04 Baseline Metrics and Indicator Set

This note defines the numbers the department collects to know whether its safety management system is in place and whether it works. It sets the baseline worksheet, the indicator definitions, the dashboards and the cautions about comparing with anyone else. Every program module names its indicators from this set; every Program Improvement Team reports them quarterly; the Gap Assessment scores Effectiveness on them; the management review reads them once a year and sets targets (ISO 45001:2018, cl. 9.1; NFPA 1550, 2024, ch. 6).

Leading over lagging. Injuries, exposures with medical follow-up, collisions and claims are the outcomes the system exists to prevent, and they must be counted. But in a department of 50 to 500 members they are rare enough that a year without one proves nothing, and a year with one proves little. Leading indicators measure whether the controls that prevent those outcomes are in place and used: fit tests current, rehab logged, exposure reports filed, mayday drilled, inspections done, corrective actions closed. A department that moves the leading indicators has done the work; the lagging indicators follow over years. Managers who are sceptical of “soft” measures accept leading indicators when each is paired with the outcome it protects, so this set pairs them (05 Design Principles and Lessons Learned, principle 7; Hollnagel, 2014).

Measure the loop. The package’s design principle is the closed loop: report, assess, act, verify, tell. So the set measures the loop itself: how long a corrective action stays open, how fast the JOHSC gets an answer, whether learning reviews finish, whether program teams report on time, whether the person who raised a concern heard back. A system whose loop measures are poor is not learning, whatever its injury rate.

Psychological safety is measured as a named group. Near-miss and good-catch reports, stop-work calls reviewed without reprisal, learning-team participation and members’ confidence in raising a concern with their officer are indicators of whether people will tell the department what it needs to know (Edmondson, 1999)DOI. They are reported together, by station and by career and volunteer status, and a rising near-miss count is reported as success, in writing (05 Design Principles and Lessons Learned, principle 8).

Denominators matter. Rates per 100 members and per 100 working fires are reported alongside counts, so a volunteer station with eight members and a career platoon with 60 can be read on the same page. Working fires, hours in the hot zone and responses are the denominators for exposure measures.

Aggregate only for health data. Medical, fitness, psychological and peer support measures are reported as counts and percentages with no individual identifiable; the custodian of the underlying records is the physician or clinician, not the department (Records, Privacy and Health Information).

Fifty-five indicators in seven groups. Type: L leading, G lagging. Frequency: how often it is computed. Owner: the role accountable for collecting and reporting it. Target guidance is a starting point for the management review, not a standard; the department sets its own target from its baseline.

Count: 55 indicators (A 9, B 7, C 6, D 9, E 11, F 5, G 8), 46 leading and 9 lagging.

Complete before the first management review and keep with the Profile Worksheet. Where data does not exist, write “not collected” and the date collection starts; that is itself a finding.

Baseline rules. Three years of lagging data where it exists (from Part D of the Profile Worksheet); one quarter of leading data collected deliberately before the baseline is declared; the baseline is declared at the first management review and never moved afterwards, so every later result is read against it.

Twelve numbers, each with the value, the target, the trend arrow against the previous three months, and a colour (green on target, amber within 10 percent, red beyond). Below them, three lines of text: the oldest open corrective action and its owner, the most recent stop-work or near miss and what changed, and the one decision the chief is being asked to make.

The chief’s page, plus: all of group D by station and by career and volunteer status; E6 and E7 inspection completion and closure by station; the near-miss and hazard report themes for the month (no names); every High and Extreme register entry re-scored or overdue; every regulator contact; the list of safety-sensitive changes awaiting consultation. The committee minutes record what the committee did with each red item.

Their own program’s indicators by district and overall, the trend against baseline and target, the register entries they own, their open corrective actions, their learning-team sessions and what they found. This is the content of the quarterly progress update (06 Program Improvement Teams section 10).

Four pages in the annual health and safety report (03 Accountability and Reporting): the department’s maturity band and stage; the category profile from the gap assessment or audit; eight indicators with three-year trends (A1, B1, C1, D1, D4, E1, G1, G2) with rates per 100 members; presumptive claims and the exposure records that support them (aggregate); the three largest safety risks and what they cost to control; the decisions council is asked to make. No names, no individual health data, written for public disclosure.

• One source of truth per indicator; the owner named in section 3 signs the number.
• Career and volunteer shown separately wherever practice differs; station-level detail available on request.
• Every red number carries an owner and a date, not an explanation.
• The dashboard is produced by the HSO from the records system; where the records system cannot produce it, the gap is itself a corrective action.

• The United States numbers are not the department’s numbers. The NFPA annual firefighter injury report and the USFA fatality reports are the richest public benchmarks, but they describe a different legal, staffing and medical system, and injury definitions and denominators differ (NFPA, n.d.; USFA, 2025). Use them to understand causes, not to set targets.
• Canada has gaps. The National Fire Information Database holds public casualty tables for 2005 to 2021 and appears dormant; there are no official national line-of-duty death statistics; the Great Canadian Fire Census is the current national data effort. Within a province or territory, the department may have to report fires and casualties to the fire marshal or fire commissioner (J34 Fire Services Legislation and Firefighter Certification), and the compensation board publishes statistics by industry (J32 Workers' Compensation and Firefighter Presumptions). Expect to benchmark against the department’s own history first.
• Small numbers mislead. In most departments the annual count of lost-time injuries is in single digits. Report three-year rolling rates and read them with the leading indicators; never set a target of “zero injuries” as a performance measure, because it suppresses reporting.
• Rising reports are not rising risk. Near-miss, hazard and exposure report counts rise when the system starts working. Say so in every report that shows them.
• Volunteer denominators. Hours worked are often unavailable for volunteers; use per 100 members and per 100 responses, and say which denominator each number uses.
• Culture and psychosocial scores are survey instruments, not audits; compare a department with itself over time, segmented by station and status, not with another department’s score (05 Safety Culture Baseline).
• Presumptive coverage lists differ by province and territory and change; a claim count says as much about the law as about the hazard (J32 Workers' Compensation and Firefighter Presumptions).

Concern-to-answer route. A member who questions a number, or reports that a control the dashboard shows as green is not working at their station, raises it through the officer, the JOHSC member or the hazard report. The HSO answers within 30 days and the station-level figure is checked at the next inspection.

Loop test. A volunteer captain says the mayday currency shown for her station is wrong because the drill was a classroom talk. By the next monthly page the HSO has checked the training records, re-scored C2 for that station as not current, and the training officer has scheduled a practical drill as a P2 item, closed within 30 days. The captain has been told, and the JOHSC pack shows the correction and why.

This chapter relies on the following legal requirements, which differ by province and territory. Check your jurisdiction’s section on each topic page before adopting it.

National Fire Protection Association [NFPA]. (n.d.). Firefighter injuries in the United States [Annual report series]. https://www.nfpa.org/education-and-research/research/nfpa-research/fire-statistical-reports/firefighter-injuries-in-the-united-states