RTW Program KPIs: The Metrics That Actually Predict Claim Outcomes

The Quarterly Meeting Where the Spreadsheet Has No Answer

Picture the scene: quarterly safety committee meeting, and the RTW coordinator walks in with the usual numbers — 14 open cases, 9 closed this quarter, 3 modified-duty assignments waiting on physician sign-off. The safety director asks one question: "Which of these numbers tells us whether next year's experience modification rate goes up or down?"

The spreadsheet doesn't have an answer. Case count is an activity log. It tells you what a coordinator did — how many files got touched, how many letters went out — but it says nothing about whether the program is actually bending claim duration or claim cost. A program can close nine cases in a quarter and still be bleeding money if those closures took 60 days longer than they should have, or if three of the nine employees reinjured themselves within a month of returning.

This is the gap between activity metrics and predictive metrics, and most RTW programs — even well-run ones — have never made the distinction explicit. This article walks through the handful of return to work program metrics that actually correlate with claim duration, claim cost, and EMR movement, and how to start tracking them even if the current system of record is still a shared spreadsheet.

Time to Light-Duty Offer: The Metric That Front-Loads Everything Else

Of all the return to work program metrics worth tracking, time to light-duty offer is the one with the most leverage, because it sits upstream of almost everything else. It's measured simply: the number of business days between receipt of a physician's documented restrictions and the date a written modified-duty offer goes to the employee.

Washington's Activity Prescription Form requires six key fields, including the primary diagnosis, specifically so the employer has what it needs to build an offer without a second round of calls to the clinic. The form exists because the restriction-to-offer gap is where cases drift. Washington's Stay-at-Work reimbursement program will not pay for days the worker didn't actually perform approved light-duty work — only approved dates count, which means every day of delay between restriction receipt and a documented offer is a day that doesn't just risk the claim going longer, it's a day that's structurally ineligible for reimbursement even if the offer eventually gets made.

The downstream cost of delay shows up elsewhere. NCCI research has found that medical-only claims which convert to lost-time cost 40 times more at 30 months' maturity than medical-only claims that never convert. A slow or undocumented offer doesn't cause that conversion by itself, but it removes the one lever — a timely, physician-matched light-duty assignment — that keeps a medical-only claim from drifting into lost-time territory in the first place. Tracking time-to-offer as a standing metric, by claim and by average across the open caseload, turns an invisible delay into a visible number a safety committee can act on.

RTW Acceptance Rate: What an Undocumented Refusal Actually Costs

The second metric is the acceptance rate on written offers — the share that are accepted, refused, or countered with a modification request, tracked by quarter and by injury type. A program with a high refusal rate isn't necessarily failing; it might mean the physical-demands matching is bad, or it might mean refusals aren't being documented in a way that would hold up if challenged.

That second failure mode is the expensive one. Litigated indemnity claims average $62,279 compared to $4,347 for non-litigated claims — roughly 14 times higher — and temporary total disability duration runs nearly 12 times longer on litigated claims (117 days versus 10 days), according to Aon's LAMBDA claims research. An offer refusal that isn't documented with the same rigor as the offer itself is exactly the kind of gap that turns a routine claim into a litigated one. Tracking acceptance rate alongside documentation completeness — was the offer in writing, was the refusal recorded with a reason, is there a timestamped audit trail — turns a soft compliance concern into a number a risk manager can watch move quarter over quarter.

The product does not adjudicate claims, determine eligibility, provide legal or medical advice, or replace the treating physician's judgment. It is an operational documentation and tracking tool for the employer's program.

Re-Injury Rate: The Metric Most Programs Don't Track at All

Few spreadsheets have a column for this one, which is part of why it matters. Re-injury rate — the share of returned employees who sustain a new or recurring injury within a defined window of resuming modified duty — is a direct signal of whether the duty match was actually appropriate, not just approved on paper.

