It Wasn’t the Decision
Where the harm in a compensation claim actually sits, and what follows for the way we work
In this issue
- Being told no wasn’t worse than being told yes. Being told nothing was worse than both.
- Delay in approving a claim, and negative contact with the insurer, outrank injury severity in producing a bad claim experience.
- How readable a scheme is matters more than how capable the claimant is.
- At three to four weeks, whether anyone had actually offered suitable duties predicted claim cost, claim duration and chronic work disability.
- Prompting clinicians to talk more about work changed nothing at all.
I’ve worked in rehabilitation for more than thirty years and I’ve sat with hundreds of injured workers, and the thing I’ve never found a way to convey to people outside this field is how far the harm travels. I’ve had clients who stopped leaving the house, who were frightened of being seen outside. I worked with a grandmother who became so unwell over the course of her claim that she locked herself in her bedroom to avoid all contact with her own children and grandchildren. I’ve sat with people doubled over in tears, trying to explain what the process was doing to them and not able to find the words for it. They get stuck, and some of them don’t recover.
Parts of that I could always point to. I can still count on one hand the number of injured workers who told me their case manager was a good one, someone who genuinely worked with them rather than to their own agenda, and rarer still is the worker who had the same case manager the whole way through. Some of my clients have had nine, ten, eleven across a single injury, telling their story again each time to someone who’s read the file but doesn’t know them. What I could never say with any precision was which parts of that process were doing the damage and which were merely unpleasant.
I have my own reference point for this. When I had cancer, the hardest part wasn’t the diagnosis, and it wasn’t the treatment plan. Both were frightening, but they were solid, and you can take hold of something solid. Nor was it the communication, which was clear. What I found nearly unbearable was the nothingness of the waiting in between, the grey zone where nobody could tell me anything yet, because there’s nothing to work with in that zone and the mind fills it in on your behalf. I’ve wondered for a long time whether that’s what a compensation system does to people, month after month, and a run of papers published since December 2025 suggests it might be. If that’s right, one of the most powerful things a scheme could do for recovery has very little to do with treatment, and a great deal to do with how long people are left not knowing.
What researchers now call system-generated mental ill health
Workers who reported negative interactions with their insurer had more than five times the odds of describing their compensation claim as a bad experience, and workers left waiting on approval had more than four times the odds. Neither injury severity nor diagnosis explained who ended up there. Those numbers come from a study Alex Collie’s group published in January 2026, which surveyed 533 injured workers about the events and interactions that had made a claim a bad experience and then interviewed twenty of them at length.
The system produces a second injury, and the authors say so in the paper. They set out, in their words, to identify “experiences, events and interactions contributing to system-generated mental ill health”, and they conclude that fixing the modifiable parts of the claim process might “reduce the risk of secondary psychological injury”. Those are horrifying phrases to find in a rehabilitation journal, because what they describe is a system built to help injured people generating injuries of its own. Those of us who’ve been around a while have watched that happen for years, and what’s different now is that we can point to the data instead of the anecdote.
Both of those predictors sit entirely within the scheme’s control, which is an uncomfortable thing to read if you work inside one. What predicted the harm was how long people were kept waiting, and how they were spoken to while they waited.
The study behind this
STUDY Sanatkar and colleagues, Australia, published online 21 January 2026
DESIGN Mixed methods. Online survey of 533 injured workers in Australia between September 2023 and July 2024, recruited through community organisations, worker support groups and social media, plus follow-up interviews with 20 of them. 503 of the 533 reported negative mental health impacts, though that recruitment route means the sample is self-selected and will over-represent difficult claims
FINDING Negative interactions with the insurer, odds ratio 5.22 (95% CI 1.89 to 14.42). Delay in claim approval, odds ratio 4.12 (95% CI 1.88 to 8.99). The full model accounted for 40 per cent of the variation in claim experiences. Neither predictor is a property of the injury
SOURCE Journal of Occupational Rehabilitation, advance online publication. doi.org/10.1007/s10926-026-10364-0
Being told nothing was worse than being told no
The harm concentrates in the phase where nothing has been decided. At the end of June 2026, Vera Camões-Costa and colleagues, working inside the IMPRovE trial, measured injured workers’ mental health every three months and matched each measurement to where the claim had actually got to at that moment, which let them compare four groups: accepted, rejected, still under assessment, and no claim lodged at all.
The people waiting on a decision were worse off than all three of the others. Set against workers whose claims had been accepted, they scored almost six points higher on stress. Set against workers whose claims had been rejected, almost six points higher on depression. And set against workers who had never lodged a claim at all, six points worse on overall mental health
The study behind this
STUDY Camões-Costa and colleagues, Australia, published online 28 June 2026
DESIGN Cohort study nested in the IMPRovE trial. 153 general practice patients with work-related mental health conditions, 422 observations, three-monthly DASS-21 and quality of life surveys matched to claim status
FINDING Claim under assessment, compared with accepted (stress, beta = 5.73, 95% CI 1.69 to 9.77), with rejected (depression, beta = 5.88, 95% CI 1.12 to 10.63), and with no claim lodged (overall mental health, beta = 6.14, 95% CI 2.87 to 9.42). No significant difference between the accepted, rejected and no-claim groups
SOURCE Journal of Occupational Rehabilitation, advance online publication. doi.org/10.1007/s10926-026-10422-7
Accepted, rejected and no claim didn’t differ from one another at all, and that’s the comparison worth sitting with. Being told no was no worse than being told yes and being told nothing was worse than both. The authors are careful to say the finding is exploratory and can’t establish causality, and they’re right to be careful about it, but it matches the Collie data, and it matches what I remember of waiting.
