Improving outcomes for women aged 70 years or above with early breast cancer: research programme including a cluster RCT

Article history

The research reported in this issue of the journal was funded by PGfAR as project number RP-PG-1209-10071. The contractual start date was in July 2012. The final report began editorial review in February 2021 and was accepted for publication in February 2022. As the funder, the PGfAR programme agreed the research questions and study designs in advance with the investigators. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The PGfAR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2022 Wyld et al. This work was produced by Wyld et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2022 Wyld et al.

Research summary

The original proposal has largely been adhered to in the programme grant, with a few minor variations.

Prospective cohort study

The study recruited 3450 (with a target of 3500) women aged ≥ 70 years, but recruitment was slower than expected and an extension was required from the original 5 years to 6 years. We also increased the number of centres from 22 to 56 to enhance recruitment. A further delay was caused by the UK Cancer Registry, where we planned to get our survival data updated before final analysis. Legal changes caused by the new General Data Protection Regulations1 meant that, although our study was legally compliant with the old legislation, it no longer complied with consent requirements in the new system. A lengthy negotiation process was required to permit access, which added 9 months to our timelines. In total, the study ran for 7 years, rather than the original estimate of 5 years.

Decision support intervention development

The development proceeded to time and target, resulting in the production of decision support interventions (DESIs) suitable to trial in the cluster-randomised controlled trial.

The cluster-randomised controlled trial

The trial started later than expected owing to a delay in receiving National Institute for Health and Care Research (NIHR) approval for phase 2 of the programme, and compounded by slower than anticipated recruitment. Therefore, an extension and an increase in centre numbers were required. However, we did ultimately recruit to target: 1330 out of the target 1328 women.

Health economic analysis

This was completed, but delayed by the registry data access issue noted in Prospective cohort study.

Variation study

This is completed and published.

Overall

The programme has met all of its targets, but was delayed by 24 months. All expected outputs have been published in scientific journals. The online tool has been Medicines and Healthcare products Regulatory Agency (MHRA) approved, is available on the web and is in regular clinical use.

Research pathway diagram

The research pathway diagram is presented in Figure 1.

FIGURE 1.

Research pathway diagram: summary of the relational timelines for the Age Gap programme grant. BCSS, breast-cancer-specific survival; ET ALONE, primary endocrine therapy; HES, Hospital Episode Statistics; OS, overall survival; S+ET, surgery plus adjuvant endocrine therapy. Light-blue shading indicates development of the DESI. Mid-blue shading represents the cohort study. Light-purple shading indicates the cluster RCT. Mid-purple shading indicates development of the models. Light-orange shading indicates analysis of variance. Mid-orange shading indicates health economic analysis.

Breast cancer in older women

Breast cancer is the most common cancer to affect women, with over 55,000 women diagnosed annually,2 and one-third of cases being in women aged ≥ 70 years. The clinical significance of breast cancer is proportionately less in older women, as breast-cancer-specific mortality is overtaken by other-cause mortality once a woman is in her eighties. However, women aged > 80 years still have a higher risk of dying from breast cancer than women in their seventies, which may be because of suboptimal treatment. 3 It is equally important to avoid overtreatment and causing unnecessary harms in women who are very frail. 4 There are currently no tools available that specifically support decision-making in older women with early-stage breast cancer who are faced with a choice of primary endocrine therapy (ET ALONE) or surgery plus adjuvant endocrine therapy (S+ET). The Adjuvant! Online tool5 and the NHS PREDICT tool6 provide survival estimates only for women who have had surgery. This programme of research has developed, validated and trialled an online tool specifically for supporting decision-making in older women.

Surgical treatment of breast cancer in older women

Surgery may be unnecessary for some frailer older women, as disease control may be achieved using antioestrogens alone (ET ALONE). Up to 90% of breast cancers in older women are oestrogen sensitive7,8 and therefore respond well to antioestrogens, potentially allowing women to avoid surgery.

A systematic review of randomised controlled trials (RCTs) evaluating the role of S+ET versus ET ALONE demonstrated no survival difference between the two treatments, but inferior local control for ET ALONE compared with S+ET. 9–12 Longer-term follow-up and a patient-level meta-analysis have recently suggested that survival is superior with S+ET than with ET ALONE (Richard Gray, University of Oxford, 2019, personal communication). These trials were methodologically flawed, and it is therefore not justified to suggest that ET ALONE is not appropriate for any older women. Overtreatment with surgery for very frail older women may be harmful. A recent study of 6000 nursing home residents with breast cancer, all of whom were treated with surgery, reported high morbidity and mortality rates and high levels of long-term functional impairment, although the authors acknowledge that some of the deterioration and deaths may have occurred regardless of their breast cancer treatment. 4

Since these trials were performed in the 1980s, the practice of ET ALONE has changed and is now usually offered only to women with oestrogen-receptor-positive (ER+) cancers, and aromatase inhibitors have largely replaced tamoxifen. 13–16 ET ALONE may, therefore, be more efficacious if potential candidates are selected appropriately, based on health status and tumour biology. From a patient’s perspective, older women express high levels of satisfaction with both S+ET and ET ALONE. 17 Factors cited in favour of ET ALONE are avoiding hospitalisation and surgery, and a desire to retain independence. 17

The purpose of the Age Gap study was to gather data to guide treatment selection for both patients and clinicians to avoid women being overtreated or undertreated.

Adjuvant chemotherapy in older women

Adjuvant chemotherapy is usually advised for high-recurrence-risk breast cancer; however, in women aged ≥ 70 years, there is less certainty about the survival benefits of adjuvant chemotherapy owing to the lack of trial data and to toxicity concerns. 18 The Oxford overview of adjuvant chemotherapy demonstrated reduced mortality for women up to age 69 years,19,20 but there were insufficient data regarding women aged ≥ 70 years to draw firm conclusions. 20

As a result of this uncertainty, in the UK, the National Audit of Breast Cancer in Older Patients (NABCOP) found a two-thirds reduction in chemotherapy use in women aged ≥ 70 years with high-risk disease. 21 Avoidance of chemotherapy in some women may be justified by concerns about poor treatment tolerance; for others, treatment may be well tolerated and improve survival. Rates of mortality and morbidity associated with chemotherapy are higher in women aged ≥ 70 years, with higher rates of haematological toxicity, febrile neutropenia22 and early discontinuation. 23 There is wide variation in chemotherapy rates across the UK,24 which suggests that age- and health-stratified guidelines are required. Generating evidence to support guideline development was one of the aims of this study.

Clinician practice variation in the UK

There is wide variation in rates of non-surgical treatment of breast cancer in the UK, ranging from 5% to 50%,21 suggesting that some women may be overtreated and others undertreated, and highlighting the need for age- and health-stratified guidelines. A similar picture is also seen in the use of adjuvant chemotherapy, with wide variation in rates of use between UK breast units. 24

One of the aims of the Age Gap study was to assess the extent and causes of these variations.

Shared decision-making and decision support

Shared decision-making is joint decision-making by a patient and health-care professional(s) in which they are both experts: the patient on what matters most to them and the health-care professional on the clinical evidence. It can be particularly beneficial in the context of treatment decisions that are preference sensitive, such as those described above for older women with breast cancer. DESIs aim to support shared decision-making and improve the quality of treatment decisions. A systematic review of DESIs showed that they lead to several benefits, such as increased knowledge, greater participation in decision-making, greater comfort with decisions and more accurate risk perceptions. 25

One of the aims of the Age Gap study was to develop DESIs to support older women and their clinicians when engaging in shared decision-making regarding breast cancer treatment.

Aims and objectives

• To provide guidance for the treatment of frailer older women with ER+ breast cancer to ensure optimal treatment with either S+ET or ET ALONE.

• To provide guidance on the treatment of fitter older women with high-recurrence-risk breast cancer to ensure optimal use of adjuvant chemotherapy with or without trastuzumab.

• To design, validate and evaluate (in a RCT) a DESI for older women faced with either of the above choices.

• To determine the health economics of S+ET compared with ET ALONE in older women with ER+ early-stage breast cancer.

• To determine the extent and causes of variation in UK practice relating to older women with breast cancer.

Work packages

• Summation of evidence of current UK practice for older women with early-stage breast cancer.

• Determination of the information needs and preferences of older women when deciding on their preferred breast cancer treatment using a mixed qualitative and quantitative methodology.

• Development of DESIs, including an online tool (using the above models), to be evaluated in a randomised trial.

• Development of outcome models in older women with breast cancer, comparing S+ET with ET ALONE, or adjuvant chemotherapy with no chemotherapy, using UK registry data.

• Acquisition of data from a cohort study of older women with breast cancer to validate the treatment thresholds identified above.

