The evidence
You've probably been told the pain is in your head — that if you really wanted to get better, you would have by now. Researchers have studied that question for thirty years. Here is what they actually found, including the parts that don't help our case. Every fact and every quote on this page links to the study it came from.
The short version
People hurt at work do measurably worse after the same operation than people with the same injury on regular insurance. That finding has been repeated by different research teams, in different countries, across different surgeries de Moraes 2012, PLOS ONE Russo 2021, IJERPH Anderson 2015, The Spine Journal. It is not in dispute.
Why it happens is in dispute — and we're not going to pretend otherwise. Researchers have not proven the comp system causes the worse outcome Spearing 2012, J Clinical Epidemiology. What they have shown is that the claims process is a serious, measurable source of stress, and that stress tracks with worse recovery years later Grant & Studdert 2014, JAMA Psychiatry. That's a strong case. It isn't a closed one.
1. The same surgery, worse results
The clearest number in this literature comes from a 2012 analysis that pooled 20 studies and 2,608 patients who had orthopaedic surgery — spine, rotator cuff, carpal tunnel, fractures. It compared people on workers' comp to people who weren't de Moraes 2012, PLOS ONE.
"Among patients who undergo orthopaedic surgical procedures, those receiving Workers' Compensation experience a two-fold greater risk of a negative outcome."
That isn't just people saying they feel unhappy. When the same team pooled the studies that used hard measuring tools — the Oswestry Disability Index, DASH, SF-36, pain scales — the gap held de Moraes 2012, PLOS ONE.
It also isn't one team's odd result. It repeats:
- Lumbar spine surgery — comp patients were 1.79× as likely to report worse pain, 2.10× as likely to be dissatisfied, and 1.68× as likely to fail to return to work (26 studies, 2,668 patients) Russo 2021, IJERPH.
- A completely separate team, different studies, published in a stronger journal, found 2.12× the risk of an unsatisfactory spine surgery outcome. Two independent groups landed on nearly the same number Anderson 2015, The Spine Journal.
- Hand, wrist, elbow and shoulder surgery — patients not on comp were about 3× more likely to improve after surgery (101 articles reviewed) Fujihara & Chung 2017, Plastic & Reconstructive Surgery.
- Carpal tunnel and rotator cuff repair — compensated patients took significantly longer to get back to work, though this rests on only 3 studies and 315 patients de Moraes 2013, Patient Safety in Surgery.
- Knee replacement — comp patients reported more pain and worse function than matched patients. In one study, only 5 of 21 returned to their old job Saleh 2004, Journal of Arthroplasty. In another, 68% of those not retired still hadn't returned to work about three years on Canadian Journal of Surgery 2005 (Ontario WSIB).
If you've been treated like you're the only one not bouncing back — you are not. This is a pattern large enough to show up in every pooled analysis anyone has run.
2. The claims process is doing something to you
This is the part injured workers already know and rarely see written down in a medical journal. In 2014, researchers followed 332 injury claimants and interviewed them six years after they were hurt. They asked how stressful the claims process itself had been — the delays, the paperwork, the assessments, being believed Grant & Studdert 2014, JAMA Psychiatry.
The people who found the process highly stressful were, six years on, significantly worse off on every measure Grant & Studdert 2014, JAMA Psychiatry:
- Worse disability — nearly 7 points on the WHO disability scale (+6.94 WHODAS)
- Worse anxiety (+1.89 HADS)
- Worse depression (+2.61 HADS)
- Worse quality of life (−0.73 WHOQOL)
The researchers then statistically adjusted for how badly people were hurt in the first place, their health before the injury, and their psychological response to it. The link got smaller — and it was still there. Being confused by what the claim required of you still predicted more disability six years later, after accounting for the injury itself Grant & Studdert 2014, JAMA Psychiatry.
Here is how common each of those stressors was among injured claimants Grant & Studdert 2014, JAMA Psychiatry:
Roughly a third of injured people were badly stressed by the paperwork and the waiting. Not by the injury — by the claim.
