I85PREP
⌘ K
OHST study materials
OHST study resources

The Zero-Incident Bonus and a Collapsing Reporting Culture

OHST · Put your knowledge to work in context.

All case studies

The Zero-Incident Bonus and a Collapsing Reporting Culture

Ridgeline Components (420 employees, two plants) cut its total recordable case rate from 6.1 to 3.4 over three years. Over the same period, first-aid-only visits logged by the plant nurse fell 71%, and near-miss cards dropped from 38 per month to 6. In January the VP of Operations launched a 'zero-incident' bonus: supervisors receive $1,500 per quarter when their area has no recordable injury. In February a press operator's hand injury was logged by the shift supervisor as a 'pinch' and treated by the nurse as first aid; the employee's outside clinic follow-up produced three sutures. The safety manager has one full-time technician and a $25,000 budget. Line supervisors say the plant has 'never been safer.'

1. What does the trend data most likely indicate, and what specific failure modes does the bonus introduce? State clearly what you can and cannot conclude.

Read the worked answer

What the numbers say

  • 420 employees at roughly 2,000 hours each is about 840,000 hours per year. At a TRIR of 3.4 that is roughly 14 recordable cases; at 6.1 it was about 26. That move is real but the annual count is small enough that a few cases of classification drift can manufacture a good chunk of the improvement on their own.
  • The problem is the combination. Genuine risk reduction usually leaves first-aid visits flat or rising, because a healthier reporting culture surfaces minor injuries earlier. Here first-aid visits fell 71% while near-miss reports fell 84%. Those two falling together is a reporting-suppression signature, not a safety signature.

Failure modes the bonus creates

  1. Classification pressure. The payout hinges on whether a case is 'first aid' or 'recordable,' which is a judgment made by a supervisor and a nurse. Three sutures is medical treatment beyond first aid, so the February case is very likely recordable and the initial classification was wrong.
  2. Under-reporting at the source. The cheapest way for a supervisor to earn $1,500 is for the employee not to report. That is precisely the behavior prohibited by the recordkeeping anti-retaliation provisions.
  3. Nurse role conflict. The nurse is being asked to make classification calls that affect a colleague's pay.
  4. No leading data to steer with. Supervisors of the two plants are being managed with a metric that arrives months late.

What you cannot conclude You cannot conclude the rate reduction is fake, that any individual supervisor suppressed a case, or that the bonus caused the February misclassification. Attribute honestly as 'the pattern is inconsistent with genuine risk reduction and creates incentives for non-reporting' and then go gather evidence: nurse log versus supervisor reports versus first reports of injury versus workers' compensation claims.

2. Propose a replacement measurement set and incentive design that the VP could actually adopt in the next quarter.

Read the worked answer

Design principle: pay for the behavior you control, measure the outcome you care about. Supervisors control reporting, hazard correction, and follow-through; they do not control whether a machine bites someone.

Balanced metric set

  • Lagging (keep, do not lead with): TRIR, DART, workers' comp claim frequency and severity, cost per claim. Use a three-year rolling average to damp small-number noise.
  • Leading: near-miss reports per 100 employees per month; percentage of reported hazards closed within 14 days; scheduled inspections completed; JHAs reviewed on schedule; training completion; corrective actions verified effective (not just closed).
  • Cultural health check: the ratio of near-miss plus first-aid reports to recordable cases. A healthy site typically runs ten or more near-miss reports per recordable. Watch the ratio; a rising TRIR with a healthy ratio is an honest site, a falling TRIR with a collapsing ratio is a hiding site.

Incentive redesign

  • Replace the 'no recordable' trigger with a participation-and-closure trigger: 90% of hazard reports closed within 14 days and near-miss reporting at or above the site's 12-month baseline. Reporting more cannot cost a supervisor money.
  • Make it team/site based rather than area based so no supervisor benefits from another area's silence, and cap the award so it cannot dwarf the cost of a lost-time injury.
  • Keep an executive-level scorecard with the lagging measures, so leadership still sees injuries.

Verification layer (cheap, high value)

  • Quarterly three-way reconciliation: nurse first-aid log vs. supervisor incident reports vs. workers' comp first reports of injury. Cases that appear in one and not the others get reviewed.
  • Audit five randomly selected first-aid-only cases per quarter against the treatment records.
  • Publish the classification rules and make the case-determination process a documented one, with the nurse reporting professionally, not to the plant manager who is paid on the result.

3. Walk through what you must do about the February case and how you would raise the incentive problem with the VP.

Read the worked answer

The February case - sequence

  1. Get the clinical facts, not the supervisor's summary: clinic note, procedure performed, work restrictions, days away or restricted. Sutures are medical treatment beyond first aid, so this is a recordable injury case unless some other specific exemption applies.
  2. Record it. The recordkeeping rule requires entry within seven calendar days of learning of the case; when a case's outcome changes or new information makes it recordable, update the log rather than leaving the original entry. Update the OSHA 300, the 300A totals for the year, and make sure whoever certifies the 300A knows the case changed.
  3. Check the employee was not disadvantaged. Confirm no discipline, no loss of bonus, no attendance point for the clinic visit. If the employee was penalized for reporting, that is a separate and more serious problem than the classification error.
  4. Close the loop with the nurse and supervisor on the classification rule, in writing, without turning it into a blame exercise. Ask the question that gets you the real answer: 'what made this look like first aid to you?'
  5. Sample backwards. Pull 15 recent first-aid-only logs and check the treatment descriptions. You are looking for lacerations with sutures, prescription-strength medications, physical therapy, and lost time logged as something else.

The conversation with the VP

  • Lead with the shared goal and the mechanism, not the accusation: 'We want 3.4 to be real, not a number that moves because cases are being relabeled. Right now the bonus gives every supervisor a reason to make the report disappear, and that is the one behavior we cannot afford.'
  • Bring the three-way reconciliation data and one or two concrete examples, and put the exposure plainly: an employee who is disciplined or pressured for reporting, or a log that is knowingly understated, is regulatory and potentially criminal exposure for the company, plus it distorts the workers' comp reserves and the insurance renewal.
  • Offer a face-saving path: keep the total bonus pool, change the trigger to reporting and closure performance, and run it as a two-quarter pilot with a review. Pilots are easier to approve than reversals.
  • Get one thing signed today: the classification determination is made against the treatment record by the safety function, and the supervisor's pay does not depend on the answer.