• Cluster-randomised trial of the DESIs’ impact on treatment allocation and outcomes for older women with breast cancer.

• Health economic analysis comparing the cost efficacy of ET ALONE with that of S+ET in older women, based on a mixture of published data, registry data and cohort study data.

• Study of the degree and causes of variation in treatment allocation between UK breast units using registry data, cohort study data and mixed-methods research.

Patient interview study to determine the preferred format and content for the surgery plus adjuvant endocrine therapy versus primary endocrine therapy decision support intervention

Questionnaire study to determine the preferred content for the surgery plus adjuvant endocrine therapy versus endocrine therapy alone decision support intervention

The above qualitative findings were quantified using a bespoke patient questionnaire used with a larger cohort. This study has been published. 27,28

Patient interview study to determine preferred format and content for the adjuvant chemotherapy versus no chemotherapy decision

Development and testing of the decision support interventions

Development of a survival outcome model for surgery plus adjuvant endocrine therapy versus endocrine therapy alone for breast cancer in older women using UK registry data

Validation of the statistical predictive model

Introduction

Adjuvant chemotherapy is recommended for high-recurrence-risk, early-stage breast cancer patients. Meta-analysis41 reported that adjuvant chemotherapy was associated with a reduced risk of recurrence (12%) and death (13%) in patients aged ≥ 70 years, compared with no chemotherapy.

A survival outcome model was, therefore, developed using retrospective analysis of data for breast cancer patients aged ≥ 70 years, from two English cancer registry regions, diagnosed between 2002 and 2012. This compared the effects of adjuvant chemotherapy or no chemotherapy in women aged ≥ 70 years. The outcomes were stratified by age, cancer stage and biology, and health status. The resultant outcome model was used in the Age Gap online algorithm. The work has been published. 33

Methods

Cancer registration records were obtained for all breast cancer diagnoses in women aged ≥ 70 years between 2002 and 2012 from two English registries. Comorbidity was derived from linked HES data and aggregated into a proxy CCI. English indices of deprivation42 (2010) were derived from postcodes.

Analyses were restricted to women diagnosed with stage I–III disease, treated with surgery and adjuvant chemotherapy or no chemotherapy. Some analyses were restricted to patients at high risk of recurrence.

Associations between patients, treatment characteristics and survival were investigated using multivariate proportional hazard regression, using Royston–Parmar, restricted, cubic spline parametric models. 38 Missing data were handled using multiple imputation with chained equations.

Results

Between 2002 and 2012, a total of 29,728 women were diagnosed with breast cancer, of whom 11,735 were included in analysis (stage I–III disease, treated with surgery). Very few patients aged ≥ 80 years received adjuvant chemotherapy, precluding meaningful analysis (n = 122; < 1%), so analyses were restricted to patients aged 70–79 years. Women who had adjuvant chemotherapy were generally younger, were in better health, and had worse disease stage and adverse biology [oestrogen receptor negative (ER–), human epidermal growth factor receptor-2 positive (HER-2+)].

The logistic regression confirmed that a younger age at diagnosis, increased nodal involvement, tumour size and grade, and having ER– or HER-2+ disease were all associated with increased probability of receiving adjuvant chemotherapy. For example, the odds ratio (OR) of adjuvant chemotherapy receipt for each year that the woman was aged ≥ 70 years was 0.763 [95% confidence interval (CI) 0.741 to 0.787; p < 0.001]. Having an ER+ cancer was associated with an OR of adjuvant chemotherapy receipt of 0.305 (95% CI 0.261 to 0.357; p < 0.001); conversely, for HER-2+ disease, the OR was 2.904 (95% CI 2.476 to 3.406; p < 0.001). Similarly, node positivity (1–3 nodes positive: OR 3.739, 95% CI 3.120 to 4.483; p < 0.001) and increased tumour size (per mm increase in tumour size: OR 1.015, 95% CI 1.011 to 1.019; p < 0.001) were also associated with increase adjuvant chemotherapy use.

Naive comparison of survival outcomes between the two treatment groups showed that BCSS and overall survival (OS) were worse for patients who received adjuvant chemotherapy than for those who did not. However, this arises from the difference in risk profiles between the two groups. When all high-risk patients were considered, adjuvant chemotherapy was associated with reduced BCSS compared with no chemotherapy. However, when the analysis was repeated with higher-risk-scoring patients only, the positive effect of adjuvant chemotherapy becomes clear.

The Royston–Parmar model results showed that adjuvant chemotherapy was associated with a significant reduction in the hazard of breast-cancer-specific mortality compared with no chemotherapy, after adjustments were made for patient-level characteristics [hazard ratio (HR) 0.74].

Discussion

This study found that adjuvant chemotherapy was associated with a reduction in breast-cancer-specific mortality for patients aged 70–79 years with high-recurrence-risk cancer, after adjustment for patient characteristics.

High-recurrence-risk disease characteristics were associated with adjuvant chemotherapy use. However, many high-risk patients did not receive chemotherapy, in particular those patients aged > 80 years. This may be justifiable given that older patients are more likely to die of other causes than of breast cancer, so the absolute gains in survival may be small.

This age-, comorbidity- and tumour-characteristic-stratified outcome model was, therefore, suitable to use in the online Age Gap tool that would be tested in the cluster-randomised controlled trial.

Introduction

An integral component of the DESI was an online tool to permit personalised survival prediction, stratified according to tumour stage, biology, fitness and frailty. The models described previously were used to ‘power’ the online tool, with stratification for patient age and comorbidity based on the CCI score. The tools were developed to be useable for both health-care professionals and patients. The development process has been summarised in a publication. 31

Methods

The design of the online tool was based on data derived from the developed models for the ET ALONE versus S+ET, and adjuvant chemotherapy models (see Developing outcome models for older patients with early-stage breast cancer and Adjuvant chemotherapy model development), supplemented with published frailty data39 and evidence from patient interviews about preferred formats for data presentation (see Defining the information needs and preferences of older women with breast cancer). The output from the tool presented data in text, pictogram or bar chart formats. In addition, in response to patient feedback, stratified outputs could be printed off in an easy-to-read format that could be inserted in the booklets developed for the tool, allowing outcomes to be personalised. A draft version was user tested with both health-care professionals and patients, and modified according to feedback.

Three focus groups were undertaken, including six surgeons, seven clinical nurse specialists and nine oncologists.

Results

Feedback from patients

Eight patients were interviewed, with varied feedback regarding data display formats, which were amended. Items of concern and the response made by the team are shown in Appendix 3, Table 2.

Figures 2 and 3 are screenshots of the final online tool for the S+ET versus ET ALONE tool used during interviews. Figures 4 and 5 are screenshots of the adjuvant chemotherapy tool. Screenshots of the booklets are shown in Appendix 6, Figure 19.

FIGURE 2.

Data collection sheet: comparing S+ET and ET ALONE. Reproduced with permission from the University of Sheffield.

Reproduced with permission from the University of Sheffield.

FIGURE 3.

Pictogram showing the survival estimates at 2 years, comparing S+ET and ET ALONE. Reproduced with permission from the University of Sheffield.

Reproduced with permission from the University of Sheffield.

FIGURE 4.

Data collection sheet: comparing adjuvant chemotherapy and no chemotherapy. Reproduced with permission from the University of Sheffield.

Reproduced with permission from the University of Sheffield.

FIGURE 5.

Pictogram showing the survival estimates at 2 years, comparing adjuvant chemotherapy and no chemotherapy. Reproduced with permission from the University of Sheffield.

Reproduced with permission from the University of Sheffield.

The tool now has MHRA approval and can be accessed online. 43

Examples of the brief decision aid for ET ALONE versus S+ET, and for the adjuvant chemotherapy or no chemotherapy decision are shown in Appendix 4, Table 3, and Appendix 5, Table 4, respectively. Images relating to the booklets are shown Appendix 6, Figure 19, and images relating to the outputs from the online tool are shown in Figure 2 and Appendix 7, Figure 20.

