Measuring RTW outcomes in Workers’ Compensation: Part 2 – Jurisdictional Approaches

In Part 1, I highlighted three different approaches to measuring return-to-work outcomes:

1.     Standardized calculation based on standardized claims payment data at specific milestone submitted by jurisdictions (AWCBC approach),

2.     Retrospective survey of workers by central researcher group using a stratified sample from each jurisdiction withing strict time frames (Safework Australia),

3.     Retrospective interview approach normalized for each participating jurisdiction (WCRI)

 

Each approach provides insights for stakeholders and policy makers. Comparability and consistency among participating jurisdictions are often primary objectives. 

 

Individual jurisdictions have different objectives in measuring their RTW outcomes, each tuned to the features, priorities, and demands of their jurisdiction. 

 

With that in mind, here are four examples of jurisdictional RTW outcome measures.

 

Texas: 2023 Return to Work

 

The Texas Department of Insurance (TDI), Division of Workers’ Compensation, Workers’ Compensation Research and Evaluation Group (WCResearch) analyzed RTW outcomes for claims between 2007 and 2020 in this report. The report focuses on initial and sustained RTW (defined as a return to work and staying at work for three calendar quarters (9 months) following injury).

 

I asked WCResearch at TDI to expand on the study methodology:

 

The return-to-work report uses data on tens of thousands of employees who received Temporary Income Benefits (TIBs) after a work-related injury, examining wage records from the Texas Workforce Commission to determine when these individuals began earning wages again. The report assesses whether this return was classified as initial, referring to the first time an employee returns to work after the injury, or sustained, defined as maintaining employment for three consecutive quarters. These findings are linked to demographic details such as age, industry, employer size, and benefit type allowing for further statistical analysis. The methodology also includes an analysis of average wages before and after the injury, along with the average number of days employees spent away from work.

 

The results over the study period include the following:

 

92% of employees were back at work within one year.

 

69% of workers back at work within six months stayed at work and achieved pre-injury earnings within two years.

 

Days away from work averaged 43 (median 29).

 

For 2020 injuries, the initial RTW rate at six months post injury was 83%, a bit higher than the average 80% for the 2007-2020 study period. The sustained RTW rate was 69%, higher than the average 63% for the study period.

 

The “sustained” RTW outcome in this report is striking.  The three quarters (9 months) of earnings post RTW is a significant threshold and the objective sources for this data provide a rigour not found in many studies that rely on interview responses.

One challenge with this study approach is the lag necessary for a year’s cohort of claims to fully develop (close) to their ultimate duration. While that is happening, changes to law, policy, practice, and economic conditions may impact current RTW results.

 

Victoria Australia: WorkSafe Victoria RTW

 

WorkSafe Victoria has made return to work a “headline” Key Performance Indicator (KPI) in their Corporate Plan 2023-2024 and Strategy 2021-2024. (https://www.worksafe.vic.gov.au/resources/corporate-plan).  “At a headline level, we measure safe and sustainable return to work outcomes for injured workers with physical and mental injuries at 16 weeks, 26 weeks, 52 weeks and 104 weeks post claim lodgment. A return to work is defined as one that has been sustained for a three-week period.”

 

In Victoria, the claims management function is delivered by third party “agents” with WorkSafe Victoria oversight. Remuneration of agents includes a performance adjustment for meeting targets set for RTW outcomes at 26 weeks. Claims management software requires case managers code RTW status upon case closure and provide for case notes or other documentation confirming RTW details.

 

The following is my summary of WorkSafeVicotria’s methodology for the26 week (182 day):

1.     Calculation:
      Number of injured workers working at 182 days [divided by]                                                                                                     
     Population (paid claims in reporting period with > 10 days weekly compensation paid (including employer excess))

2.     RTW Status Assessment made 182 days post injury.

a.     Assessment of whether the worker has continued to remain at work for three weeks (21 days) following the return to work.

b.     Based on recorded fielded responses in systems (ACCtion, Fineos) and documentation of RTW (email, letter, Novus note, …).

3.     RTW Status Assessment and Validation at 21 days are subjected to random sample audit.

 

Note the exclusion of the 10 days of weekly compensation) in this calculation. Shorter duration absence in this range typically resolves with 100% RTW. Eliminating these from the calculation denominator and the numerator will tend to lower the overall RTW rate than if these cases were included.