"Being told no was no worse than being told yes.
Being told nothing was worse than both."
Being told nothing was worse than both."
REHAB REIMAGINED | DR. CAROLINE HOWE
So what is it about the waiting that does the damage?
People can’t read the system they’re waiting inside, and that turns out to matter more than anything about the person doing the waiting. A Dutch study from March 2026 surveyed 168 people applying for work disability benefits and measured what its authors call social insurance literacy across four domains. Three belong to the individual: the ability to obtain information, to understand it, and to act on it. The fourth belongs to the scheme, and it’s simply how comprehensible the system is.
System comprehensibility was by a clear margin the strongest thing in the model (justice B = 0.72; fairness B = 0.57), while the claimant’s own ability to obtain or to understand information wasn’t associated with perceived justice or fairness at all. Only two of the four domains were associated with either outcome. We’ve spent years producing fact sheets and information packs on the assumption that the problem is what the worker knows, and this data suggests the problem is what the system is like to read.
At week four, the useful question is whether anyone has made an offer
Three to four weeks into a back injury claim, only three things predicted claim cost, claim duration and chronic work disability, and one of them was whether anyone had actually offered an accommodation. The other two were age and disability level. That comes from a Canadian study out of Saskatchewan published in April 2026, and if the waiting is the exposure, then anything that shortens it or fills it starts to look like treatment.
Accommodation is the North American term for what we’d call suitable duties: reduced or modified hours, a change to the tasks or the equipment, or a graduated return plan, so that someone can keep working or come back sooner. It sits close to what the Disability Discrimination Act calls reasonable adjustments, though in a compensation claim it usually arrives as suitable duties worked out between the employer, the treating doctor and the insurer. What Bruno and Passmore measured was whether an accommodation or an early return to work program had been offered, so the variable is simply whether anyone had put a realistic way back to work in front of the person.
We can’t change age and we can’t change disability level. Whether an offer has been made by week four we absolutely can, and Paul Bruno and Steven Passmore say so directly, describing it as a modifiable risk factor available to insurers, employers and workers alike.
The study behind this
STUDY Bruno and Passmore, Saskatchewan, Canada, published online 12 April 2026
DESIGN Prospective study. Questionnaire at three to four weeks post-claim linked to workers’ compensation board cost and duration outcomes. 368 claimants, a 16.4 per cent response rate from 2,247 invited, which the authors note as a limitation
FINDING Age, disability level and whether an accommodation or early return to work program had been offered predicted all three outcomes. Recovery expectations additionally predicted claim duration
SOURCE Journal of Occupational Rehabilitation, advance online publication. doi.org/10.1007/s10926-026-10386-8
Employers under-accommodate because they can't capture the return
The reason is money, and specifically that a firm which pays to accommodate a worker often doesn’t get to keep the benefit. An American paper published in Econometrica in March 2026 gives the most convincing account of this I’ve read. Naoki Aizawa, Corina Mommaerts and Stephanie Rennane used twelve years of Oregon workers’ compensation records linked to state earnings data, covering roughly 71,500 disabling claims, and exploited a change in the wage subsidy Oregon pays employers who accommodate an injured worker.
Accommodation rates moved with the subsidy rate, and workers who received one had significantly higher employment and earnings a year later. Employers under-accommodate because workers leave, so the firm can’t capture the return on what it spends, and because imperfect experience rating blunts the premium signal that might otherwise make up the difference. Suitable duties behave like an investment with a market failure attached rather than like a compliance obligation, which means premium design and wage subsidy arrangements are suitable duties policy, whether we describe them that way or not.
Prompting clinicians to talk more about work changed nothing
Prompting both clinicians and patients to pay more attention to work produced no effect on anything, across 1,091 people and twelve months of national registry data. In December 2025, Ingvild Bardal and colleagues randomised Norwegian outpatients to a nudge, where both patient and clinician received a mapping of motivation for work, barriers to return and work environment, against a mapping of health factors only. No effect on sickness absence days, none on return-to-work rates, none on disability benefits.
In a service where clinicians already raise work, prompting them to raise it more adds nothing, which is the authors’ own reading and the useful one. Talking about work isn’t the active ingredient, and doing something about the delay and the offer may well be.
We've generalised the participatory approach further than the evidence goes
The participatory approach works for low back pain and isn’t supported for mental health claims, which is awkward, because most Australian schemes apply it to both. A Dutch meta-analysis from June 2026 pooled eight randomised trials of that approach, where a worker and supervisor sit down together, identify the obstacles and agree the solutions. The low back pain effect carries moderate certainty; for mental health and mixed complaints the certainty is very low, and the evidence doesn’t support effectiveness.
