Key takeaways
- Critical illness cover pays on a named list with severity tests attached. The ABI model wording excludes prostate tumours graded below Gleason 7, and cancer in situ outright.
- Aviva declined 5.2% of its 2025 critical illness claims because the definition was not met, against 3.9% for misrepresentation. On income protection that order reverses: 2.6% against 5.8%.
- The industry floor for a heart attack is a blood test result — troponin T above 200 ng/L. Aviva’s and Legal & General’s current wordings carry no number at all.
- Three insurers paid between 89.3% and 94% of critical illness claims in 2025, on books ranging from 1,426 to 5,586 paid claims each.
- A diagnosis below the cancer threshold gets a partial payment: Legal & General caps it at £25,000, Aviva at the lower of £25,000 or 25% of the cover amount.
The policy pays on a named list, and every name carries a severity test
Does critical illness cover pay out if you're diagnosed with cancer? Not always, and the reason isn't how ill you are. It's whether a consultant's report matches a definition printed in your policy document.
That's the product in one sentence, and it's behind most disappointed claims. The Association of British Insurers — the trade body whose minimum standards most UK insurers sign up to — puts it plainly. Critical illness insurance is "insurance cover that pays out a specified sum when the policyholder or other insured person is diagnosed with a serious illness of a specified kind". A specified kind. Not a serious illness.
Three conditions are compulsory. The ABI Guide to Minimum Standards for Critical Illness Cover, published in September 2022 and updated in April 2023, says a product can only be called critical illness insurance if it covers cancer, heart attack and stroke to at least the model wording set out in the Guide. Those three "amounted to over 80% of all critical illness claims in 2016", the ABI's own figure. Each of the three carries a severity threshold, and two of them are measured in a laboratory.
The Guide is a floor, not a ceiling. Insurers "must use that model wording and cannot amend that wording to provide cover which is less generous" — but they're free to go further, and on two of the three core conditions the big names have. That matters more than the floor itself, and it's where this piece ends up.
Cancer: the exclusions run longer than the definition
The ABI's cancer definition is a single sentence: "Any malignant tumour positively diagnosed with histological confirmation and characterised by the uncontrolled growth of malignant cells and invasion of tissue."
Then come eight bullet points of what isn't covered. The first sweeps out everything early: cancers histologically classified as pre-malignant, cancer in situ, having borderline malignancy, or having low malignant potential. A cancer in situ is a real diagnosis with real surgery attached. Under the model wording it doesn't trigger the main benefit.
Prostate cancer is the clearest case. The wording excludes "all tumours of the prostate unless histologically classified as having a Gleason score of 7 or above or having progressed to at least TNM classification cT2bN0M0 or pT2N0M0 following prostatectomy". A Gleason score is a pathologist's grade of how abnormal the cells look under a microscope. A Gleason 6 prostate cancer is diagnosed, registered and malignant. It doesn't meet the definition.
The same structure repeats down the list. Urothelial tumours of the bladder and urinary tract need to reach at least T1N0M0. Thyroid tumours need T2N0M0. Melanoma confined to the epidermis, the outer layer of skin, is out. Aviva's Critical Illness+ conditions reproduce the whole list and add "non-invasive" to the excluded classifications.
These exclusions do more work than every other definition combined, because cancer is where the claims are. Cancer was 65% of critical illness claims across the market in 2025 on ABI data, 54.8% of Aviva's, and 64% of Royal London's.
Heart attack: the floor is a blood test number the biggest insurers have dropped
The model wording asks for typical clinical symptoms, new ECG or imaging changes, and "the characteristic rise of cardiac enzymes or Troponins recorded at the following levels or higher: Troponin T > 200 ng/L (0.2 ng/ml or 0.2 ug/L), Troponin I > 500 ng/L (0.5 ng/ml or 0.5 ug/L)". Troponin is a protein released into the blood when heart muscle dies. Two things are named as not covered: myocardial injury without infarction, and angina without infarction.
That number has moved a long way. Legal & General's heart attack wording effective March 2007 required "Troponin T > 1.0 ng/ml" or "AccuTnI > 0.5 ng/ml or equivalent threshold with other Troponin I methods". The current floor for troponin T is a fifth of the 2007 level.
