Matters of life and death are always emotive, and the CervicalCheck affair has dominated headlines. A storm of finger-pointing and posturing has ensued, but in all the sound and fury it’s easy to lose sight of the facts.
In media coverage smears have been presented as a test to detect cancer in individuals. But this is a misunderstanding – the purpose of a smear test is to identify individuals with markers that might indicate future disease potential. It is explicitly not an individual diagnostic cancer test but a population-wide screening tool designed to reduce cancer incidence in asymptomatic healthy people.
To this end CervicalCheck has been remarkably effective – prior to its introduction Irish cervical cancer rates were rising by 4 per cent annually. Since introduction this trend has reversed, and we have witnessed an annual drop in incidence of 7 per cent.
The distinction between a screening and a diagnostic test is important. Screening seeks risk-markers in a population. Some people with these markers will never develop cancer, while screening misses others. Why then don’t we simply perform more screening, more often? Because screening is not zero-risk, and false positives are extremely common, leading to over-diagnosis with needless invasive procedures.
False results
Over-screening is unethical and ineffective, and the benefits and harms of any screening programme must be carefully weighed up, adjusted as evidence dictates. Such adjustment has recently occurred in America, where guidelines from the American Cancer Society now recommend screening every three years (the EU norm) rather than annual checks.
Crucially, false results are not indicative of ineptitude – the reality is they are unavoidable, even with excellent tests.
For example, the combined ELISA Western blot test for HIV infection boasts a sensitivity of approximately 99.99 per cent. Yet for the typical Irish person, a positive test means one has a 50 per cent chance of having HIV.
To understand this seemingly paradoxical assertion, we must factor in disease prevalence in a low-risk population, about one in 10,000. Were 10,000 people tested, one would have the virus and test positive. Of the remaining 9,999, we would expect a single false positive. We are left with two positive results, only one of which is true – thus the chances of a positive test being a “true” positive is one in two or 50 per cent.
This surprising result is an inescapable consequence of Bayes’ theorem, the mathematical framework central to assessing conditional probability. It frequently yields counter-intuitive results. The inescapable reality is that false positives and negatives are statistically unavoidable.
At the heart of the present outrage lies the revelation that some women were not given their audit results. In many outlets this has been framed as withholding cancer diagnosis from afflicted women, condemning them to a death sentence. This is an explosive claim, and has caused wide-scale panic and anger. But it pivots on a very serious misunderstanding.
When a patient is diagnosed with cervical cancer, audits are conducted on previous smears to ascertain whether potential markers were missed. These audits are performed to ensure screening quality is maintained.
In 209 cases, re-analysis suggested prior smears could have yielded different results. In several cases there was unacceptable delay or outright failure to convey this to patients and their families. This colossal communications failure ought be investigated, but it has snowballed into assertions that it delayed treatment and cost lives. This simply isn’t true – the audits took place after cancer diagnosis. For families affected, the realisation that earlier detection might have been possible is heart-breaking. Yet the reality is that false negatives are entirely unavoidable. This hasn’t stopped a circus of politicians of all stripes perpetuating misinformation further.
With a certain dark irony, some of the political voices hand-wringing now are among those who cast doubt over the safety and efficacy of the HPV vaccine, an intervention which would prevent over 90 per cent of cervical cancers.
We cannot overlook the fact that Ireland’s historical record on women’s health has been punctuated by misogynistic episodes, from symphysiotomy to the Eighth Amendment. The Cervical Check debacle doesn’t quite fit this narrative, but wariness is understandable. Serious questions must be answered, but it’s imperative we don’t fall victim to misunderstanding, that we are guided by facts rather than emotion. The difficult reality is that medical testing is imperfect – in the words of Gerd Gigerenzer “we have to learn to live with uncertainty”.
Dr David Robert Grimes is a physicist, cancer researcher and Maddox prize-winning science writer