not be confused with proof of control.

Use a 3-level model for the first version: high potential, serious potential, and local potential. If your team already has a severity weighting model, connect the precursor score to that model so the monthly review does not run on two competing languages.

Step 5: How do you test whether controls really worked?

Controls really worked only when the team can show that they were present, understood, used, and verified at the point of risk. A signed permit or completed checklist is evidence of administration, not evidence of control effectiveness.

Named sources such as the Campbell Institute's 2022 work on leading indicators for SIF prevention emphasize control effectiveness and high-consequence exposure, not generic activity volume. The trap is counting 100 inspections without knowing whether any inspection looked at the control whose failure can kill someone.

For each high-potential precursor, require 2 forms of evidence: document evidence and field evidence. Document evidence may be a permit, isolation certificate, lift plan, or training record, while field evidence may be a photograph, supervisor verification, worker interview, or instrument reading.

Step 6: Link each precursor to one owner

Each precursor needs one accountable owner because shared ownership usually becomes delayed ownership. The owner is not blamed for the event, but is responsible for restoring the control and proving that the restoration happened.

James Reason's latent-failure model helps keep this step disciplined because it directs attention to the conditions that made the failure possible. The review should ask why the control was weak, where the same condition may exist, and whose authority is needed to remove the condition.

Assign the owner closest to the decision that can change the system. A supervisor may own a verification gap, maintenance may own a backlog gap, procurement may own a contractor equipment gap, and the plant manager may own a production-pressure pattern that repeats across areas.

Step 7: What should enter the executive review?

Only precursors that show high potential, repeated control degradation, cross-area exposure, or unresolved authority conflict should enter the executive review. Everything else belongs in the local action tracker.

In more than 250 cultural transformation projects, Andreza Araujo observes that executives lose trust in safety data when every item arrives with the same urgency. The better route is escalation discipline, because senior leaders need the few signals that require power, capital, or trade-off decisions.

4 escalation filters keep the review focused: credible fatality potential, repeated degradation, weak owner authority, and overdue control restoration. If the topic already appears in weak signal metrics for board questions, it should not be buried in a local spreadsheet.

Step 8: Close the 14-day loop

The 14-day loop closes when each high-potential precursor has a classification, owner, control-restoration action, due date, and verification method. Without those 5 elements, the review has produced discussion rather than protection.

Andreza Araujo's position is practical: safety does not improve because an organization counts more things, but because it makes better decisions from the things it counts. That is why the final review should be short, operational, and uncomfortable enough to challenge the green areas.

Run the final meeting for 45 minutes. Spend 10 minutes on the pattern, 20 minutes on the highest-risk controls, 10 minutes on owner conflicts, and 5 minutes on what enters the next monthly dashboard.

Comparison: dashboard review vs SIF precursor review

Each month without a SIF precursor review lets high-potential events compete for attention with low-consequence activity counts, which means the next executive safety meeting may celebrate the number while missing the exposure.

Conclusion

A SIF precursor review is not another dashboard layer, but a disciplined routine for finding fatal-risk exposure before injury statistics confirm the failure.

If your organization needs to redesign safety metrics around serious-risk exposure, Andreza Araujo's team can support the diagnostic, calibration, and executive-review model. Start with leading indicator quality, connect it to the right safety metric denominator, and request support at Andreza Araujo.