More striking, the two wordings quoted here have dropped the threshold altogether. Aviva's Critical Illness+ conditions ask only for "the characteristic rise of cardiac enzymes or Troponins". Legal & General's guide asks for "the characteristic rise of biochemical cardiac specific markers such as troponins or enzymes". Neither prints a number. Someone who has read about troponin thresholds may not have one in their own contract — and someone holding an older policy may.
Stroke: the floor says permanent, and permanent means for the rest of your life
Here the model wording is harshest. A stroke is "death of brain tissue due to inadequate blood supply or haemorrhage within the skull resulting in permanent neurological deficit with persisting clinical symptoms".
The Guide's glossary removes any ambiguity about "permanent": "expected to last throughout the insured person's life, irrespective of when the cover ends or the insured person retires". The deficit itself has to be "present on clinical examination". Explicitly not covered: a transient ischaemic attack, and "an abnormality seen on brain or other scans without definite related clinical symptoms". A scan that shows a stroke, in someone who recovered, doesn't clear the floor.
Both retail wordings quoted here are far better than that. Legal & General's stroke definition asks for "neurological deficit with persisting clinical symptoms lasting at least 24 hours". Aviva pays on permanent deficit, or on scan evidence plus "neurological deficit with persistent clinical symptoms lasting at least 24 hours". On stroke, the gap between the minimum standard and the market is the difference between a lifelong disability and a bad day. Both are ABI-compliant, because the Guide only sets a minimum.
Severity tests aren't confined to the core three. Legal & General's deafness definition requires hearing loss "greater than 70 decibels across all frequencies in the better ear using a pure tone audiogram". Partial loss isn't covered. Every condition on the list has a line like this somewhere in it.
A diagnosis below the threshold gets £25,000, not the cover amount
Insurers don't simply refuse the early-stage cases. They pay a smaller, capped amount for them.
Legal & General includes additional cover for carcinoma in situ of the breast and low grade prostate cancer, and "if a valid claim is made we will pay up to a maximum of £25,000". Its low grade prostate definition is the mirror image of the exclusion — a tumour "having a Gleason score of between 2 and 6 inclusive" that has reached at least T1N0M0.
Aviva's structure is the same shape. Its additional cancer benefits pay "the lower of £25,000 or 25% of the cover amount", rising to £30,000 if upgraded critical illness benefit is bought. On a large policy that partial payment is a fixed budget rather than a proportionate share of the sum insured, so the gap between clearing the threshold and missing it widens as the cover amount rises.
The scale is visible in Aviva's 2025 numbers. It paid £1,337,042 across 67 less advanced cancer claims, against £388m across 5,586 claims on the main benefit, where the average payout was £69,559. A capped partial payment is closer in size to an emergency cash reserve sized to income volatility than to the sum the policy was bought for.
Declines cluster on the definition, not on dishonesty
Aviva publishes a breakdown that separates the two reasons; Royal London and Zurich don't. In 2025 Aviva paid 90.7% of critical illness claims. Of the 9.3% it didn't pay, the split was: definition not met 5.2%, misrepresentation 3.9%, other 0.2%.
Aviva defines the category itself: "this is when the claim is for an illness that either isn't named in the cover or hasn't reached a severe enough point to meet the policy criteria."
Now compare the same insurer's other two product lines in the same year. On income protection, 10% of claims weren't paid — misrepresentation 5.8%, definition not met 2.6%, other 1.7%. On life insurance, 1.3% weren't paid — misrepresentation 1%, definition not met 0.3%. The chart below plots all six figures.
Critical illness is the only one of the three where the definition beats the application form as a reason for a decline. That's the practical difference between critical illness and the product buyers most often confuse it with. Income protection turns on whether you can work, which is why the incapacity definitions and the answers on the application form decide an income protection claim. Critical illness turns on a histology report or a troponin assay. Different products, different failure modes, and buyers routinely hold both.
Three insurers paid between 89.3% and 94% of critical illness claims in 2025
The market numbers for 2025, from the insurers' own claims reports:
- Zurich paid "nearly 94% of critical illness claims", which was "1,426 claims worth £125.4 million" — up from "1,352 claims" worth "£109.7 million" in 2024, a 20% increase in value.
- Aviva paid 90.7%, on 5,586 claims and £388m. Its largest single critical illness claim in 2025 was £1.4m.
