DART & LTIR, Down and to the Right - A 90-Day Path With Computer Vision

Driving down your Days Away, Restricted, or Transferred (DART) rate and Lost Time Injury Rate (LTIR) takes a clear plan grounded in real data. This article lays out a 90-day plan that uses computer vision to connect everyday activity on the floor with fewer recordable incidents.

The program connects what you see on the floor with the numbers you report at the end of the month. It gives EHS leaders the hard data needed to show progress and get leadership on board.

This 90-day plan will help you:

• Set clear safety baselines and realistic reduction targets.
  
  
• Put computer vision to work in high-risk zones to spot unsafe behaviors.
  
  
• Run quick weekly triage meetings to assign and track fixes.
  
  
• Show your monthly progress against a clear DART and LTIR glide path.

‍

Set Baselines and Targets to Reduce DART and LTIR

A practical safety plan starts with knowing exactly where you stand.

To start, you need to pin down your current rates, total hours worked, and how serious recent incidents have been. With that information, you can set a baseline for the whole site and even for specific lines, areas, and shifts.

From there, you can set a quarterly reduction target based on your headcount and expected work hours.

‍

DART and LTIR Formulas

OSHA sets the formulas for DART and LTIR. They both use the number 200,000, which is what 100 full-time employees would work in a year.

• DART Rate = (Total DART-recordable cases × 200,000) ÷ Total hours worked
  
  
• LTIR = (Total lost time cases × 200,000) ÷ Total hours worked

To get the math right, you need to classify every case the same way according to OSHA recordkeeping standards. You also have to precisely track all hours worked across all employees and contractors.

‍

Baseline Math: Example

Let's walk through an example. Say a facility had 450,000 work hours last quarter. During that time, they recorded 9 DART cases, and 5 of those were lost time cases.

1. DART Rate Calculation: (9 cases × 200,000) ÷ 450,000 hours = 4.0
2. LTIR Calculation: (5 cases × 200,000) ÷ 450,000 hours = 2.22

For some extra context, if the site had 12 total recordable incidents, its Total Recordable Incident Rate (TRIR) would be 5.33. The severity breakdown shows that of the 9 DART cases, 5 led to days away from work, and 4 ended in restricted duty or a job transfer.

‍

Target Setting - Glide Path

With a DART rate of 4.0, leadership might approve a 90-day goal to get it down 20% to 3.2. If you expect another 450,000 hours for the next quarter, this new rate means you can have a maximum of 7 DART cases. You can plot a glide path to track your progress month by month:

• Baseline: 9 cases
  
  
• Month 1 Target: 8 cases
  
  
• Month 2 Target: 7 cases
  
  
• Month 3 Target: 7 cases (maintaining the new rate)

The glide path gives leadership a visible line to follow and helps teams turn a percentage target into a concrete case count.

‍

Enable Rules Where Risk Concentrates

Instead of trying to do everything at once, focus your computer vision rules on the areas with the highest risk. Start with the big ones like Personal Protective Equipment (PPE) compliance, people entering restricted areas, and vehicle-pedestrian interactions.

Map each rule to a specific location and shift, then fine-tune its sensitivity to catch repeat behaviors without creating a lot of noise. And make sure to document all changes so you have a clear audit trail.

‍

PPE Compliance Rule Packs

Activate PPE Detection for important gear like hard hats, high-visibility vests, and gloves. The system tracks compliance data by area and crew, helping you find clusters of non-compliance. These insights let you put targeted fixes in place, like coaching or better signs, before a lack of PPE leads to a recordable injury.

R

estricted-Area Incursion Rules

Use Area Control to create virtual fences around hazardous zones like machine perimeters, electrical panels, or robotic work cells. The platform monitors who goes in and out by time of day and shift, showing you the data on trend dashboards. These patterns often point to simple fixes, like adding physical barriers or changing pedestrian routes, that can quickly cut down on unauthorized entries.