Introduction

The purpose of this study was to recruit a large prospective cohort of older women with early-stage breast cancer with baseline information regarding health, cancer stage, biology and treatment. These data were analysed using propensity score matching to adjust for baseline variation in health and fitness to determine outcomes (survival and QoL) between women treated with S+ET versus ET ALONE, or adjuvant chemotherapy versus no chemotherapy. This would allow us to determine optimal age- and health-stratified thresholds for treatment allocation. Several specific, focused analyses were performed, which are summarised in subsequent sections, including:

• general cohort methods and results

• surgery outcomes (see Outcomes from surgery for older women with early-stage breast cancer)

• overall and BCSS outcomes for women having S+ET versus ET ALONE (see Endocrine therapy alone versus surgery plus adjuvant endocrine therapy: survival analysis)

• QoL outcomes for women having S+ET versus ET ALONE (see Quality-of-life variation between primary endocrine therapy and surgery plus adjuvant endocrine therapy)

• overall and BCSS outcomes for women following surgery who did or did not have adjuvant chemotherapy (see Adjuvant chemotherapy versus no chemotherapy in women with high-recurrence-risk breast cancer using both unmatched and propensity-score-matched analyses)

• QoL outcomes for women following surgery who either did or did not have adjuvant chemotherapy (see Quality-of-life impacts of chemotherapy)

• treatment allocation and survival outcomes in women with or without cognitive impairment (see Analysis of the impact of cognitive impairment on outcomes for older women with early-stage breast cancer)

• health economic outcomes for women who had S+ET versus ET ALONE (see Cost-effectiveness analysis comparing surgery plus adjuvant endocrine therapy with primary endocrine therapy).

General methods

General cohort results

Conclusion

The Age Gap cohort study recruited 3414 women aged ≥ 70 years, demonstrating that recruitment in this age group is feasible. Recruitment was slightly skewed towards younger, fitter women, which must be considered when interpreting the generalisability of the results.

Various specific analyses of the cohort study are presented in the subsequent sections.

Introduction

This analysis was performed to describe the allocation of surgery (mastectomy vs. wide excision, and sentinel node biopsy vs. axillary clearance) and the safety and adverse events associated with it. This may enhance understanding of the impacts of surgery. These data have been published. 60

Methods

For this analysis, surgery was classified as major (i.e. mastectomy and/or ALND) or minor (i.e. BCS with or without SLNB). In the case of patients who had a bilateral procedure for invasive breast cancer, surgery was assessed as two unilateral procedures.

Outcomes

Mortality related to surgery was defined as death within 30 days of surgery or surgery being documented as contributing to cause of death. Death from breast cancer was assessed by death certification and expert review of all causes of death. Causes of death were categorised as breast cancer-related or other causes. Complications, obtained by follow-up to 2 years, were categorised using the CTCAE28 and as systemic (i.e. atelectasis, stroke, infarction, deep vein thrombosis/embolism, arrhythmia, allergic reaction or somnolence) or local (i.e. lymphoedema, neuropathy, functional difference, wound pain, wound, necrosis, infection, haematoma, seroma or haemorrhage).

Results

This analysis examined the data from 2816 women who underwent surgery. Of these, 62 had bilateral surgery; therefore, 2858 surgical events are reported. The median age of the surgical cohort was 76 years (range 70–95 years). The majority underwent breast conservation surgery (n = 1716) and 1138 underwent a mastectomy. For axillary surgery, 575 women underwent axillary clearance, 2203 underwent sentinel node biopsy and 76 underwent no axillary surgery.

There were no deaths directly attributable to surgery, although there were three deaths within 30 days of surgery (to which surgery theoretically may have contributed). There were 551 adverse events recorded from these 2858 procedures (19.3%), although the majority of these were simple wound complications. Only 2.1% of women had more serious systemic adverse events, such as cardiac, respiratory or cerebrovascular events. QoL was adversely affected by surgery, with impacts clearly seen in the post-operative period (the 6-week time point; Figure 6). These were usually transient, but some residual negative impact was seen in some domains up to 2 years. Negative QoL impacts were more likely with major surgery (i.e. mastectomy, axillary clearance). There was also a long-term negative impact on physical function, implying a lack of resilience in this age group.

FIGURE 6.

Selected mean QoL domain scores following different types of surgery. (a) Axillary clearance vs. sentinel node biopsy: global health status; (b) mastectomy vs. breast conservation: global health status; (c) axillary clearance vs. sentinel node biopsy: functional QoL score; (d) mastectomy vs. breast conservation: breast symptoms; (e) axillary clearance vs. sentinel node biopsy: arm symptoms; and (f) mastectomy vs. breast conservation: body image. ANC, axillary node clearance; BCS, breast conservation surgery; Mx, mastectomy; SLNB, sentinel node biopsy. Surgery usually occurred between 0 and 1.5 months.

Conclusions

In this cohort, breast surgery is generally safe and well tolerated, with few serious adverse events and no deaths. However, there is a negative impact on QoL that is more significant in women who have more major surgery. These risks need to be communicated to women before surgery as part of shared decision-making. We are in the process of developing the online tool to include risk-stratified QoL and adverse event outcomes to be displayed alongside survival outcomes.

Introduction

This study has used real-world data reflecting current UK practice to determine whether or not there is a group of older women who may be safely offered ET ALONE with little negative impact on their risk of death from breast cancer. The analysis corrected for treatment selection bias using propensity score matching and included only women with ER+ cancers. This analysis has been published. 61

Methods

The cohort recruited women aged ≥ 70 years with operable breast cancer and collected detailed baseline data regarding health, frailty, tumour stage and biology. Normal treatment allocation was permitted to derive a real-world cohort. Analysis of unadjusted outcomes is reported; in addition, to correct for baseline variation in patient allocation to these two treatment arms, propensity score matching was used to create a matched cohort of women with similar baseline characteristics. 61

Results

Overall survival (unmatched analyses)

Of the 486 patients who received ET ALONE, 203 (41.8%) died during follow-up, compared with 336 out of 2307 (14.6%) of the S+ET patients. Patients treated with ET ALONE had inferior OS compared with those treated with S+ET (unadjusted HR 0.27, 95% CI 0.23 to 0.33; p < 0.001), but adjusting for case mix via multivariable Cox regression reduced this difference (adjusted HR 0.83, 95% CI 0.63 to 1.09; p = 0.18). The survival curves are shown in Figure 7.

FIGURE 7.

Kaplan–Meier survival curves for women with ER+ breast cancer treated with either S+ET or ET ALONE in matched and unmatched cohorts. (a) OS: unmatched; (b) OS: matched; (c) BCSS: unmatched; and (d) BCSS: matched.

Discussion

Patient characteristics differed significantly between groups, with ET ALONE generally allocated to older, frailer women than those allocated S+ET. Propensity matching allowed creation of well-matched cohorts of women, although some degree of bias persisted, as demonstrated by the residual variation in OS, where the impact of non-breast-cancer causes of death was still slightly higher in the ET ALONE group than the S+ET group. As a result, although S+ET was still associated with an OS benefit, this difference disappeared in both BCSS and recurrence-/progression-free survival at 52 months’ follow-up.

Surgery plus adjuvant endocrine therapy was the primary treatment modality in 82% and ET ALONE in 18% of women, which is slightly lower than a recent national audit,62 which reported a ET ALONE rate of 24%. This variation may reflect selective recruitment of the less frail into the trial and the technical challenge of recruiting women receiving ET ALONE, in whom treatment often starts on the day of diagnosis, leaving a very narrow window for recruitment.

This study suggests that older women with a short life expectancy of ≤ 5 years derive little survival benefit from S+ET, but women likely to survive > 5 years may start to see endocrine resistance develop on ET ALONE. This accords with previous RCTs63–65 in which survival outcomes at 5 years were equivalent between S+ET and ET ALONE, but outcomes diverged over the longer term.

One of the key aims of this work was to identify a subgroup of women who may safely be offered ET ALONE. It would appear that women with a life expectancy of ≤ 5 years (aged ≥ 85 years with significant comorbidities, or aged > 90 years) should be offered an informed choice, highlighting the potential differences in survival, adverse events and QoL.

Quality-of-life outcomes between treatment groups are presented in Quality-of-life variation between primary endocrine therapy and surgery plus adjuvant endocrine therapy.

Background

There are very few published data regarding QoL outcomes in older women comparing S+ET and ET ALONE. One previous randomised trial comparing ET ALONE and S+ET measured psychiatric morbidity using the General Health Questionnaire-28 (GHQ-28)67 score up to 2 years after diagnosis. 68 The study found that 9 out of 49 (18%) surgical patients and 2 out of 49 (4%) ET ALONE patients had scores indicative of psychiatric morbidity, whereas, by 2 years, there were 6 out of 49 (12%) patients with scores indicative of psychiatric morbidity in both arms. The GHQ-28 has no domains that would reflect the specific impacts of breast cancer. The European Organisation for Research and the Treatment of Cancer (EORTC) QoL instruments are much more specific to the impacts of breast cancer treatment (i.e. the EORTC-QLQ-BR2353) and the QoL concerns relevant to older people (i.e. the EORTC-QLQ-ELD1454). The purpose of this study was to explore the QoL differences between S+ET and ET ALONE in older women with early-stage breast cancer using highly specific QoL tools.