Early intervention can't be early if the claim starts late
Claims often don’t start when injuries do, which means early intervention timed from lodgement isn’t early at all. Alison Sim and colleagues interviewed nineteen Australian researchers, insurance staff and regulators in May 2026 about why evidence-based early intervention keeps failing to land in compensable settings, and the theme I keep returning to is the one they call invisible obstacles. Everything we design as early intervention starts its clock at lodgement. If the injury happened weeks or months before that, the grey zone has already been running for some time before the scheme is even aware the person exists.
Put time to decision in the report, and make the wait legible
Time to decision belongs in our reports as a clinical variable, sitting alongside the pain scores and the functional capacity, because on this evidence it behaves like one. A file that records eleven weeks under assessment is describing an exposure, not an administrative detail, and it deserves the same escalation we’d give a symptom score heading the wrong way. It costs nothing to start recording it.
Where we can’t shorten the wait, we can change what the wait is like, and Vervenne’s data says the lever is comprehensibility rather than the volume of information we push at people. A worker who can tell you what happens next, who decides it and roughly when, is in a different position from one who can’t, and that’s deliverable inside a conversation we’re already having.
What the handovers were actually doing
Each handover restarts the not knowing, and I think that’s what my clients with their eleven case managers were describing all along. I’d always understood the handovers as a continuity problem, and they are, but a new person also means a fresh reading of the file, another round of questions about whether you’re really as limited as you say you are, and another stretch of time in which nothing is settled. I’ve been counting the case managers for thirty years. I suspect I should have been counting the days.
The grandmother who locked herself away from her own children and grandchildren didn’t get there from her injury alone. She got there somewhere inside a process that nobody was measuring, and we have the evidence now to know which part of it to measure first.
Sources, with the date each paper first appeared online
Sanatkar, S., Pritchard, E., Callaway, L., Cruickshank, G., Grant, G., Godic, B., MacEachen, E., Thompson, J., Di Donato, M., Gray, S., & Collie, A. (2026). Factors associated with negative experiences and mental ill health during a workers’ compensation claim: A mixed methods study. Journal of Occupational Rehabilitation. Advance online publication, 21 January 2026. https://doi.org/10.1007/s10926-026-10364-0
Camões-Costa, V., Nolidin, K., Collie, A., Chakraborty, S., Kenardy, J., Brijnath, B., Mortimer, D., Enticott, J., & Mazza, D. (2026). Association between mental health of patients with work-related injury and their compensation claim status: Exploratory findings from the IMPRovE trial. Journal of Occupational Rehabilitation. Advance online publication, 28 June 2026. https://doi.org/10.1007/s10926-026-10422-7
Vervenne, D., Brouwer, S., Ziech, M., Coffeng, J., Krause, B., Engel, O., de Vries, H., & Abma, F. (2026). Associations between social insurance literacy and perceived justice and perceived fairness in a work disability claim setting. Journal of Occupational Rehabilitation. Advance online publication, 17 March 2026. https://doi.org/10.1007/s10926-026-10369-9
Bruno, P., & Passmore, S. (2026). Early prognostic factors for claim cost and claim duration following a work-related back injury in Saskatchewan, Canada. Journal of Occupational Rehabilitation. Advance online publication, 12 April 2026. https://doi.org/10.1007/s10926-026-10386-8
Aizawa, N., Mommaerts, C., & Rennane, S. (2026). Firm accommodation after workplace disability: Labor market impacts and implications for subsidy design. Econometrica, 94(2), 341-374. Published at issue, 1 March 2026. https://doi.org/10.3982/ECTA22565
Bardal, I., Aars, N. A. P., Trichet, L. O., Brandseth, O. L., Terjesen, C., Irgens, E., Hansen, B. O., Kristoffersen, A., Höper, A., Jenssen, O. R., Brinchmann, B., & Mykletun, A. (2025). Expanding the focus on work factors in an outpatient setting: Does a nudge of patients and clinicians have an effect on return-to-work and benefits? Findings from the NSAC Nudge multicentre randomised controlled trial. Journal of Occupational Rehabilitation. Advance online publication, 12 December 2025. https://doi.org/10.1007/s10926-025-10343-x
de Smalen, A. W., Maarleveld, J. M., Burchell, G., Huysmans, M. A., Juurlink, T. T., Beekman, A. T. F., Anema, J. R., & Schelvis, R. M. C. (2026). Effectiveness of the participatory approach on return-to-work of sick-listed workers: A systematic review and meta-analysis. Journal of Occupational Rehabilitation. Advance online publication, 19 June 2026. https://doi.org/10.1007/s10926-026-10410-x
Sim, A., McNeilage, A. G., Richardson, C., & Ashton-James, C. E. (2026). Barriers and facilitators to implementing early intervention programmes in compensable settings: A multi-stakeholder qualitative study in Australia. Journal of Occupational Rehabilitation. Advance online publication, 5 May 2026. https://doi.org/10.1007/s10926-026-10408-5
Join the conversation
Get real-time psychosocial risk strategies, legislative compliance breakdowns, and mature workforce insights delivered straight to your professional feed by Dr. Caroline Howe.