- Royal London paid "£198m in critical illness claims to 2,959 customers, with an average payment of £67,000 (89.3% payout rate)". Cancer was 64% of its claims, heart attack 9%, stroke 8%.
Market-wide, the ABI reported that individual critical illness claims totalled £1.25 billion in 2025, at an average payout of £67,000, while 97.9% of all individual protection claims were paid. That 97.9% is the number most people see, and it isn't the critical illness number.
The FCA published the product-by-product split in January 2026, using 2024 ABI claims data: term assurance 96%, critical illness 91%, whole of life 100%, individual income protection 86%, all protection products 98%. Critical illness sits second from bottom. Roughly one claim in eleven didn't pay in 2024, against one in twenty-five for term assurance.
The strongest case that the definitions aren't the problem
There's a serious argument that all of the above describes a product working as designed, and it deserves a hearing.
Start with the economics. The FCA's January 2026 interim report on the pure protection market found that accelerated critical illness has a projected claims ratio of 59% of premiums — behind whole of life at 66% and term assurance at 60%, but ahead of standalone critical illness at 49% and income protection at 40%. On margin it looks tighter still: "net cash flow margins range from 3% (accelerated critical illness) to 18% (underwritten whole of life)". Accelerated critical illness is the thinnest-margin product in the study. The severity thresholds are a large part of why a young buyer can afford six figures of cover.
Second, that 5.2% overstates the severity story. Aviva's own definition of the category covers claims "for an illness that either isn't named in the cover or hasn't reached a severe enough point". A claim for a condition that was never on the list lands in the same bucket as one that missed a threshold by a hair, and nobody outside the insurer can separate them.
Third, the floor has moved consistently in the claimant's favour. The troponin threshold fell from the 2007 level, and the wordings quoted here have since dropped it. The Guide's stated aims are to "create clear expectations of the scope and limitations of critical illness cover and avoid customer disappointment" and to "minimise the number of disputed claims". Over 91% of critical illness claims were paid across the market in 2021, on the ABI's figures, and the 2025 insurer numbers are in the same region.
All of that is true. What it doesn't do is displace the finding: definition-not-met is still the largest single reason critical illness claims get declined, and critical illness is the only Aviva product line where that's the case.
What this claims data cannot tell you
The decline split rests on one insurer, one year, one book. Aviva's critical illness figures include total permanent disability and children's benefit, which have their own definitions and their own failure modes. Royal London and Zurich publish payout rates without a reason breakdown, so their 89.3% and 94% can't be decomposed the same way.
Payout rates aren't comparable across insurers in any case. Different books, different policy vintages, different mixes of standalone and accelerated cover, different average ages. Reading 94% against 89.3% as a quality ranking would be a mistake.
The denominator is also incomplete. Nobody counts the people who read their definition, concluded they'd fail it, and never claimed. Those cases don't appear in any decline statistic.
The wordings quoted here are current. An in-force policy sold years ago carries the wording it was sold with — Legal & General's own conditions history shows the heart attack definition changing repeatedly, and the March 2007 version is still out there on policies written then. The ABI's minimum standards apply to new policies from each insurer's adoption date, not retrospectively.
And the FCA's margin and claims-ratio figures are projections from firms' own models on 2024 new business, not realised profit. They're the best public estimate available, not an audited result.
What would change the picture
The single change that would settle the argument is disclosure. If insurers split "definition not met" into "condition not covered at all" and "condition covered but below the severity threshold", the 5.2% would resolve into a mis-selling problem and a threshold problem, and the two need different fixes. Right now they're reported as one number.
A second change would come from product design. If severity-tiered payouts — where the amount scales with stage rather than switching on at one line — displaced the £25,000 bolt-on as the market norm, the binary that produces these declines would soften on its own. Some upgraded products already move partway there.
The third is the ABI's own review cycle, which the Guide sets out and which has already lowered the troponin bar once. A threshold that moves in one review can move again.
The number worth watching isn't the headline payout rate, which has sat above 89% at every insurer here. It's the share of declines attributed to the definition, and whether it keeps sitting above the share attributed to the application form. Whether any of this matters to a particular household is a separate question, and it turns on the difference between risk tolerance and risk capacity — what a balance sheet could absorb without the lump sum, rather than how the diagnosis would feel.