What workers told the researchers
Two systematic reviews sifted 1,006 studies down to the 13 strongest that recorded injured workers' own accounts of dealing with insurers Kilgour 2015, J Occupational Rehabilitation. Their conclusions are worth reading slowly, because they are the researchers' words, not ours:
"The majority of interactions were negative and resulted in considerable psychosocial consequences for injured workers."
"Interactions between insurers and injured workers were interwoven in cyclical and pathogenic relationships, which influence the development of secondary injury in the form of psychosocial consequences instead of fostering recovery."
Pathogenic means disease-causing. Secondary injury means a second injury — one you didn't have when you walked in.
The same team's companion paper found that even your doctor gets pulled into it Kilgour 2015 ("Healing or Harming?"), J Occupational Rehabilitation:
"Injured workers reported that HCPs [healthcare providers] could play both healing and harming roles in their recovery… Insurer and compensation system processes exerted an influence on the therapeutic relationship."
Their recommendation was not that injured workers try harder. It was that insurers change how they administer claims Kilgour 2015 ("Healing or Harming?"), J Occupational Rehabilitation.
3. The part that argues against us
We're including this because leaving it out would make everything above easier to dismiss.
The same 2017 study that found comp patients were 3× less likely to improve after hand surgery argues that a good chunk of the gap is a measuring problem Fujihara & Chung 2017, Plastic & Reconstructive Surgery. Its actual finding, in the authors' words:
"70 percent of studies found that patients receiving WC had significantly worse postoperative results than uncompensated patients, whereas only 42 percent of studies that measured preoperative versus postoperative improvement were influenced by WC."
On objective physical measures — what the joint can actually do — it drops to 17% Fujihara & Chung 2017, Plastic & Reconstructive Surgery.
Here's what that likely means, and it matters: injured workers tend to arrive at surgery already worse off — in more pain, in worse general and mental health. So their finishing score looks worse even when their improvement is comparable. That's a real difference in starting point, not a broken thermometer. It does not mean injured workers are fine. But it does mean the raw "twice as likely to fail" number carries some of that head start built into it.
Anyone arguing against you will find this study. Better that you find it first.
What this evidence cannot tell you
We'd rather you hear the limits from us than from an insurer's lawyer.
- None of it proves cause. Every study here is observational — it watches what happens, it doesn't run an experiment. The spine researchers say plainly they can't tell whether the effect comes from compensation status itself "or rather… some specific aspects that are necessary to obtain the compensation status" Russo 2021, IJERPH. The six-year study's authors write: "our analyses show associations, not cause-and-effect relationships" Grant & Studdert 2014, JAMA Psychiatry.
- There's a chicken-and-egg problem. People in the six-year study were asked about claim stress at the same interview that measured their recovery Grant & Studdert 2014, JAMA Psychiatry. So it's possible that recovering badly makes you remember the claim as more stressful, rather than the other way round. Researchers have published a formal critique making exactly this point about this whole field Spearing 2012, J Clinical Epidemiology. It has not been settled.
- The strongest stress evidence isn't American. It's Australian Grant & Studdert 2014, JAMA Psychiatry and Canadian Canadian Journal of Surgery 2005 (Ontario WSIB). Those systems are less adversarial and less litigated than US state comp. Applying it here is reasonable inference — it is not a finding about Idaho.
- The quality is honest but limited. The 2012 authors say none of their 20 studies met high methodological standards de Moraes 2012, PLOS ONE. The 2021 spine authors rate their own evidence "low" quality and warn their numbers may be overestimates Russo 2021, IJERPH. The knee studies are tiny — 21 and 22 patients Saleh 2004, Journal of Arthroplasty Canadian Journal of Surgery 2005 (Ontario WSIB).