 

WorkSafe Victoria’s data relies on the integrity of the information encoded in claims management systems about initial and continued employment at three weeks.  Audits of data entries to confirm the integrity of the coded information is a critical component of this measure.

 

A key feature WorkSafe Victoria’s RTW outcome reporting is the separate tracking of physical and mental injuries.  Mental injury cases tend to have longer recoveries. If a jurisdiction has a larger proportion of mental injury claims, a combined RTW measure will result in a lower percentage of RTW at 26 weeks.

 

For physical injuries, the result at 26 weeks is 73.60%, but for mental injuries, the return-to-work rate was 41.60%. This highlights a significant difference between mental injury cases and other injury types.

 

WorkSafeBC: Key Performance Indicator

 

WorkSafeBC is the operating name for the Workers’ Compensation Board of British Columbia. Its operational priorities include “Maximize overall recovery and post-injury earnings for injured workers”. A key performance indicator (KPI) to track progress against this priority objective is included in their Annual Report and Service Plan.

 

The KPI, “Improve return-to-work outcomes” carries the description: “The percentage of B.C. workers who return to work within six months (26 weeks) of their work-related injury.” The calculation method as stated in the 2024 Annual Report and 2025-2027 Service Plan (p. 34) is described as follows: 

 

“This KPI measures duration (by 26 weeks), whether the return to work is voluntary (the worker does not object), and whether it is safe and durable, with no subsequent inability to work for 30 days. To track the percentage, we compare the number of claims meeting these criteria with the total number of wage-loss claims open for more than 26 weeks.” [emphasis added].

 

Fielded case management data are used to determine status on closure. The absence of a claim re-opening indicating an inability to work infers durability.  There is no explicit confirmation of sustained RTW or return to full earnings.  

 

The KPI results reached a high of 81.3% in 2022 before falling to 78.5% in 2023 and rising again to 79.7% in 2024. The target going forward is to achieve 81.0% return to work by 26 weeks. (Note:  AWCBC records 85.00% (2022) and 83.62% (2023) under 25.5-Percentage of Wage-Loss Claims off Wage-Loss Benefits at 180 days (%)

 

There is no waiting period in British Columbia’s workers’ compensation system, so the results capture many very short duration claims.  These cases tend to have very high successful RTW outcomes.  Self-insurance is allowed in BC but not self-administration, so the results include for these deposit-class employers (governments, certain large employers) are also reflected in this measure.

 

South Australia: return-to-work milestones

 

When you rebrand your workers’ compensation authority “ReturnToWorkSA”, the messaging is clear. As you might expect, RTWSA has measures that reflect RTW status at various milestones such as 4, 12, 26, 52, and 78 weeks.

 

The data-rich report on ReturnToWorkSA Insurer Statistics FY2024 provides RTW outcomes for the 26-week milestone, similar to others noted in this part.  This time, there are three RTW status outcomes noted: “Not at work”, “Partially at work”, and “Fully at work”.  This reporting reflects a level of importance on “stay at work” outcomes within the system.  

 

The four-year time series provided for this milestone reflects a trend toward fewer workers not working and an increase in the number and proportion of workers fully RTW:

 

Not at work 5.7%,

Partially at work 3.5%,

Fully at work 90.9%. 

 

This is a population-based study reporting on 12,121 cases in the 2024 result.  According to RTWSA:

 “The milestone is measured from the workers first date of work incapacity, or injury date if the worker does not have work incapacity. Improvement has been achieved with a multifaceted program of work aimed at improving services to injured workers and employers.”

 

Note the subtlety of a worker suffering a work injury but not a work incapacity. The accepted claim covers the medical costs, but the worker may continue working on modified duties as part of a stay-at-work program, an approach actively promoted by RTWSA.

 

Summary Comment

 

Jurisdictional approaches to measuring RTW outcomes are often specifically crafted to address particular needs and priorities.  Where the national measures must find commonality among the jurisdictions, individual jurisdictions can allow the unique data systems and to provide greater depth.  Each approach starts with a purpose, and each design has its limitations.  In general, the results highlighted by the measure lag significantly behind changes in policy or initiatives to improve outcomes. 