Methods

Quality-of-life assessment was performed on the matched and unmatched cohorts from the ET ALONE versus S+ET analysis using the validated European Organisation for Research and the Treatment of Cancer Quality of Life Questionnaire (EORTC-QLQ) instruments described in The Age Gap prospective observational multicentre cohort study general methods and results.

Owing to significant baseline variation between the unmatched groups in all QoL domains, analyses are largely descriptive and map the changes caused by treatment in each group. In the smaller matched cohort (matched as described in Endocrine therapy alone versus surgery plus adjuvant endocrine therapy: survival analysis), comparative analyses are presented.

Results

EORTC-QLQ-C30: unmatched analysis

The number of patients completing the EORTC-QLQ-C30 global health status at baseline was 1902 out of 2854 (67%): 1644 received S+ET (70% of the total of 2354 patients) and 258 received ET ALONE (52% of the total of 500 patients). Patients treated with ET ALONE had worse baseline scores than those treated with S+ET across all domains of the EORTC-QLQ-C30. In the S+ET group, a steeper change in mean scores between baseline and 6 weeks is apparent in the role functioning, social functioning and fatigue domains (Figure 8).

FIGURE 8.

Summary of key changes in QoL during treatment for breast cancer using the EORTC instruments. (a) Role functioning (EORTC-QLQ-C30 domain); (b) social functioning (EORTC-QLQ-C30 domain); (c) fatigue (EORTC-QLQ-C30 domain); (d) arm symptoms (EORTC-QLQ-BR23 domain); (e) breast symptoms (EORTC-QLQ-BR23 domain); (f) systemic therapy side effects (EORTC-QLQ-BR23 domain); (g) burden of illness (EORTC-QLQ-ELD14 domain); (h) joint stiffness (EORTC-QLQ-ELD14 domain); and (i) family support (EORTC-QLQ-ELD14 domain). Surgery took place between baseline and 6 weeks.

Discussion

Surgery has a clinically relevant impact on a range of QoL domains. Most of these impacts return to baseline levels by 24 months, but there may be some permanent impairment. In ET ALONE patients, the pattern of changes are those of a slow decline across all domains over the 2-year period. These impacts need to be considered and discussed with older women when deciding treatment options. Considering the survival benefit seen in fitter, older women with breast cancer, for most women, despite the largely transient negative impacts of surgery, surgery will be the better option.

However, for older, frailer women with significant comorbidities, for whom any survival benefit may be small, these QoL impairments become more significant. For some of these women, surgery will have no benefit and impair the quality of their remaining lifespan. Older women need to be made aware of this, and we plan to develop the online decision tool to include QoL data so that this may be factored into the decision-making process.

Introduction

The available evidence relating to adjuvant chemotherapy in older women suggests that there is a reduced level of benefit for older women compared with younger women. Nevertheless, benefit is present in some women aged between 70 and 80 years, although there are very limited data for those aged > 80 years. 69 The objectives of this analysis were to determine health-status-stratified outcomes for older women (aged ≥ 70 years) with breast cancer according to whether or not they received adjuvant chemotherapy. This analysis has been published. 70

Methods

Results

Overall survival

The median follow-up of 52 months is reported for 1495 out of 1520 high-risk patients (adjuvant chemotherapy, n = 371; no chemotherapy, n = 1124). Receiving adjuvant chemotherapy was associated with a longer OS than not receiving chemotherapy, but the difference was not statistically significant when adjusted for other covariates: unadjusted HR 0.545 (95% CI 0.40 to 0.73; p < 0.001) and adjusted HR 0.87 (95% CI 0.58 to 1.28; p = 0.47; Figure 9).

FIGURE 9.

Kaplan–Meier survival curves for adjuvant chemotherapy or no chemotherapy in matched and unmatched groups. (a) OS: unmatched 1; (b) OS: matched 1; (c) BCSS: unmatched 2; (d) BCSS: matched 2; (e) metastatic recurrence-free survival: unmatched 3; and (f) metastatic recurrence-free survival: matched 3.

In the matched population, adjuvant chemotherapy was associated with a longer OS than no chemotherapy, although this was not statistically significant: unadjusted HR 0.79 (95% CI 0.50 to 1.3; p = 0.32) and adjusted HR 0.81 (95% CI 0.50 to 1.31; p = 0.38; see Figure 9).

Breast-cancer-subtype-specific analysis of survival

Breast cancers that are ER– and/or HER-2+ are particularly chemosensitive. Therefore, additional exploratory analyses were performed in these subgroups (Figure 10).

FIGURE 10.

Kaplan–Meier survival curves for OS and BCSS in patients treated with adjuvant chemotherapy or no chemotherapy in subgroups with ER– or HER-2+ cancers. (a) OS: HER-2+; (b) OS: ER–; (c) BCSS: HER-2+; and (d) BCSS: ER–.

There were 369 patients with ER– breast cancer with known mortality status, of whom 132 (35.8%) received adjuvant chemotherapy. In a propensity-score-matched analysis of 136 patients, adjuvant chemotherapy was associated with improvements in OS (HR 0.20, 95% CI 0.08 to 0.49) and BCSS (HR 0.12, 95% CI 0.03 to 0.44).

There were 326 patients with HER-2+ breast cancer and known mortality status, of whom 156 (47.9%) received adjuvant chemotherapy with or without trastuzumab. There were fewer deaths from breast cancer and other causes in those who received adjuvant chemotherapy with or without trastuzumab, but, in a matched analysis of 137 patients, the differences were not statistically significant for OS (HR 0.63, 95% CI 0.27 to 1.48) or BCSS (HR 0.50, 95% CI 0.16 to 1.63).

Discussion

Older women treated with adjuvant chemotherapy are younger and fitter and have higher-risk breast cancer (i.e. higher stage, more adverse biology) than those not treated with chemotherapy. When matching for these baseline characteristics is performed, the OS and BCSS benefits disappear. However, in metastatic-recurrence-free survival, a small difference persists in favour of adjuvant chemotherapy in the population of all women, unselected for tumour biology. This apparent benefit did not translate to a survival benefit at 52 months’ follow-up, but longer-term follow-up will be needed to clarify this. 72

It is recognised that the benefits of adjuvant chemotherapy are small for patients with ER+, human epidermal growth factor receptor-2 negative (HER-2–) breast cancer. This study found a benefit from adjuvant chemotherapy in women with ER– breast cancer, with a reduction in the number of breast cancer deaths. These data are consistent with an analysis of US Surveillance, Epidemiology, and End Results (SEER) data, which suggested that the benefits of adjuvant chemotherapy in older women were restricted to those with ER– breast cancer. 73

These data relate to women aged between 70 and 79 years. No comment can be made about women aged > 80 years, as there were insufficient patients aged > 80 years receiving adjuvant chemotherapy to permit analysis.

It is clear that the risks of adjuvant chemotherapy must be considered and balanced against these survival benefits for certain subgroups of women. The present study found that mortality rates from adjuvant chemotherapy were very low and side effects were consistent with previous analyses in this setting. 18 QoL outcomes are reviewed in Quality-of-life impacts of adjuvant chemotherapy.

Introduction

There are few data about QoL outcomes during adjuvant chemotherapy in older women. 74 This study has evaluated QoL in the Age Gap cohort, comparing women who did and did not receive adjuvant chemotherapy. Considering the high priority that older women give to maintaining their QoL, these data are of great clinical value. This article has been published. 75

Methods

Results

Impact of adjuvant chemotherapy on quality-of-life according to adjuvant chemotherapy use (unmatched): EORTC-QLQ-C30

The impact of adjuvant chemotherapy on each domain of the EORTC-QLQ-C30 is shown in Figure 11. As shown in the figure, significant differences in baseline scores between groups were seen and we adjusted for these. The impact of adjuvant chemotherapy was significant on most domains at 6 months, including global health status (MD –9.20, 95% CI –11.95 to –6.44; p < 0.001), physical functioning (MD –8.05, 95% CI –10.21 to –5.89; p < 0.001), role functioning (MD –17.59, 95% CI –21.24 to –13.95; p < 0.001), cognitive functioning (MD –5.55, 95% CI –7.97 to –3.13; p < 0.001), social functioning (MD –18.72, 95% CI –22.17 to –15.27; p < 0.001), and financial problems (MD 3.28, 95% CI 1.16 to 5.39; p = 0.002), all of which largely persisted to 12 months, but had returned to near baseline levels by 18 months.

FIGURE 11.

EORTC-QLQ-C30, EORTC-QLQ-BR23 and EORTC-QLQ-ELD14 selected mean scores and CIs in women treated with adjuvant chemotherapy or no chemotherapy. (a) Global health status/QoL; (b) social functioning; (c) appetite loss; (d) fatigue; (e) body image; (f) systemic therapy side effects; (g) future perspective; (h) upset by hair loss; (i) family support; (j) burden of illness; (k) worries about others; and (l) worries.