- Big questions have no answer here yet. We found no verified evidence on lifetime disability rates by insurance type, opioid prescribing, suicide risk, the effect of denied surgery, or anything specific to Idaho. We're still working on those. We won't post numbers we haven't checked.
So what do you do with this?
You use it to stop blaming yourself. The pattern is real, it's documented, and it's big. When an adjuster implies you're an outlier, you're not — you're the median.
And you protect yourself where the research points: the stress that predicted worse recovery six years out came from not understanding what the claim required and from waiting Grant & Studdert 2014, JAMA Psychiatry. Those are the two things you can attack. Read your denial letter — it has to tell you why. Write everything down. Call your state ombudsman; it's free and it's their job. Start on our get help page, find your state's offices on the state page, and learn the playbook being run on you on the tactics page.
You're not imagining it. It's in the journals.
Every source, with links
Read them yourself. Every link goes straight to the journal that published it, to the National Library of Medicine's PubMed record, or to the free full text — never to us. Where a study is free to read in full, we've said so.
Pooled 20 prospective studies and 2,608 patients: workers' comp patients had roughly double the risk of a bad outcome (RR 2.08) after orthopaedic surgery. The conservative estimate in this field — the authors excluded weaker retrospective studies on purpose and still found it.
26 studies, 2,668 patients. Comp patients: 1.79× worse pain, 2.10× more dissatisfied, 1.68× more likely to fail to return to work. The authors rate their own evidence low-quality and note they cannot tell whether compensation status or the process of obtaining it drives the effect.
An independent team, different studies, stronger journal — and nearly the identical figure: 2.12× the risk of an unsatisfactory spine surgery outcome. Independent replication is what makes the spine finding hard to wave away. (Abstract free; full text paywalled by the publisher.)
Read this one closely — it cuts both ways. Comp patients were 3.17× less likely to improve after hand/wrist/elbow/shoulder surgery. But the authors' real argument is that the gap shrinks from 70% of studies to 42% when you measure improvement instead of final score, and to 17% on objective physical measures.
Compensated patients took significantly longer to return to work. Handle with care: only 3 studies and 315 patients, with no adjustment for job type or workplace factors — the authors say compensation "predicts," not causes, and concede the small sample limits how far it travels.
The single most important study for injured workers. 332 claimants followed six years. Those who found the claims process highly stressful had significantly worse disability, anxiety, depression and quality of life — and the link survived adjustment for how badly they were hurt. 33.9% were highly stressed just trying to understand what the claim required.
Reviewed 1,006 studies, kept the 13 strongest recording workers' own accounts. Found most worker–insurer interactions were negative, and described the relationship as "cyclical and pathogenic," producing "secondary injury" instead of fostering recovery. Qualitative — it shows mechanism and lived experience, not effect size.
The companion review. Injured workers said providers could play "both healing and harming roles"; insurer processes "exerted an influence on the therapeutic relationship." Concludes that changing insurer administrative demands could improve worker recovery.
The strongest published attack on this whole field, and we're linking it on purpose. Argues that studies claiming compensation harms recovery ignore the possibility that poor recovery causes the claim experience to be recalled as worse — not the reverse. Unresolved. (Abstract free; full text paywalled.)
Comp patients scored worse than age-matched controls, and only 5 of 21 returned to their old job — though both groups improved and every patient said they'd have the surgery again. Tiny sample, matched on age only, and now two decades old. (Abstract free; full text paywalled.)
38 comp patients vs 38 matched controls who started out the same before surgery. Afterwards comp patients had significantly more pain, worse function and less knee bend. Of those not retired, 68% had not returned to work at ~3 years. Small, single-hospital, Canadian single-payer.
How this page was built. Five research passes across the medical literature, 23 sources read in full, 110 individual claims pulled out and checked one by one against the original papers. Anything that failed the check isn't here — including one knee-surgery claim we dropped entirely. Numbers are quoted with the authors' own caveats attached, and every fact links to its source. If you find an error, tell us and we'll fix it. This page is information, not medical or legal advice.