The story behind the studies begins with the premise that breast cancer is the most common form of cancer in the UK – equating to 15% of all cancers reported. The NHS Breast Screening Programme (NHSBSP) is directed towards diagnosing these cancers at the earliest possible stage to, in short, improve life expectancy through early detection and treatment. But, as with most things in life, not all people are made equal, and specifically some women are at greater risk of developing this disease compared to others. In simple terms this has prompted us to ask should we screen and feedback to women their breast cancer risk and thereafter should women of lower risk be screened less frequently going forwards. Concurrently should high risk women be screened more frequently and be offered potential preventative interventions including risk reducing treatment medication. Thus began the study looking at the feasibility of implementing risk-stratified breast screening in England – the BC-Predict Study.

Importance of patient and public involvement

This blog post is co-authored by the lay patient representative on the BC-Predict and associated sub-studies and wishes recognises the importance of having insight from a member of the general public who is an end user of the screening programme. It gives perspective on user access to the screening programme and the importance of meeting needs for every single end user where possible to ensure that the benefits of screening are maximised for everyone, including NHS services. The lay representative is very appreciative of having had this involvement.

Offering risk assessment and information in a new setting

This study used recognised risk assessment methods (the Tyrer-Cuzick risk algorithm, breast density derived from raw mammography data, data taken from DNA in saliva samples, and a self-report questionnaire) to generate 10-year breast cancer risk for women attending at 5 NHSBSP sites in Northwest England and a further 2 sites within the study offered standard NHSBSP screening. The study had a phased introduction in September 2019. Alarm bells immediately ring, and you will realise that COVID-19 inevitably had a massive impact given the effect of the pandemic on the NHSBSP. Nevertheless, the study proceeded and women were offered BC-Predict via letter shortly after their separate routine NHSBSP mammogram invite. If they chose to participate in the study, they subsequently received their 10-year risk of breast cancer feedback via letter. Women were given one of four 10-year risk categories: low (<2% i.e., less than 2% 10-year risk), average (2-4.99%), above average (moderate; 5-7.99%) or high (≥8%). Importantly women who fell within the moderate or high risk category were then offered risk reducing medication – and this amounted to 62/80 (77.5%) of women at high risk and 43/68 (63.2%) of those at moderate risk.

At this point we want to focus on the difficulties encountered with recruitment. COVID-19 was the number one problem as screening closed for a period and even when it re-opened the obligation moved to the invitees to make an appointment. In short it meant that there was now a shorter period for recruitment, and as the breast screening programme was recovering from the COVID-19, research was not a priority at the sites; their concern was that they didn’t want anything that could negatively influence a woman’s decision not to come for screening.

This significantly slowed recruitment to the study, but COVID-19 was not the only issue. Overall BC-Predict uptake was a disappointing 9.4% of those invited, however, having a HCP at one site to engage with potential participants when attending for their routine mammogram made a significant impact on recruitment to 50.2% for this subgroup of invitees. Furthermore, those who were offered a paper rather than online questionnaire received more favourable uptake and uptake was better in areas with a much lower mean deprivation score. Additionally, although screening invitations do not have ethnicity recorded, in the BC Predict self-reporting questionnaires, 86% of women identified as white British. This was despite work with women from British-Pakistani origins to identify and overcome what might be regarded as barriers to having a risk assessment in this demographic.

Utility of nested studies

This study led to additional sub-studies within the overarching premise viz. including "What do women think about having received their breast cancer risk as part of a risk-stratified NHS Breast Screening Programme? A qualitative study" and an analysis of the "Psychological impact of risk-stratified screening as part of the NHS Breast Screening Programme: Multi-site non-randomised comparison of BC-Predict versus usual screening."

It is, perhaps of no surprise, that these sub-studies encountered similar recruitment challenges.

To extrapolate what do women think about receiving their breast cancer risk as part of a risk-stratified NHS Breast Screening Programme, participants were sourced through the BC-Predict Study. They received their risk feedback, initially by letter, at an appropriate time following the risk assessment with the offer of a telephone interview to follow up. In total, 346 women were invited to participate in an interview (145 low-risk, 103 average-risk, 53 moderate-risk and 45 high-risk), and 40 took part. The women were aged 47-71 years (median 58.5), generally living in less deprived areas (based on IMD decile) and mainly reported having White British/Irish ethnicity.  Of the women with moderate or high-risk, 8 attended appointments with a HCP to discuss their risk. The pandemic did interfere with the provision of appointments with HCP but overall, the concept of risk-stratified screening seemed acceptable to women who took up this offer, and little evidence of adverse emotional effects was found.

We found that it was difficult to reach those women from socially deprived areas and/or of ethnic minority backgrounds despite trying to proactively recruit women based on information in the BC-Predict project. It could have been a longer-term hangover from COVID-19 and people had too many competing priorities to take part in more research.

A sub–set of the BC-Predict study women numbering 5901 were also invited (between Nov 2020 to July 2021) to complete questionnaires to investigate whether there were differences in self-reported harms and benefits between women offered risk-stratification (BC-Predict) compared to women offered standard NHSBSP. These questionnaires were completed at baseline, at 3-months and at 6-months. The target was to address potential psychological impact of risk stratified screening – specifically addressing 3 objectives:

*Are there differences in self-report measures of harms and benefits between women offered BC-Predict compared to women offered NHSBSP, controlling for baseline values?

*Are there differences in self-report measures of harms and benefits between women who accept BC-Predict compared to women who decline BC-Predict, controlling for baseline values?

*Are there differential changes in self-report measures of harms and benefits for the four groups of women provided with different risk estimates (i.e., high, moderate, average and below average) by BC-Predict?

 The women received a first communication by letter approximately 7 days after their initial mammogram appointment date, which directed consenting participants to an online portal, though paper copies were available on request. Women who did not respond received a follow up letter approximately 2 weeks later.

 Of 5901 women invited only 662 consented to the study (11.2%) and of those taking part (over the 3 questionnaires) n=358 were offered BC-Predict and n=304 were offered NHSBS. Both groups were of a similar age and follow-up rates were reasonable at 3-months: 77.2% (BC-Predict group 80.7% and NHSBSP group 73.0%) and 6-months 71.5% (BC-Predict group 72.6% and NHSBSP group 70.1%).

 The positive outcome of the research indicated that there was no evidence that offering BC-Predict resulted in any effect on psychological harms, such as general anxiety or cancer worry, or benefits such as screening knowledge, compared to offering NHSBSP. Its limitations however, as before, included the number of participants viz. 11.2% questionnaire response rate of women invited to participate, in part impacted by the reduced number of invitees predominantly due to the effects of COVID-19.

 All is not lost

 In a perfect world and in a non-COVID-19 environment, more could possibly have been done to reach women including those ‘hard to reach’ groups through greater community outreach, better public information about risk reducing treatment medication (so women know there are multiple strategies to manage breast cancer risk) and greater pro-activity by Healthcare Professionals engaging with women at mammography appointments, to name a few.  But that is not to say that the study didn’t provide good data and demonstrates that breast cancer risk stratification can be delivered as part of the NHSBSP, and it supports the uptake of preventative medicines for women at high risk of breast cancer.