Cervical Screening: False Negatives and Interval Cancers in Ireland, 1973 - 2007
It's well known hundreds of women were diagnosed with cervical cancer (2008-2015), after normal smears, known as CervicalCheck Scandal.
There's a VITAL question media and politicians didn't seem too fussed to ask... "So how many women did this happen to from 1973 - 2007?"
Attempting to answer a difficult question.
It's vital to ask and (at least try) answer that question.
Screening tests have sensitivity (% of true positives identified) and specificity (% of true negatives identified). Given neither can be 100%, there will always be False Negatives (FN) and False Positives (FP).
How many got cancer after normal smear results *before* the scandal?
In order to ascertain whether the recent scandal, represents an abject failure of the Organized Screening Service, outsourced US Labs etc..
You should maybe want to compare it to the Opportunistic Screening that preceded it.
If 300-500 cancers, following normal smears, is absurdly high, then throw the kitchen sink of blame at CervicalCheck, the screening service, and absolve the limitations of the Pap test itself.
To calculate how many women were diagnosed with invasive cervical cancer from 1973-2007, after a normal smear (FN's + Interval Cancers), you need to know:
-How many women in Ireland were diagnosed with cervical cancer?
-How many women in Ireland died from cervical cancer?
-How many smears were read?
All of that is easier said than done.
There was no National Cancer Registry pre-1994, so it's a bit of a mission to try estimate the number of women diagnosed with cervical cancer.
The easy part is 1994 - 2007, with 2,864 women diagnosed with cervical cancer, according to National Cancer Registry.
Southern Tumour Register was established in 1978 and was the only cancer registry in Ireland, covering Cork and Kerry. They also only disaggregated in-situ and invasive cervical cancer in 1983.
Still, you can take the 1983 - 1990 incidence of invasive cervical cancer in Cork and Kerry as a tentatively fair proxy of national incidence, in the period.
There's a lot of research and multiplication involved, as I needed these rates and numbers of cervical cancer deaths from CSO from 1973 onwards, which should be noted are underestimates of cervical cancer deaths, for reasons I'll later elaborate on.
Once all the rates are known, then have to use the 20+ years old female population for each individual year, to extrapolate nationally, from 1973 - 1982.
For the amount of smears read, it requires a lot of research; which, as ever, means tracking down and compiling the excellent work of so many people, over many decades.
To achieve this I drew from the following sources:
-A wonderful 1986 report compiled by Ethel McKenna for the Department of Health, gives exact figures for the years 1971 - 1985 for St. Luke's Hospital (then-National Cytology Laboratory) and estimates for Holles Street, Coombe, University College Hospital Galway, Cork Mercy etc.
-Screening figures from Grainne McGuckin's amazing 1992 dissertation for the Department of Finance, which obtained figures from the Department of Health.
-Screening figures from 1993 Working Group on Cervical Screening Interim report.
-Screening figures from 1996 Cervical Screening Committee report.
-Screening figures from Marian O' Reilly's brilliant 2001 Cytology Laboratory Survey.
After all that research and maths, which took a few weeks, I'm left with these estimates.
1973 - 2007:
Smears: 5,088,341
Women diagnosed with cervical cancer: 7,896
Deaths from cervical cancer: 2,702
How many got cancer after normal smears?
Now on to the question of how many of them were diagnosed with cervical cancer in the same circumstance as the recent scandal.
Which is to say how many of them were diagnosed with cancer, after a normal smear result?
To answer this question, I needed to draw on the 1990 study by the Department of Obstetrics and Gynaecology of the National Maternity Hospital, Holles Street.
The study reviewed the screening history of 100 Irish women, who sadly had radical hysterectomies between 1981 - 1989, due to invasive cervical cancer.
24% of those women had a normal smear, before being diagnosed with cervical cancer, allied to 11% whose screening history was unknown. Suggesting anywhere between 24% - 35% diagnosed with cancer, after a normal smear.
Due to the fact it's analyzing 100 women, out of the 1,500+ cervical cancers diagnosed in Ireland in the 1980's, it's necessary to compare a range of 24% - 35% to other countries in the 1980's.