 

Texas can focus on sustainable RTW outcomes, WorkSafe Victoria can differentiate results for mental and physical injury cases, and ReturnToWork South Australia can provide data on partial RTW and include successful stay at work outcomes that might be excluded from other designs.

 

Note the different approaches in determining sustainability.  In Part 1, we saw Safework Australia and WCRI ask workers directly about their work status as the time of interview. Their retrospective methodology builds in significant distance from initial RTW to assess sustainability of the RTW.  In Part 2, we see WorkSafeBC uses a case criterion “with no subsequent inability to work for 30 days” to exclude non-durable RTW outcomes from its measure, while WorkSafeVictoria uses an audited 21 day record (or direct/indirect inquiry) to assess whether the worker has continued to remain at work.

 

Again, there is no one “right way” to measure RTW outcomes.  Those jurisdictions willing to participate in studies, establish well defined measures, and post transparent results allow leadership, stakeholders, and policy makers assess performance and achieve better outcomes for injured workers and their families. 

In Part 3, we will look at RTW outcome trajectories. 

 

[This post was prepared as a resource for DMCCT- Evaluating DM Programs & Assessing RTW Processes, Pacific Coast University for Workplace Health Sciences]

 

What can we take from the ProPublica/NPR investigative reports on Workers' Compensation?

The ProPublica/NPR investigative reports have highlighted what is wrong with workers’ compensation.  Make no mistake; there is much to criticize among the US and Canadian workers’ compensation systems.  Does that mean we should throw out the current systems and start again?

Marjorie Baldwin, (Professor, Arizona State University, and Chair of the Study Panel on Workers’ Compensation Data of the National Academy of Social Insurance (NASI.org)) recently responded to the main issues highlighted by this investigation.  Her post, “Workers’ Compensation:  Critical Questions, Elusive Answers” addresses some of the obvious issues.  The journalistic approach of focusing on individuals to illustrate the issues effectively shines a light on vivid examples of poor benefits, bad adjudication and abusive processes that re-victimize the victims of work-related injuries, illness, and disease.  The scholarly examination of underlying policies at the root of these failures may not grab headlines but it is critical to public policy development.  Headlines don’t tell the whole story.  Professor Baldwin’s point that stakeholders need “a more informative accounting of how the system performs” succinctly summarizes both what is needed and what has been missing from much of the discussion.

It is not that there isn’t good information out there.  The NASI report on Workers’ Compensation:  Benefits, Coverage and Costs 2012 provides a starting point.  The AWCBC Key Statistical Measures provides similar data for the Canadian workers’ compensation jurisdictions.  The work by IAIABC and WCRI to provide objective data on the Workers’ Compensation Laws that ultimately determine the benefits, costs and coverage on both sides of the boarder is another important information resource. 

And it’s not as if there is no objective yardstick on what a workers’ compensation system ought to do.  The 1972 Report of the National Commission on State Workmen’s Compensation Laws made recommendations that provide clear guidance.  ProPublica/NPR journalists the National Commission’s recommendations to design workers’ compensation laws; public policy analysts in Canada, the US and other countries often take the measure of workers’ compensation systems using the National Commission’s recommendations. 

Objective assessment of systems’ performance against those recommendations reveals two things.  The first is the point of the ProPublica/NPR reports:  Worker’s compensation is failing in some states.  The second point is really the corollary.  Despite the poor performance of some jurisdictions, there are workers’ compensation systems that are providing benefits that meet or exceed most of the recommendations of the National Commission. 

Workers’ compensation is not one “system”.  There are more than sixty North American jurisdictional attempts at fulfilling a common social policy objective that is the foundation of the Grand Bargain, the Historic Compromise.  It is plainly wrong to extrapolate grievous failings from a few jurisdictions to every workers’ compensation system. 