Conclusions

Adjuvant chemotherapy has a clinically and statistically significant, largely transient, negative effect on many domains of QoL and this must be taken into consideration when advising patients about whether or not to undergo adjuvant chemotherapy. As seen in the adjuvant chemotherapy survival analysis, survival benefit is largely seen in women with adverse biology (ER– disease), aged between 70 and 79 years. For women outside this age range and with ER+/HER-2– disease, the risk-to-benefit ratio must be carefully considered. It is clear that, for women with adverse tumour biology, in whom a survival benefit may be obtained, these ‘costs’ may be justified. Overall, the impacts of adjuvant chemotherapy are transient, but there is still some residual negative impact at 24 months that may never fully resolve. Older women considering adjuvant chemotherapy must be made aware of the potential survival benefits of adjuvant chemotherapy, and also the potential risks and negative QoL impacts.

Background

One of the key subgroups of women considered for non-standard care are those with cognitive impairment. Rates of dementia are increasing as population ageing occurs, and severe dementia is a significant cause of death. The management and outcomes for women with dementia and breast cancer were of interest. This work has been published. 76

Methods

Results

Survival outcomes

Overall survival at the median 52 months’ follow-up reduced from 81% for women with normal cognition to 36% for those with severe dementia. BCSS also decreased from 40.6% for women with normal cognition to 14% for those with severe dementia. However, because of the heterogeneity of patient characteristics (age, treatment, health status, physical function, tumour size), these differences cannot be ascribed to dementia. On matching, when adjusted for age, treatment type and physical function, the significance of the variation in OS was reduced, but not fully eliminated. The effect of these adjustments on BCSS and progression-free survival (PFS) were no longer significant. This suggests that dementia is a significant predictor of all-cause mortality, but not disease-specific mortality (Figure 12).

FIGURE 12.

Kaplan–Meier survival curves for women with or without dementia, with both unmatched and matched cohorts. (a) OS: survival proportion; (b) OS: matched on age, treatment and NPI; (c) OS: matched on age, treatment, NPI and functionality; (d) BCSS: unmatched; (e) BCSS: matched on age, treatment and NPI; and (f) BCSS: matched on age, treatment, NPI and functionality.

Conclusions

This analysis has demonstrated variation in the treatments that older women with cognitive impairment receive compared with those that women with normal cognitive function receive. The association is confounded by the link between older age and a higher rate of comorbidity in women with cognitive impairment. The use of ET ALONE is higher in women with cognitive impairment, particularly those with severe impairment.

Reasons for the lower rates of surgery in the cognitively impaired groups compared with the non-cognitively impaired groups may include clinicians offering ET ALONE because of a perceived risk, patient and carer preferences for low-risk treatment options, and the patients’ carers’ wish to optimise QoL. Other factors may be clinician preference, and the complexities of surgical consent in patients with cognitive impairment.

There is no difference in the rates of progression-free survival between the groups at 52 months’ follow-up. These data suggest that non-standard treatment may be appropriate for some older women with severe dementia, whose life expectancy is likely to be short, and that women with mild and moderate dementia are still likely to benefit from standard care. Dementia does not appear to be an independent predictor of inferior cancer outcomes.

Introduction

The objective of this study was to use the Age Gap data, combined with other evidence, in a de novo cost-effectiveness model, using individual-level simulation in a partitioned survival model framework, to compare S+ET and ET ALONE. This analysis has been published. 77

Methods

Results

Surgery plus adjuvant endocrine therapy delivered more QALYs than ET ALONE, and was also cheaper, for a fit, 70-year-old, node-negative patient, that is, S+ET dominates ET ALONE for this subgroup. The ICER remained below £10,000 for most patient subgroups, and increased for the oldest, least fit patients only. S+ET was also cost-effective compared with ET ALONE at a threshold of £20,000 for all patient groups, except for patients aged ≥ 90 years, with a comorbidity score of 2 who were node negative, and for all patients aged ≥ 90 years with a comorbidity score of ≥ 3. The cost-effective benefit of S+ET reduced significantly with increasing age and rates of comorbidity.

Conclusions

This study shows that the cost-effective benefit of S+ET reduces with age and comorbidities, and, ultimately, becomes non-cost-effective in an unfit population aged ≥ 90 years. This correlates with the Age Gap survival analyses (see Endocrine therapy alone versus surgery plus adjuvant endocrine therapy: survival analysis).

Introduction

There are no information resources specifically tailored to older women or the breast cancer treatment choices they may face. We have developed and piloted a series of DESIs specifically to support decision-making in older women with breast cancer,27,31 and have used a cluster-randomised controlled trial methodology to determine their impact on shared decision-making and in improving treatment allocation and QoL in older women.

Methods

Results

Recruitment and patient characteristics

Patient flows are shown in Figure 13. There were 46 sites recruiting to the RCT (intervention, n = 21; usual care, n = 25) between December 2015 and June 2018.

FIGURE 13.

Cluster-randomisation and patient recruitment summary.

The median follow-up duration was 36 months. The baseline characteristics of the two arms were well matched.

Kaplan–Meier survival outcomes

Survival

The median follow-up was 36 months. Survival was similar in the two arms, with an estimated 2-year survival of > 90% in both arms and a HR of 1.07 (95% CI 0.80 to 1.43; p = 0.633) in favour of the usual-care arm (Figure 14). Of the 184 deaths (intervention, n = 94; usual care, n = 90), two-thirds were unrelated to breast cancer, giving a HR for BCSS of 0.88 (95% CI 0.54 to 1.44; p = 0.609) in favour of the intervention arm (see Figure 14). Sixty patients (usual care, n = 35; intervention n = 25) experienced recurrence (HR 0.86, 95% CI 0.51 to 1.43; p = 0.556; see Figure 14).

FIGURE 14.

The OS, BCCS and time to recurrence for ITT and per-protocol analyses. (a) ITT analysis: OS; (b) ITT analysis: BCCS; (c) ITT analysis: time to recurrence; (d) per-protocol analysis: OS; (e) per-protocol analysis: BCCS; and (f) per-protocol analysis: time to recurrence.

Discussion

This large, multicentre, cluster-randomised controlled trial is the first to evaluate the effects of two DESIs on clinical decision-making in older women with operable breast cancer. In the ITT analysis, use of the tools did not affect global QoL at 6 months, but the tools were intended for only those women who had been offered either of the relevant choices. Preplanned per-protocol analysis of women offered either choice showed a small improvement in QoL. Treatment choices were altered, with lower rates of S+ET and adjuvant chemotherapy at intervention sites than at usual-care sites. Patient knowledge was enhanced by using the DESIs, suggesting that the DESIs improve informed decision-making.

Analysis of survival outcomes at a median of 36 months’ follow-up showed that, in the ITT analysis, there was minimal survival difference; however, the per-protocol analysis did show some separation of the survival curves, indicating a small potential survival benefit among participants who underwent treatment consultations, as intended by the intervention. Nevertheless, longer follow-up is required.

The DESIs are now available online,43 following granting of MHRA approval. We plan to develop the online tool to incorporate QoL and adverse event outcomes, stratified by patient characteristics and treatment choice, considering the differences in these outcomes demonstrated in the cohort study QoL analyses and the patient feedback about the importance of these outcomes for their decision-making.

Introduction

Medical Research Council guidance on the evaluation of complex interventions were used to investigate the context, implementation and mechanisms of impact of the intervention to explore how the intervention led to its effect. 93 Shared decision-making was a short-term outcome by which the intervention was expected to improve QoL through greater confidence in making the best decision, and with less regret about the choice made. 44 Levels of shared decision-making between the two trial arms were also, therefore, measured.

Methods

In brief, the process evaluation interviewed 73 women and 10 health-care professionals from recruitment sites. It developed and used a range of questionnaire tools to evaluate the dose, reach, fidelity, adaptations, and barriers to and facilitators of the use of the Age Gap tool. Site use of the DESIs was also recorded.

Results

The DESIs were generally highly regarded by both patients and clinicians, and helped facilitate shared decision-making. However, engagement with the DESIs in clinical practice was suboptimal for a number of reasons, and this limits our understanding of their value and likely adoption in routine practice.

Conclusion

The process evaluation sought feedback about the dose, reach, practicality, acceptability and usefulness of the DESIs. The results were favourable, but engagement was limited in some centres.