A 1984 study in Canada showed 28.6% of women diagnosed with cancer after a normal smear, a 1983 study in Sweden showed 31.3%, while this 1987 study in USA showed 37% of women diagnosed after a normal smear.
I feel due to the fact there was Opportunistic Screening, with irregular recall and poor follow-up to colposcopy, that 30% - 40% range is probably accurate for 1980's Ireland.
These tests were being performed at (I think) the same sensitivity and specificity requirements as the 1980's and before the implementation of Liquid Based Cytology and, later, HPV-testing in Ireland during the 21st century.
Overall, I estimate approximately 2,460 women (range 2,060 - 2,860) were diagnosed with cervical cancer in Ireland in the same circumstance, after a normal smear, in the years 1973 - 2007.
That estimate is a combination of False Negatives (abnormality was present but missed by screening) and Interval Cancers (no abnormality present and invasive cancer developed between smears).
CervicalCheck audit period, 2008 - 2015:
Women diagnosed with cervical cancer: 2,415
Women diagnosed with cervical cancer, after a normal smear: 450 (18.6%)
Pre-CervicalCheck period, 1973 - 2007:
Women diagnosed with cervical cancer: 7,896
Women diagnosed with cervical cancer, after a normal smear: 2,460 (31.3%)
Discussion:
Still, technology alone isn't the reason for the dramatic fall in false negatives and interval cancers.
A well-organized screening programme, with regular recall and excellent follow-up to Colposcopy can make a big difference in reducing the number of cancers missed by screening and increasing diagnosis rates at the earliest possible stage of cancer.
That a much higher number and proportion of invasive cervical cancers occurred after a normal smear in the 1970's - 1990's was probably in-part due to frequency of smears.
-Increased early diagnosis in Stage 1 of cancer and improved survival outcomes.
-Reduced (a) false negatives (b) rate of interval cancers (c) rate of invasive cancers (d) deaths.
One commonality between shambolic Opportunistic and professional Organized screening:
They can't now and couldn't ever prevent every woman screened from getting cervical cancer.
There was always - and will always be - some women that screening sadly couldn't, can't and won't save.
Within all of that 1973 - 2000 chaos of Opportunistic Screening, and the subsequent stability of 2001 - 2015 Organized Screening, from the ICSP trial to CervicalCheck, there was countless thousands of women prevented the ordeal of cancer and/or early death between 1973 - 2015.
People look at the amount of women that screening "fails" but if you spare 100,000 lives, not just from death but from stress, anxiety, life-altering, fertility-ending treatment, and fail to spare 2,000 lives, it's only a failure if you can't see the wood for the trees.
The lives saved never make the 6 o'clock news because objective progress and good news, rarely rises above subjective scandal and bad news.
What happened in the CervicalCheck scandal, False Negatives and Interval Cancers, always happened in Ireland and always happened everywhere else in the world for that matter.
Measured conversations, on what screening can do and what it could ever have done, would benefit everyone.
The best it can do today, yesterday, tomorrow, is save most lives.
If that's the best it can do, it's worth encouraging and protecting screening.
Bibliography:
Irish Medical Journal, Volume 2, June, 1990: Cytological screening history of patients with early invasive cervical cancer, Department of Obstetrics and Gynaecology, National Maternity Hospital, Holles Street: https://archive.imj.ie/Archive/Cytological%20screening%20history%20of%20patients.pdf
Cytology Laboratory Survey 2002 prepared by Marian O'Reilly: https://www.lenus.ie/bitstream/handle/10147/44786/6355.pdf?sequence=1&isAllowed=y
Department of Health, McKenna, Ethel. 1986. The organisation of the cervical screening service in Ireland: https://www.lenus.ie/bitstream/handle/10147/574996/organisationofcervicalscreeningserviceireland1986.pdf?sequence=1&isAllowed=y
Future approaches to screening for cervical cancer in Ireland.: Dissertation for MSc (Public Sector Analysis) December 1992, Grainne McGuckin: https://www.lenus.ie/bitstream/handle/10147/560432/cervicalcancerscreening.pdf?sequence=1
Report of the Department of Health Cervical Screening Committee, 1996: https://www.lenus.ie/bitstream/handle/10147/44463/7673.pdf?sequence=1&isAllowed=y
