By Dr Laura Heavey
Specialist in public health medicine
Every year we collect data (figures) from our colposcopy clinics to see whether the CervicalCheck programme is meeting its standards (aims). This data goes into our CervicalCheck programme report. This year, when we checked the data before we put it in the report, we noticed that in the past we had made a mistake in the way we counted the outcomes for women attending colposcopy clinics. An 'outcome' is the diagnosis the woman gets at colposcopy – the most important diagnosis the woman gets is recorded and counted.
The mistake means that we have over-reported the number of women that were picked up on screening with CIN (abnormal cells on the cervix) or cancer. This change in data doesn’t affect our programme standards.
Importantly, it doesn’t in any way affect the women diagnosed with cervical cancer and CIN. This means that:
- There has been no delay for women in any aspect of their care because of this reporting mistake.
- The women were all informed of their results and they have all received the appropriate treatment and follow up.
- It is the way we counted them up after they had had their diagnosis and treatment that has changed.
This happened because there are two ways that figures like this are collected in the health services. Sometimes a clinic will give figures for 'number of patients' and sometimes they count 'number of visits'. Number of 'visits' is usually higher than the 'patient' figure, because a patient can be seen more than once for the same ‘outcome’ or diagnosis. The 'people' figure only counts someone once in a year, regardless of how many times they are seen at the service.
The mistake also affected how we counted the number of cancers diagnosed in colposcopy. Again, this has no effect on the management or care of people who were diagnosed. What it means is that a very small number of cases were either not counted or in some cases, counted twice.
Since we noticed these differences, we have recounted our colposcopy figures and are noting the corrections in all of our annual reports between 2008 and 2020.
We have prepared a short report with all of the corrected figures going back to 2008.
To make sure this error in our figures doesn’t happen again, we have developed a new protocol to guide the data analysis for our annual reports from now on. We also found that the problem was partly related to the way we take the figures from our database. We will also change the way we do this to improve the accuracy for the future.
The figures were used in the programme reports, in some newspaper articles and on our website. It does not affect the number of cases of cervical cancer reported by the National Cancer Registry, as they use different methods for getting their data on the number of people who have developed cervical cancer. We have also checked how we count figures in our other screening programmes, and their figures aren’t affected by this.
We are sorry that this has happened and have worked quickly to correct the mistake. We are committed to keeping you informed when we find we have made mistakes, and letting you know the steps we are taking to fix them. It is always our intention to provide high-quality, accurate data and on this occasion we did not meet our own high expectations. We have learned from this and believe the measures we have taken will make sure we do not make this error again.