Yes, there are failures.  Workers’ were promised compensation for work-related injuries but there are jurisdictions where between a third and a half of all workers with lost-time work-place injuries are entitled to no compensation for lost wages—and that does not take into account the issue of claims suppression and under-reporting.  Large proportions of the labour force—particularly agricultural workers and domestics—are excluded from coverage in some jurisdictions.  Middle-to-high wage earners may have less than half their earnings unprotected by workers’ compensation insurance in states/provinces with low maximum benefits and very low indemnity rates.  These inequities are not only unjust, they undermine the social contract and threaten the social policy (and possibly legal) basis of the “exclusive remedy”. 

The ProPublica/NPR reports force policy makers to acknowledge these failures and hopefully seek out those jurisdictions that live up to the bargain.  Those jurisdictions that come closest to meeting the National Commission recommendations cover nearly everyone who works for someone else and even offer coverage to those who are self-employed; they cover high wage earners and provide compensation that restores 80-90% of spendable (after tax) income.  They provide timely decisions and are accountable for their errors in the application of law and interpretation of policy.  They seek and earn a measure of social licence for what they do and do it at a cost that is affordable and sustainable.

H. James Harrington (author of Business Process Improvement among others) said:

Measurement is the first step that leads to control and eventually to improvement.
If you can't measure something, you can't understand it.
If you can't understand it, you can't control it.
If you can't control it, you can't improve it.

NASI, WCRI, IAIABC, AWCBC and others provide objective measurement of jurisdictional and national performance.  The National Commission recommendations provide a standard against which the measures from each jurisdiction may be assessed and understood.  Measuring the performance of each system against those recommendations can be the first step in addressing the failures, controlling the excesses and improving outcomes for injured workers and their families without a wholesale scrapping of all systems. 

No system is perfect.  Among the sixty-plus systems in North America, however, there are a few that have come close to meeting the recommendations of the National Commission.  They are proof that the Grand Bargain, the Historic Compromise can achieve the social policy objective: to protect workers from work-related injury, disability, illness and death in a compassionate and sustainable way that still allows the economic activity and innovation necessary for societies to operate and thrive. 

Improvement is not only possible, it is essential—not only because it is the morally correct thing to do but also because every failure erodes the public confidence in all workers’ compensation systems everywhere.

Rather than taking a defensive posture, insurers and policy makers can thank the ProPublica/NPR journalists for raising the level of discourse, highlighting the disparities that exist and illustrating the need for genuine improvement in under-performing systems.

What does a rising trend in injury costs mean?

The costs of work-related injuries are felt by individuals, families and communities.  The personal and human costs are immense but difficult to quantify.  Injury costs associated with workers’ compensation claims, however, are financial and easily quantifiable. 

Changes in injury claim costs get the attention of employers and workers’ compensation administrators.  The obvious concern is that rising claims costs might be an indicator that claims managers are “giving away the farm”, not managing the claims very well.  I have rarely found this to be the root cause of such a trend.

Rising injury claim costs can be related to external factors.  Healthcare cost inflation is outstripping overall inflation, for example; even if all else remains the same, this factor alone may account for increased claims costs.  The external employment environment may also be a factor.  We know from research that claim duration increases as the employment moves from high demand to steady and from steady to declining demand (given relatively constant labour force supply). 

Occasionally injury claim costs rise because of closer adjudicative scrutiny.  I recall one claims director promoting very early intervention and a medical report every two weeks before a payment could be issued.  The result was increased medical tests and physician visits, which added significantly to costs and did not result in shorter duration. 

I was consulting with one large employer who was considering terminating his third party administrator (TPA) because indemnity and medical costs per claim were on the rise.  As we discussed the situation, it became clear that demographics not claim management practices were the main driver.  As this very large firm had automated and improved efficiency, it had hired fewer workers.  Mandatory retirement had also been eliminated so more workers were working longer (automation actually facilitated a longer work career).  Taken together, the result was an aging workforce and one with a greater level of co-morbidity (diabetes, high blood pressure, obesity).  While injury rates were lower, duration and medical costs were higher every year.  Clearly, these issues—not the TPA’s claims management practices—were driving claim costs upward. 

Recently, I was corresponding with a coordinator for a booming resource extraction firm where recruitment and retention were big issues.  The employer made a point of meeting with all the crews, raising awareness about workers’ comp and disability benefits as well as their extensive EFAP program and easy, on-line extended benefit program.  The result was increased utilization in WC, EFAP, LTD and extended health driving premiums higher.  This was both expected and welcomed because it contributed to a more important corporate objective:  increased employee retention and lower turnover.  The higher premium costs were much less than the cost of recruiting and training of new employees. 