‍

Vehicle Safety Interventions

Forklift incidents are a major cause of serious injuries. With Vehicle Control, you can track near misses between forklifts and pedestrians, identify speeding, and find blind-spot hazards.

This objective data helps you adjust traffic routes and staging areas to improve flow and reduce the close calls that often happen before big accidents. In some deployments, better route and fleet decisions have helped teams defer purchases worth an estimated $25,000 to $50,000 per forklift.

‍

Camera Coverage and Zones

A clear rollout starts with camera and zone setup. Follow these steps to confirm coverage, tag feeds, and validate detection quality so the system provides accurate, actionable data.

• Verify existing AI cameras have clear views and consistent lighting, then tag each feed to a named zone and assign it to the site shift schedule inside the Protex platform.
  
  
• Run short detection tests and quick feedback loops to check detection quality, fine-tune camera angles and rule sensitivity to reduce false positives, and log any configuration changes so you can trace when and why rules were adjusted.

‍

Security and Privacy

Meeting information security and privacy expectations is essential for video-based systems. Here is a summary of common controls and compliance steps to keep your technical setup aligned with company policies.

• Process video on-site using edge computing so raw footage never leaves your facility, and send only anonymized metadata to the cloud for centralized analysis and reporting. This reduces exposure, supports your internal information-security policy, and enables trend detection and cross-site comparisons without exposing individual identities.
  
  
• Design the solution to align with security standards such as ISO 27001 and comply with privacy regulations such as GDPR, as outlined in our security and privacy overview, and keep audit logs and access controls ready for compliance checks or third-party reviews.

‍

Weekly Triage of Repeaters and Root Causes

Gather core EHS, operations, and maintenance staff for a quick 30-minute meeting each week. The goal is to review a short list of the top repeat behaviors, hotspots, and rising trends the platform has found. Use this meeting to assign fixes with clear owners and deadlines.

‍

Hotspot and Repeater Lists

To keep everyone focused and build momentum, filter your data to show the top 5 to 7 repeat issues. This might be a specific walkway where people aren't using the designated path or a work cell where PPE compliance is always low on a certain shift.

‍

Near-Miss Visibility Feed

Review a feed of near-miss events, complete with timestamps, locations, and short video clips. This hard evidence helps the team spot patterns that line up with later recordable incidents. Seeing these early events allows you to prioritize fixes to facility layout, work schedules, or standard operating procedures.

‍

Corrective Action Ownership

For every issue you find, assign a person to own the fix, set a firm due date, and decide on a clear way to check that it's done. Tracking how long these actions take and how well they work keeps everyone accountable and prevents important fixes from falling through the cracks.

‍

Toolbox Talks with Video

Use 30 to 90-second video clips of near misses or unsafe behaviors in your pre-shift huddles. This visual proof makes hazards feel real and helps reinforce safer ways of working. It's also a chance to give a shout-out to teams that have successfully fixed repeat issues.

‍

Escalation and SLA Cadence

Set service-level agreements (SLAs) for closing out high-priority issues and create a formal way to escalate actions that aren't done on time. Put out a weekly scorecard showing on-time closures and overdue items to keep things visible and drive accountability. The Protex.ai Reporting & Workflows feature make this process simple.

‍

Month-End DART and LTIR Glide Path

At the end of each month, it’s time for a review that you can take straight to leadership. Compare your actual DART and LTIR performance to the glide path you set up. Look at your case count against the plan and check the severity of any incidents that happened. Turn any gaps into specific action items for the next 30 days and estimate the impact they will have on case counts and rates.

‍

Glide Path Math

Show how the actual hours worked changed your rates and figure out the case difference needed to get back on track. For example, if you had one more case than planned, turn that gap into a clear target for preventing one additional recordable incident next month.

