A Safety Scorecard That Looks Fine and Is Not
Ridgeway Metalworks runs one plant with 480 employees and worked 1,040,000 hours last year. The OSHA 300 log closed with 11 recordable cases, 7 of them DART cases; the prior year had 6 recordables and 2 DART cases. The safety manager resigned five months ago and the plant manager has been certifying the log himself. The safety committee meets quarterly, its minutes list only housekeeping items, and it has never reviewed the 300 log. Two of this year's recordables are ergonomic injuries in the same assembly cell; a third entry is a forklift pedestrian near miss that a supervisor logged as first aid. The corporate VP has approved $150,000 for one initiative: a guarding upgrade for six press brakes, a lift-assist system for the assembly cell, or a consultant-led behavior-based observation program. The plant manager favors the consultant because it 'changes culture.'
1. Calculate the year's TRIR and DART rate, then explain what those numbers do and do not tell you about where to act.
Read the worked answer
TRIR = (recordable cases x 200,000) / hours worked = (11 x 200,000) / 1,040,000 = 2.12. DART = (7 x 200,000) / 1,040,000 = 1.35.
What the rates say: a TRIR near 2.1 is unremarkable for metal fabrication, so a benchmark-only comparison would conclude the site is fine. What they do not say: the year-over-year movement (6 to 11 recordables, 2 to 7 DART) is a near doubling of cases and a tripling of days-away/restricted cases, and the DART share (7 of 11, about 64%) is unusually high. Either previously missed cases are now being recorded, or severity is genuinely rising; both demand investigation before any trend claim is made.
The cluster matters more than the rate. Two ergonomic cases in one assembly cell is a design problem, not a worker problem. The forklift near miss is not recordable at all because no injury occurred, but logging it as first aid shows supervisors do not understand 29 CFR 1904.7, which predicts both under-recording and misclassification. Reconcile the 300 log against clinic, workers' compensation, and first-aid records before drawing conclusions about trend.
2. Recommend how the $150,000 should be spent and justify the choice using the hierarchy of controls and the loss data.
Read the worked answer
Use the hierarchy of controls and the actual loss data. The only demonstrated injuries are the two assembly-cell ergonomic cases, and a lift-assist is an engineering control that removes the exposure rather than asking the worker to behave differently. The press brakes have produced no injuries yet, but an unguarded point of operation is a low-frequency, high-severity exposure (amputation), so it cannot be deferred indefinitely. The consultant-led behavior program is the weakest of the three: it is administrative, its effect decays without reinforcement, and if observations feed discipline it can suppress reporting, which 29 CFR 1904.35(b)(1)(iv) prohibits. A shrinking log would then look like success while risk stays constant.
Practical allocation: fund the lift-assist now and use remaining scope to start press-brake guarding, sequencing the machines with the shortest cycle times and no presence-sensing devices. If the guarding package exceeds the remaining funds, commit it as capital in the next cycle, and in the interim verify existing guards, add presence sensing where feasible, and prohibit bypass without a documented management-of-change review. If the consultant is retained, buy supervisor coaching tied to hazard correction, not a culture campaign.
3. The plant manager asks for a 90-day plan to close the program governance gaps. What should it contain?
Read the worked answer
Days 0-30: name an interim safety lead with stop-work authority and a defined reporting line to the plant manager. Personally reconcile the 300 log against clinic, workers' compensation, and first-aid logs; add or reclassify entries as needed; confirm the log is certified by a company executive and retained five years (1904.32, 1904.33). Brief all supervisors on 1904.7 recordability criteria and on the distinction between a near miss and first aid, and open a no-blame near-miss channel separate from the injury log.
Days 30-60: rebuild the committee with trained members, a management representative who holds budget authority, and monthly meetings. Adopt a standing agenda: log review, incident and near-miss review, corrective-action status, and hazard reports from the floor. Publish corrective-action closure rates so the committee has visible output.
Days 60-90: shift the dashboard to leading indicators, for example percent of machine guards verified each month, near-miss reports per 100 employees, corrective actions closed on time, and training completion. Set a 12-month target for DART reduction, not only TRIR, and add an annual recordkeeping audit. Do not tie any indicator to individual discipline.