Introduction

In the UK, there is considerable variation in rates of ET ALONE in women aged ≥ 70 years,21 with rates of between 12% and 40%. 97 Previous authors have used UK registry data to identify factors affecting whether or not surgery is performed in older breast cancer patients, but, to the best of our knowledge, to date, none have studied the variation in treatment allocation based on individual hospitals and clinicians. 98–100 Our research analysed UK practice relating to women aged ≥ 70 years with ER+ breast cancer to determine whether or not the variation observed at hospital- and clinician-levels persists following adjustment for patient characteristics, and tumour stage and biological subtype.

Methods

Records on operable ER+ breast cancers diagnosed in women aged ≥ 70 years between 2002 and 2010 were acquired from two UK cancer registries (WMCIU and NYCRIS). Data on patient and tumour characteristics were included, which were derived from the registries, English Indices of Multiple Deprivation 2010 and HES. A proxy CCI score36 (excluding cancer) was calculated for each patient. 37 The proportion of patients undergoing surgery was calculated for each clinician and hospital. Multivariable logistic regression was used to estimate the probability of a woman undergoing surgery based on age, proxy CCI score, level of socioeconomic deprivation, tumour detection method, tumour size, tumour grade, tumour node metastasis (TNM) stage and nodal status. Missing data were handled using multiple imputation101 to produce 25 imputed data sets, and combining the results. 102

Expected rates of surgery were calculated for each health-care professional and breast unit by summing individual patient probabilities, determined from the logistic regression model. Rates of surgery were risk adjusted by dividing the observed rate by the expected rate for each health-care professional and hospital, and multiplying by the UK national rate. 103 Data are displayed as funnel plots with 95% (2 SD) and 99% (3 SD) SD lines. Logistic regressions were performed in IBM SPSS Statistics, version 21 (IBM SPSS Statistics, Armonk, NY, USA), and multiple imputations were performed using R (version 3.0.1). The analysis was also repeated on the Age Gap cohort data using similar methodology.

Results

Cancer registry analysis

Cancer registration records were obtained for 23,960 patients. After applying the exclusion criteria, 17,129 records remained for analysis. In total, 9955 out of 17,129 (58.1%) patients were treated surgically. The median age was 79 years (range 70–103 years).

The unadjusted rates of surgery varied between hospitals and clinicians, with 39 out of 68 (57.4%) hospitals and 73 out of 167 (43.7%) clinicians falling outside the 95% limits on the funnel plots (Figure 15).

FIGURE 15.

Funnel plots showing unadjusted and adjusted rates of surgery at hospital and clinician levels using registry data, or at hospital level using Age Gap cohort data. (a) Hospital level (registry data): unadjusted; (b) hospital level (registry data): adjusted; (c) clinical level (registry data): unadjusted; (d) clinician level (registry data): adjusted; (e) hospital level (Age Gap data): unadjusted; and (f) hospital level (Age Gap data): adjusted.

Adjusting for case mix did not significantly reduce the variation in surgery rates, with 30 out of 68 (44.1%) hospitals falling outside the 95% limits (see Figure 15). However, at clinician level, adjusting for case mix did reduce the variation in surgery rates, with 17 out of 167 (10.2%) clinicians falling outside the 95% limits on the funnel plot (see Figure 15).

Discussion

This study found considerable variation in the rates of surgery despite case mix adjustment in both the registry and Age Gap data analyses. Analysis of the reasons for this variation confirmed that thresholds for comorbidity, tumour characteristics, dementia and physical function vary widely. Age, comorbidity, frailty, dementia, tumour biology and stage all have an impact on treatment allocation. This underlines the need for fitness-stratified guidelines. The reasons for this variation are explored in Health-care professional interviews and questionnaire study to determine the reasons for treatment variation in older women with early-stage breast cancer.

Methods

A mixed-methods study design was used, comprising initial interviews with a range of health-care professionals, including breast surgeons, specialist nurses and geriatricians. 97 Interviews were recorded, transcribed and analysed using the framework approach. 104

A postal questionnaire was designed using the themes identified in the qualitative interviews to ensure content validity. Face validity and acceptability were assessed during piloting with five health-care professionals. The survey was posted to clinician and nurse members of the UK Association of Breast Surgeons (ABS). In addition, this study attempted to determine the most significant variables influencing this choice in UK health-care professionals using a discrete choice experiment (DCE)105 that used a series of hypothetical scenarios. Key scenario variables included patient age, comorbidity, cognitive status, functional status, and tumour stage and biology. A geriatrician then estimated the probable life expectancy for each scenario and categorised it as < 2 years, 2–5 years or > 5 years, although this information was not shown to participants. For each scenario, individuals were asked to state whether they would choose S+ET, ET ALONE or had no preference. The final DCE had 17 scenarios. 95 Analysis was by multinomial regression using Stata^®, version 13 (StataCorp LP, College Station, TX, USA).

An example of a scenario is shown in Figure 16.

FIGURE 16.

Example of a discrete choice scenario.

Results

Results

Results: questionnaire study

The response rate was 40.2% (258/641). Most respondents (75%) were breast surgeons; 21% were breast care nurses. The results (Figures 17 and 18) show wide-ranging opinions regarding optimal patient management, with some scenarios having a three-way split between surgery, ET ALONE and undecided. It was clear that, in women with > 5 years predicted life expectancy, surgery was preferred, but, in women in poorer health, variability was greater. All five variables, patient age, comorbidity, cognitive status, functional status, and tumour stage and biology, influenced the decision, but comorbidity, cognition and tumour size were most clearly associated with a preference for ET ALONE.

FIGURE 17.

Patient and tumour characteristics, and their importance in shaping the advice health-care professionals would give to an older patient for whom a choice of ET ALONE or S+ET may be considered (n = 258).

FIGURE 18.

Treatment preferences according to estimated life expectancy of each scenario.

Conclusion

This study found wide variation of opinion among health-care professionals regarding the use of ET ALONE for older women, which may be one of the factors underlying wide UK variation. The lack of guidelines on the characteristics of women for whom ET ALONE may be appropriate means that such guidelines are urgently needed. It is hoped that the Age Gap online tool will aid clinical decision-making and harmonise selection of treatment.

Limitations

The Age Gap study has several limitations to consider when interpreting the findings. The most significant is the duration of follow-up. In the cohort analysis comparing S+ET and ET ALONE, all women had ER+ cancers, for which the natural history of the disease is slow. At present, median follow-up is 52 months and, ideally, analysis should be repeated at a median follow-up of 10 years. Previous comparisons of S+ET and ET ALONE, although showing a trend in favour of S+ET at 5 years’ follow-up, show a benefit to S+ET when analysed at longer follow-up. Future research plans to revise the analysis with 10-year data will be made. As a result, the online tool demonstrates outcome at 2 and 5 years only, and, again, this will be updated when modelling has been extended to 10 years.

Another limitation is that, although the Age Gap cohort study is very large, with > 3400 women recruited, when propensity score matching is applied, the sample size inevitably falls, meaning that the matched analysis was smaller. This reduces the chance of statistical significance, although the matched population is still of a reasonable size and the findings are robust. With any matching process there is always a trade-off between the degree of matching and the size of the matched sample. It is felt that the present analysis has optimally matched, while maintaining an adequate sample size.

One limitation of the cluster trial is the fact that median follow-up is only 36 months, which is too short for truly meaningful interpretation, for the same reasons outlined for the cohort study.

Another problem with the cluster trial was the reduction in power caused by the fact that completion of QoL questionnaires was optional. This was chosen to reduce trial burden in this group of often frail women, but, consequently, although our headline recruitment was adequate, the number completing questionnaires was not. Most women who did not elect to complete QoL forms were older and less fit, and therefore more likely to be the type of women for whom ET ALONE is preferable. If we had mandated 100% form completion, we would have selected a population of women for whom S+ET would be preferable, which would have skewed the survival analysis. There was, therefore, no optimal way to design the trial to meet both needs.

Recruitment was slightly skewed in favour of younger and fitter women, as we would expect in this challenging research demographic. However, the fact that the trial managed to recruit > 3400 women, with the oldest woman aged 102 years, is a testament to the support of older patients and also the study design being as accessible as possible for this group.

The models used to populate the online tool require further validation work, particularly the adjuvant chemotherapy model, which we have yet to validate against an external data set. It was also not possible to undertake the health economic modelling relating to adjuvant chemotherapy outcomes owing to lack of time (due to significant delays in recruitment and in gaining access to registry data). Again, this work is planned for the future. The models themselves, although using the best available evidence, did not have a direct frailty measure included, instead making use of imputation of external frailty impacts. Similarly, data held by the UK Cancer Registry during the period studied had missing data for ER status and use of endocrine therapy, although the model inferred ER status from whether or not endocrine therapy was used to ensure the data set was as complete as possible.