‍

Severity Mix Analysis

Track the full range of safety events, from near misses and first aid cases to restricted duty and lost time injuries, and point out any shifts toward less serious outcomes. And celebrate the wins where repeat behaviors were stopped before they could lead to an injury.

‍

Line-Zone-Shift Variance

Compare performance across different lines, areas, and shifts to see who's doing great and which spots still need work, and use those insights to rebalance resources and adjust your rule coverage toward the areas that need it most. You can find success stories from high-performing areas in our case studies.

‍

90-Day Progress Dashboard

Present your progress on a single dashboard that gives leadership a clear view of how the safety program is performing. This view should include your main metrics, filters, and the ability to drill down into the underlying data.

‍

Single-View KPI Board

Your dashboard should have a baseline-versus-target chart with a glide-path overlay for DART and LTIR. List the following KPIs and show their weekly movement to keep teams engaged. Teams often pair the KPI board with predictive safety analytics to forecast incident risk and staffing or process needs:

• DART and LTIR Rates
  
  
• Case Counts (First Aid, Recordable, DART, Lost Time)
  
  
• Near-Miss Volume
  
  
• PPE Compliance Rate
  
  
• Restricted-Area Incursion Rate
  
  
• Vehicle Interactions
  
  
• Open vs. Closed Corrective Actions
  
  

‍

Line-Zone-Shift Breakouts

Allow for quick pivots by area and shift to guide decisions on staffing, process changes, or schedule adjustments. Use the dashboard to highlight the top three improvements made in the last month to build positive momentum.

‍

Audit Trail and Exports

The system automatically records who took action, when, and how. Provide CSV and PDF exports from the reporting dashboards to support OSHA logs, insurance reviews, and internal audits. Those exports replace a lot of manual spreadsheet work, which for many teams runs to tens of hours a month.

‍

FAQs: DART, LTIR, Near Misses

When you see events as they happen instead of weeks later in a report, it changes how fast you can react. These FAQs explain how that visibility ties into your DART, LTIR, and near-miss performance.

‍

What is a good DART benchmark?

A good DART rate really depends on your industry and how complex your work is. Instead of getting hung up on a single number, compare your performance against your industry’s average and your own past results. The main goal is to see a steady downward trend.

‍

How fast can DART and LTIR improve?

You'll get visibility into unsafe behaviors and near misses almost immediately. Many teams start to see changes in leading indicators (like unsafe behaviors and near misses) within the first few weeks as fixes go in. Depending on your hours worked and baseline, that early shift can translate into measurable movement in DART and LTIR by the second or third month, but timelines will vary by site.

‍

How do near misses relate to recordables?

An increase in detected near misses is actually a good sign. It means your detection is getting better, not that your site has become less safe. This improved visibility lets you fix underlying problems before they result in recordable injuries.

‍

What is the difference between TRIR, DART, and LTIR?

These are all ways of looking back at incident frequency per 200,000 hours worked.

• TRIR (Total Recordable Incident Rate): This is the big one. It includes all work-related injuries and illnesses that need medical treatment beyond first aid.
  
  
• DART Rate: This is a smaller group within TRIR. It only includes cases that result in Days Away, Restriction, or Transfer.
  
  
• LTIR (Lost Time Injury Rate): This is an even smaller group within DART. It only includes cases that result in days away from work.

For more on how OSHA and the BLS define and track these metrics in current data, see our latest health and safety stats report.

‍

Get the 90-Day DART/LTIR Playbook

Ready to build your own 90-day plan? Download our complete playbook. It comes with checklists, sample dashboards, triage meeting agendas, and KPI calculation spreadsheets.

‍

Request the Playbook

Fill out a short form to get the full PDF and spreadsheet pack. We also offer a quick-start call to help with site mapping and rule calibration.

‍

See the Unified Dashboard

Schedule a live demo to see how PPE Detection, Area Control, and Vehicle Control data all come together in our unified Reporting & Workflows dashboard. To learn more about the platform and approach, visit Protex AI.

‍