Injury type matters here. NCCI data show knee injury claims are twice as likely to convert from medical-only to lost-time as other claim types, which suggests some categories of injury are simply less forgiving of an imprecise duty match. A physical-demands matching process grounded in a real occupational taxonomy — rather than a coordinator's best guess at what "light duty" means for a given job — is the structural fix for this. That's the specific gap TransitionalDuty's duty-matching engine is being built to close: restrictions get matched against ONET's physical-demands data (ONET content is licensed CC BY 4.0 by the U.S. Department of Labor and requires attribution, not a public-domain dataset free of any sourcing obligation) across more than 900 occupations, rather than against a coordinator's memory of what a given job title usually involves. Tracking re-injury rate by body part and by job-match type is how a program finds out whether its matching logic is actually working, instead of assuming it is because no one complained.

Medical-Only to Lost-Time Conversion Rate: The Leading Indicator

If time-to-offer is the metric with the most leverage, conversion rate is the metric with the most predictive power, because it's the pivot point where most of the eventual cost gets decided. NCCI's claim-size data show that roughly 75% of lost-time claims land in the $1,000–$50,000 range, and about 80% of ultimate claim dollars across the system sit in the $10,000–$500,000 range — meaning the difference between a claim that stays medical-only and one that converts to lost-time is often the difference between the low end and the high end of that distribution.

Portfolio composition matters for forecasting this. In Illinois' 2022 composite claims report (covering 2021 claims data), 60.2% of opened claims were medical-only, with attorney representation present in 16.7% of opened claims. A program that knows its own medical-only share and tracks how many of those claims convert each quarter has a genuine early-warning indicator — one that moves months before the claim-cost line item does, and well before the EMR recalculation reflects it.

That lag matters for how these numbers get used. EMR is typically calculated from the prior three years of claims history, excluding the most recent completed policy year, and recalculated annually — a structure explained in more detail in our piece on how experience modification rates actually work. Conversion rate is one of the few return to work program metrics that shows up on a quarterly basis rather than a multi-year lag, which makes it useful as an interim signal while waiting for EMR to catch up.

Building a Return-to-Work Program Metrics Dashboard That Predicts Outcomes

None of these four numbers — time-to-offer, acceptance rate, re-injury rate, conversion rate — mean much in isolation. Together, tracked by quarter and benchmarked against the prior year, they form a return to work program metrics dashboard that a safety committee or finance team can actually use to anticipate claim-cost movement instead of just reporting on it after the fact.

A few practical notes on building that dashboard, whatever system it lives in:

• Benchmark against industry injury rates, not just internal history. Our guide to BLS and OSHA benchmarking for RTW programs covers how to pull the right comparison rates for a given industry code, so a 2.3 TRC rate doesn't get judged against a different sector's baseline.
• Track metrics per case, not just in aggregate, once caseloads grow. Averages hide the two or three cases quietly driving most of the cost. Our piece on managing concurrent RTW cases covers what breaks first once a coordinator is juggling more than a handful of open files at once.
• Connect the metrics to budget, not just to compliance. A rising conversion rate or a lengthening time-to-offer has a real cost; the RTW program ROI and cost-of-inaction breakdown walks through how to frame that cost for a budget conversation without needing a dollar figure invented for the occasion.

These are operational and tracking metrics, not legal or medical determinations — a program's acceptance rate doesn't establish whether a refusal was legally effective, and a conversion-rate trend doesn't determine whether a given claim is compensable. Those calls still belong to the physician, the employer, and the claims examiner. What the dashboard does is make the pattern visible early enough to act on it.

Where to Go From Here

If the current system for tracking any of this is a spreadsheet with a tab for "open cases" and nothing else, the fastest next step isn't new software — it's a framework for turning these four metrics into a budget conversation with real numbers attached to real decisions. The ROI calculator is a quick way to see how time-to-offer and conversion-rate trends translate into rough cost terms for a program's own caseload.

For a more structured version of that exercise — one built for building the actual budget case a finance committee will want to see — the RTW Program ROI & Budget Case Workbook walks through how to turn these same metrics into a one-page business case, with worksheets for time-to-offer, acceptance rate, re-injury rate, and conversion rate side by side with the budget ask they justify.