Suggestions for future work

The Age Gap team are currently using the QoL data from the cohort study to develop an enhanced online tool that not only provides survival outcomes, but also shows the personalised impact of treatment on QoL and adverse events. This enhanced online tool will then be prospectively evaluated on a further cohort of older women. It is also planned to revise the Age Gap models of survival within the online tool with survival outcomes up to 10 years.

It is also hoped that the adjuvant chemotherapy cost-effectiveness analysis will be completed.

A major theme that emerged from the qualitative data was the desire of older women to maintain their QoL and independence, and they are often prepared to trade this against length of life at any cost. Pilot data have been acquired that show that this is the case, especially for the older women in this age group, and this work will be published separately. It is planned to undertake further mixed-methods analysis of this issue, which will be important for understanding the reasons why many women make treatment choices that may have potentially adverse cancer outcomes.

Another area of interest that merits further study is the role of clinicians in directing decision-making, and how their conscious or unconscious bias may have an impact on treatment choices. We have already explored this with the variation-in-practice interviews and questionnaires, and Malcolm WR Reed and Jenna Morgan are currently undertaking a more detailed investigation of the role of unconscious age bias in treatment selection, which will be published separately.

Patient and public involvement

The Age Gap programme has benefited from a high level and quality of PPI support and engagement across all aspects of the work. We had three experienced and active PPI representatives on the TSC who attended every TSC (twice yearly) since the study began. We had PPI representatives provide input into the original grant application to assist in developing the concept of the project and ensuring it was patient focused and acceptable to patients. The PPI representatives have been involved in discussions about study development and design at each stage, making many valuable suggestions about feasibility, acceptability, patient focus and ethics. They have also been involved in the drafting of all study protocols for each of the many stages of the project and editing all patient facing documents such as consent forms and patient information sheets. They have advised on the conduct of the trial and helped with raising the profile of the study by presenting at academic meetings. They have been involved in the interpretation of qualitative interviews; their perspective has been slightly different from that of the medical researchers and has added depth to our interpretation. They were particularly helpful in the development of the decision support tools, with excellent feedback about design and content. In addition, they have provided detailed comments and editing on all publications arising from this work and are credited with full authorship on all key papers. They have also provided valuable feedback on the final report and have assisted in our plans for follow-on funding applications.

Contributions of authors

Lynda Wyld (https://orcid.org/0000-0002-4046-5940) (Professor of Surgical Oncology) was the chief investigator. She contributed to study development, funding acquisition, trial design, recruitment, governance, management, budget and staff management, writing reports, data analysis and interpretation, and drafting of publications.

Malcolm WR Reed (https://orcid.org/0000-0001-7442-2132) (Professor of Surgical Oncology) was the co-chief investigator. He contributed to study development, funding acquisition, trial design, recruitment, governance, writing reports, analysis and drafting of publications.

Karen Collins (https://orcid.org/0000-0002-4317-142X) (Emeritus Professor of Health Services Research) was the lead for the qualitative study elements. She contributed to study development, funding acquisition, trial design, recruitment, governance, writing reports and drafting of publications.

Sue Ward (https://orcid.org/0000-0002-3344-0913) (Senior Research Fellow in Health Economics and Decision Science) contributed to the development of registry and cohort study models, health economic analysis, data analysis, and drafting of publications.

Geoff Holmes (https://orcid.org/0000-0002-8861-9638) (Senior Research Fellow in Health Economics and Decision Science) contributed to development of registry and cohort study models, health economic analysis, data analysis and drafting of publications.

Jenna Morgan (https://orcid.org/0000-0002-5337-5581) (NIHR Clinical Lecturer in Surgery) was the lead for the variation elements of analysis. She contributed to study development, design, governance, data analysis and drafting of publications.

Mike Bradburn (https://orcid.org/0000-0002-3783-9761) (Senior Medical Statistician) was the lead statistician for the project. He contributed to development of the statistical analysis plan, data analysis, data interpretation, writing reports and drafting of publications.

Stephen Walters (https://orcid.org/0000-0001-9000-8126) (Professor of Medical Statistics) was the senior statistician for the project. He contributed to study development, funding acquisition, trial design, oversight of statistical analysis and drafting of publications.

Maria Burton (https://orcid.org/0000-0002-5411-8181) (Senior Research Fellow in Health and Social Care Research) undertook the process evaluation, and was the lead for the qualitative work during the Age Gap tool development, qualitative interviews, trial design, statistical analysis and interpretation and drafting of publications.

Kate Lifford (https://orcid.org/0000-0002-9782-2080) (Research Associate in Health Psychology) was the co-lead for the process evaluation and qualitative analyses, development and evaluation of the online tool, qualitative interviews, data analysis and interpretation, and drafting of publications.

Adrian Edwards (https://orcid.org/0000-0002-6228-4446) (Professor of General Practice) was the co-lead for development of the decision support tool and its evaluation, and lead for the process evaluation and its analysis and interpretation, methodological expertise on decision support tools, data interpretation and drafting of publications.

Kate Brain (https://orcid.org/0000-0001-9296-9748) (Professor of Health Psychology) was an advisor on development and evaluation of the decision support tool, funding acquisition and editing of publications.

Alistair Ring (https://orcid.org/0000-0003-1230-2973) (Consultant Medical Oncologist and Honorary Reader in Breast Cancer Clinical Trials) was the lead for adjuvant chemotherapy and trastuzumab analyses, including study design, conduct, analysis, interpretation and write-up. He was the study advisor for all systemic therapy elements.

Esther Herbert (https://orcid.org/0000-0002-1224-5457) (Medical Statistician) contributed to statistical design, analysis and interpretation, and writing reports and publications.

Thompson G Robinson (https://orcid.org/0000-0003-2144-2468) (Professor of Stroke Medicine) was an advisor on geriatric medicine and assessment, funding acquisition, study design and management, data interpretation, and drafting and editing of publications.

Charlene Martin (https://orcid.org/0000-0002-0988-0353) (Post-Doctoral Research Associate) was the study manager, study monitor, and lead for analyses relating to cognitive impairment, statistical analysis and interpretation, data quality assurance, trial governance and ethics, and drafting and editing of publications.

Tim Chater (https://orcid.org/0000-0002-1138-0147) (Senior Data Manager) contributed to the set-up, management and monitoring of the Age Gap database, data security and quality assurance. He also contributed to the editing of publications.

Kirsty Pemberton (https://orcid.org/0000-0002-3666-2386) (Data Manager) contributed to the set-up and monitoring of the Age Gap database, data security and quality assurance. She also contributed to the editing of publications.

Anne Shrestha (https://orcid.org/0000-0002-5879-5534) (Research Fellow in Surgery) contributed to the design and analysis of QoL data, data collection, trial initiation visits, trial management and drafting and editing of publications.

Anthony Nettleship (https://orcid.org/0000-0003-0658-3668) (Information Technology Specialist) contributed to the design, development and set up of the Age Gap decision tool, tool monitoring and troubleshooting, and editing of publications.

Paul Richards (https://orcid.org/0000-0002-0320-9083) (Research Fellow in Health Economic Modelling) contributed to the development of initial model of outcomes for ET ALONE versus surgery and adjuvant chemotherapy versus no chemotherapy, and editing of publications.

Alan Brennan (https://orcid.org/0000-0002-1025-312X) (Professor of Health Economics and Decision Sciences) contributed to trial development, funding acquisition, study design, trial management, health economic and outcome model development oversight and validation, drafting and editing of publications.

Kwok Leung Cheung (https://orcid.org/0000-0003-2973-0755) (Professor of Breast Surgery and Medical Education) was an expert on geriatric oncology, primary endocrine therapy, trial steering group member and editing of publications.

Annaliza Todd (https://orcid.org/0000-0002-5888-8263) (Age Gap Trial Monitor) contributed to study monitoring, data quality assurance, data source verification, database management, research governance, and drafting and editing of publications.

Helena Harder (https://orcid.org/0000-0002-7296-8227) (Research Fellow) contributed to qualitative interviews and analysis of adjuvant chemotherapy interviews for the Age Gap study, and editing of publications.

Riccardo Audisio (https://orcid.org/0000-0002-7538-4809) (Professor of Surgical Oncology) contributed to the trial design and development; was an expert on geriatric oncology, surgery in the older patient and geriatric assessment; was a Trial Steering Group member; and contributed to editing of manuscripts.

Nicolo Matteo Luca Battisti (https://orcid.org/0000-0002-1063-1717) (Medical Oncologist) was an expert in chemotherapy in older patients, statistical analysis and interpretation of adjuvant chemotherapy, trastuzumab and radiotherapy data, and drafting and editing of publications.

Juliet Wright (https://orcid.org/0000-0002-8482-081X) (Professor of Elderly Medicine) was an expert on frailty assessment and geriatric medicine, was a member of the Trial Steering Group, and contributed to editing of manuscripts.