Of course, a trend toward higher injury claim costs may also be related to internal factors.  Increasing caseloads, for example, may be caused by staffing reductions or increased claim volume with insufficient increases in staffing; either way, the result may be less attention per claim and rising injury claim costs.  Other internal changes to systems, policies and procedures may also contribute to higher costs.  One insurer back-filled a significant number of case manager positions with less experienced staff so the more senior staff could be assigned to a particular crisis situation.  At the best of times, the handover of a particular caseload from one case manager to another may contribute to longer claim duration and increased utilization as the new case manager becomes accustomed to the caseload.   When amplified by the simultaneous transfer of many caseloads, the impact on performance measures such as claim cost can be significant.

Whether you are looking at the performance of an overall system or a specific firm, rising injury claim costs cannot be viewed in isolation.  Its meaning must be interpreted within a broader context. 

How do you develop leading indicators for occupational safety and health?

Most of us are familiar with the concepts of “lagging” (sometimes referred to as “trailing”) and “leading” indicators from the world of economics. GDP and “average duration of unemployment claims” tell us about where we have been and, therefore, are generally considered lagging indicators of the relative health of the economy. Housing starts and permits are great examples of leading indicators. If these are rising, the demand for labour and supplies to build the new housing units is likely to rise in the near future. As the housing units are completed, demand for consumer goods like furnishings to fill them is also likely to rise.



The power of leading indicators is obvious to those gauging current conditions and making plans. If housing starts are rising, retailers of furniture and appliances are more likely to increase orders and hire new staff; manufacturers are likely increase production and inventories in anticipation of rising demand.


In workers’ compensation and OH&S, traditional measures tend to be lagging indicators. Injury rates, injury counts, and “days injury free” are, at best, lagging indicators of safety—they may tell us something about where we were but little (if anything) about where we are going. These measures are heavily weighted to the past and may mask serious safety and health risks in the current workplace. Developing leading indicators at the operational, sectoral and even jurisdictional levels helps focus resources and attention where it is most needed and provides early signals of the effectiveness of current programs or initiatives.


To design a leading indicator, you need a logic model, a framework that takes into account the near-term, mid-term and long-term objectives that will lead you to your goal.


Suppose your goal is a safer, healthier workplace and you have an objective of reducing strain injuries in your manufacturing plant. You might want to start by identifying the factors that lead to these injuries. Ergonomics is an obvious factor but you could get more granular or more general in your consideration. Loads, repetitions, and workstation design might be factors at the individual level while work procedures, the pace of work, and safety culture might be important factors at the operational or corporate levels.


Now that you have a model of how the injuries occur, you can think about interventions at the causative level that will contribute to greater prevention. Perhaps you have been convinced as I have that safety culture is vitally important and you have initiatives to improve safety culture in your operation. Annual external audits or random quarterly surveys could help you determine both the current climate and trend over time. If your model is correct, improvements in your safety culture will lead to outcomes like improved adherence to safe work procedures, more safety-oriented content in supervisor-worker interactions, more rapid time from hazard identification to removal—all of which have been proven to reduce injuries and make workplaces safer and healthier.


Other examples of leading indicator metrics for the objective of reducing strain injuries I’ve come across in industry include:

• % of workstation ergonomic evaluations completed

• % of employees/supervisors trained in ergonomics

• % of ergonomic action items addressed

• % of employees engaged in fitness and wellness program


Developing a logic model and selecting a leading indicator forces you to understand your business, how injuries occur and what research tells us will prevent them. That understanding is critical for good management as well as OH&S.



Don’t bother developing logic models, selecting leading indicators, and continually measuring indicators if you think it will be easy. Making the time and effort is hard but worthwhile.



My favourite quote on this topic makes the point very well:

"Measurement is the first step that leads to control and eventually to improvement.

If you can't measure something, you can't understand it.

If you can't understand it, you can't control it.

If you can't control it, you can't improve it.”


- H. James Harrington (Author, columnist, a Fellow of the British Quality Control Organization and the American Society for Quality Control).