11 January 2012
11 January 2012, 11:00 – 12:00 ** Phone conference **
Present
Professor Julietta Patnick (Chair) Professor Dame Valerie Beral Kath Moser
Professor Sir Richard Peto Professor Amanda Ramirez Professor Malcolm Reed Professor Sir Mike Richards Sarah Sellars
Margot Wheaton Dr Robin Wilson Richard Winder
Apologies
Dr Lucy Carpenter
Documents circulated in advance:
- Agenda
- Draft revised trial protocol (Jan 2012)
- Minutes of TMG meeting 3 February 2011
- Correspondence regarding complaint received regarding the NHSBSP Age Extension trial
- Correspondence with the HTA regarding the proposal to extend the NHSBSP Age Extension trial to older age groups
Plus for background information (4 papers on digital equipment + publication on the number needed to screen to prevent one breast cancer death):
1. Welcome and introductions
2. Minutes of the last meeting (3 February 2011) and matters arising
The minutes of the last meeting (3 February 2011) were accepted as an accurate record of the meeting.
Matters arising:
Sarah Sellars circulated a background document about digital screening. 74 (92.5% of units) breast screening units have at least one digital mammography machine while just over half (42) are fully digital; 6 units have no digital machines.
Action: Sarah Sellars to circulate recently compiled table with details of the sites with digital mammography.
3. Trial update
a) progress to date (KM)
As of October 2011, 40 of the 73 breast screening units due to randomise had started randomising. It is estimated that 330,000 women had been randomised by the end of October 2011. SS said that by the end of January 2012, 47 units will have started randomising. No trial data are available yet.
A complaint has been received from a patient randomised in at the age of 72 regarding the information provided in the patient information leaflet.
It was agreed that in the future, complaints should be referred to Professor Janet Darbyshire, Chair of the Data Monitoring and Ethics Committee (DMEC), to advise on such matters. Any complaints received by individual breast screening units, and regarding this trial, should also be referred to the Chair of the DMEC.
Action: KM to refer this complaint to Janet Darbyshire.
b) protocol
The protocol has been substantially revised. The plan is for it to be agreed, re- submitted to the ethics committee and published soon. It was agreed that members of the TMG would be listed as co-authors on the published protocol.
The protocol should refer via a link to the version of the standard NHSBSP leaflet (that accompanies all invitations to screening) that is currently in use. This would therefore allow for any future changes that may occur in this standard NHSBSP leaflet.
Patient information leaflet
The trial patient information leaflet has not yet been updated. Any revisions will need ethical approval and will be published as part of the protocol. It was clarified that in areas participating in the Age Extension Trial, every woman invited for screening receives the trial patient information leaflet.
Action: The revised trial patient information leaflet will need to be submitted for ethical approval and will form part of the published protocol.
Richard Peto highlighted two proposed major scientific changes to the protocol:- increasing the power of the trial, and extending the trial at older ages.
Increasing the power of the trial
Even with two rounds of randomisation (i.e. recruiting over 6 years), it may still be difficult to guarantee a reliable result for the women in the younger age group (47- 49). In order to increase the power of the study, he advocated excluding from the main comparisons any woman who – at the point of randomisation - on the basis of objective information is unlikely to accept an invitation for breast screening:
- women who did not accept their invitation to attend cervical screening
- women who have moved out the area
- women who have died
- women who developed breast cancer prior to batch creation
Excluding women clear of breast cancer when first invited for routine screening at the average age of 51 would also improve the power of the trial, but it is not yet clear whether it is feasible to do this.
The possibility was raised of undertaking a detailed sub-group analysis into the effect of the type of mammography screening used (digital or analogue). Another sub-group analysis could look into the characteristics of tumours detected (e.g ER positive or ER negative).
An end point committee will need to be convened to determine objectively which deaths can be attributed to breast cancer, as they are sometimes miscertified as lung and bone cancer or pneumonia, particularly in old age.
Action: Any comments regarding the trial protocol should be sent to Kath Moser before the end of January 2012.
4. Proposal to extend the trial at older ages
A proposal has been submitted to the Health Technology Assessment (HTA) programme to extend the trial by inviting the women randomised in aged 71-73 for a further two extra screens (at ages 74-76 and 77-79 years). This is important because it would provide a bigger signal as to the effect of continuation of screening and provide a clear result by 2020, with no chance of a false negative result. It would also mean that the effect of screening for each separate age group could be seen.
In response to the queries raised by the HTA regarding how extending the trial to older ages would increase the power of the study, Richard Peto read the reply he provided to the HTA (see below).
Thanks for your thoughtful letter. I'd like to talk over my concerns, but meanwhile please read the attached revision of the age extension study protocol, and in particular the "Statistical power" section on page 6. We could well get a false negative for the effects of adding an early screen at ages 47-49, in which case the effects at ages 70+ need to be overwhelmingly reliable. At these older ages, however, if women were randomly allocated 0, 1 or 2 extra screens after 70 (in the ratio 2:1:1) then a trend test would have about the same power as randomisation of 0 vs 1 extra screens (in the ratio 2:2), but a trial of 0 vs 2 extra screens (in the ratio 2:2) would have a lot more power, especially for breast cancer deaths at ages 75-84. The appropriate criterion isn't just "power to detect at p=0.01", however; what we need is a narrow CI on the effects, and a p=0.01 result has a lower confidence limit of a negligibly small effect. A study of at least 2 extra screens vs none would offer this, and one of screening to age 79 would do so even better. If we get massively statistically reliable evidence about size of the effects of 2 extra screens, then we could infer reasonably reliably what just one would have offered, achieving what your HTA call was for in a more unambiguous way than a 3-way randomisation would.
Extending the trial for these two extra screening rounds would not affect the trial costs (for follow-up, record linkage etc). However there would be extra costs to breast screening units in providing more mammograms to the older women (1/6 of the women in the trial). This cost would increase gradually over time, and would not reach the full amount for a few years. The cost of these extra screens is estimated at £12m for 2 extra screening rounds.
It was agreed that it would be of great benefit to extend the trial for two extra screens in the upper age group. Subject to ethical approval, this could be tested in the 5 pilot sites by inviting the women randomised in aged 71-73 for a further screen 3 years later; this would indicate what issues might arise, as well as the uptake rates and costs. There may be some reduction in treatment costs due to cancers being detected earlier.
5. Any other business
The group were told about a separate trial that is being undertaken to look at treatment versus no treatment in low risk DCIS. This trial should shed light on the debate over under diagnosis and over–treatment.
6. Date of next meeting
It was agreed that the next meeting of the Trial Management Group should be January 2013 and should be coordinated with the meeting of the Data Monitoring and Ethics Committee.
Action: Kath Moser to circulate an email to fix a date for the next meeting of the Trial Management Group.