Richard Simcock (https://orcid.org/0000-0002-9828-4114) (Consultant in Clinical Oncology) was an expert advisor on radiotherapy in older women, was a member of the TSC, and contributed to editing of manuscripts.

Christopher Murray (https://orcid.org/0000-0001-6939-4384) (Managing Director) contributed software and IT expertise, and to funding acquisition, study design and development; was a co-developer of the Age Gap online tool, and was a member of the TSC.

Alastair M Thompson (https://orcid.org/0000-0003-0604-7267) (Professor of Surgery) was chairperson of the DMEC, and contributed to trial oversight and safety monitoring, progress monitoring, trial design and development, and editing of publications.

Margot Gosney (https://orcid.org/0000-0001-9004-7571) (Professor of Elderly Care Medicine) was a member of the DMEC, and contributed to trial oversight and safety monitoring, progress monitoring, trial design and development and editing of publications.

Matthew Hatton (https://orcid.org/0000-0003-2778-7926) (Consultant Clinical Oncologist and Honorary Professor of Clinical Oncology) was a member of the DMEC, contributed to trial oversight and safety monitoring, progress monitoring, and trial design and development, was an advisor on radiotherapy issues, and contributed to drafting and editing of publications.

Fiona Armitage (https://orcid.org/0000-0003-4311-0093) (Clinical Nurse Specialist) was an advisor on nursing issues, and contributed to trial set-up and trial initiation visits, development and evaluation of the online tool, and data interpretation and analysis.

Julietta Patnick (https://orcid.org/0000-0002-5500-3335) (Visiting Professor of Cancer Screening) contributed to funding acquisition, study development and editing of publications.

Tracy Green (https://orcid.org/0000-0002-7198-2828) (PPI representative) was a member of the TSC, and contributed to trial design, trial development, review of trial materials, ethical review and editing of publications.

Deirdre Revill (https://orcid.org/0000-0002-5074-4177) (PPI representative) was a member of the TSC, and contributed to trial design, trial development, review of trial materials, ethical review and editing of publications.

Jacqui Gath (https://orcid.org/0000-0001-8875-5894) (PPI representative) was a member of the TSC, and contributed to trial design, trial development, review of trial materials, ethical review and editing of publications.

Kieran Horgan (https://orcid.org/0000-0003-4173-1777) (Consultant Breast Surgeon and Honorary Professor of Surgery) was the recruitment lead at Leeds Teaching Hospitals NHS Trust; contributed to patient recruitment, consenting, follow-up, data collection and editing of publications; was chairperson of the NABCOP; and contributed expertise in the management of breast cancer in older women.

Chris Holcombe (https://orcid.org/0000-0002-6014-8053) (Professor of Surgery and Consultant Breast Surgeon) was recruitment lead at Liverpool; contributed to patient recruitment, consenting, follow up, data collection and editing of publications, and contributed expertise in the management of breast cancer in older women.

Matt Winter (https://orcid.org/0000-0001-6192-9874) (Consultant Medical Oncologist and Honorary Reader) was recruitment lead at Sheffield, and contributed to patient recruitment, consenting, follow-up, data collection and editing of publications.

Jay Naik (https://orcid.org/0000-0002-4793-8878) (Consultant Medical Oncologist) was recruitment lead at Harrogate, and contributed to patient recruitment, consenting, follow-up, data collection and editing of publications.

Rishi Parmeshwar (https://orcid.org/0000-0002-8162-2746) (Consultant Breast Surgeon) was recruitment lead at Morecambe, and contributed to patient recruitment, consenting, follow-up, data collection and editing of publications.

Publications

Ring A, Battisti NML, Reed MWR, Herbert E, Morgan JL, Bradburn M, et al. Bridging the Age Gap: observational cohort study of effects of chemotherapy and trastuzumab on recurrence, survival and quality of life in older women with early breast cancer. Br J Cancer 2021;125:209–19. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC8292504/pdf/41416_2021_Article_1388.pdf (accessed September 2021).

Martin C, Shrestha A, Morgan J, Bradburn M, Herbert E, Burton M, et al. Treatment choices for older women with primary operable breast cancer and cognitive impairment: results from a prospective, multicentre cohort study. J Geriatr Oncol 2021;12:705–13. URL: www.geriatriconcology.net/action/showPdf?pii=S1879-4068%2820%2930528-2 (accessed September 2021).

Battisti NML, Hatton MQ, Reed MWR, Herbert E, Morgan JL, Bradburn M, et al. Observational cohort study in older women with early breast cancer: use of radiation therapy and impact on health-related quality of life and mortality. Radiother Oncol 2021;161:166–76. URL: https://orca.cardiff.ac.uk/142497/ (accessed September 2021).

Morgan JL, Shrestha A, Reed MWR, Herbert E, Bradburn M, Walters SJ, et al. Bridging the age gap in breast cancer: impact of omission of breast cancer surgery in older women with oestrogen receptor-positive early breast cancer on quality-of-life outcomes. Br J Surg 2021;108:315–25. URL: https://academic.oup.com/bjs/article/108/3/315/6184678 (accessed September 2021).

Dharmarajan KV, Presley CJ, Wyld L. Care Disparities Across the Health Care Continuum for Older Adults: Lessons from Multidisciplinary Perspectives. American Society of Clinical Oncology Educational Book 41. Alexandria, VA: American Society of Clinical Oncology; 2021. URL: https://eprints.whiterose.ac.uk/174377/ (accessed September 2021).

Wyld L, Reed MWR, Collins K, Burton M, Lifford K, Edwards A, et al. Bridging the age gap in breast cancer: cluster randomised trial of two decision support interventions for older women with operable breast cancer on quality of life, survival, decision quality, and treatment choices. Br J Surg 2021;108:499–510. URL: https://academic.oup.com/bjs/article/108/5/499/6182547 (accessed September 2021).

Holmes GR, Ward SE, Brennan A, Bradburn M, Morgan JL, Reed MWR, et al. Cost-effectiveness modelling of surgery plus adjuvant endocrine therapy versus primary endocrine therapy alone in UK women aged 70 and over with early breast cancer. Value Health 2021;24:770–79. URL: https://eprints.whiterose.ac.uk/cgi/request_doc?docid=2230501 (accessed September 2021).

Burton M, Lifford KJ, Wyld L, Armitage F, Ring A, Nettleship A, et al. Process evaluation of the Bridging the Age Gap in Breast Cancer decision support intervention cluster randomised trial. Trials 2021;22:447. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC8278730/pdf/13063_2021_Article_5360.pdf (accessed September 2021).

Wyld L, Reed MWR, Morgan J, Collins K, Ward S, Holmes GR, et al. Bridging the age gap in breast cancer. Impacts of omission of breast cancer surgery in older women with oestrogen receptor positive early breast cancer. A risk stratified analysis of survival outcomes and quality of life. Eur J Cancer 2021;142:48–62. URL: www.ejcancer.com/action/showPdf?pii=S0959-8049%2820%2931274-0 (accessed September 2021).

Todd A, Martin C, Morgan J, Herbert E, Bradburn M, Burton M, et al. Age specific recruitment and retention to a large multicentre observational breast cancer trial in older women: the Age Gap trial. J Geriatr Oncol 2020;12:714–23. URL: www.geriatriconcology.net/action/showPdf?pii=S1879-4068%2820%2930487-2 (accessed September 2021).

Battisti NML, Reed MWR, Herbert E, Morgan JL, Collins KA, Ward SE, et al. Bridging the Age Gap in breast cancer: impact of chemotherapy on quality of life in older women with early breast cancer. Eur J Cancer 2021;144:269–80. URL: www.ejcancer.com/action/showPdf?pii=S0959-8049%2820%2931359-9 (accessed September 2021).

Ward SE, Holmes GR, Morgan JL, Broggio JW, Collins K, Richards PD, et al. Bridging the Age Gap: a prognostic model that predicts survival and aids in primary treatment decisions for older women with oestrogen receptor-positive early breast cancer. Br J Surg 2020;107:1625–32. URL: https://bjssjournals.onlinelibrary.wiley.com/doi/full/10.1002/bjs.11748 (accessed September 2021).

Lifford KJ, Edwards A, Burton M, Harder H, Armitage F, Morgan JL, et al. Efficient development and usability testing of decision support interventions for older women with breast cancer. Patient Prefer Adherence 2019;13:131–43. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC6338238/pdf/ppa-13-131.pdf (accessed September 2021).

Ward SE, Holmes GR, Ring A, Richards PD, Morgan JL, Broggio JW, et al. Adjuvant chemotherapy for breast cancer in older women: an analysis of retrospective English cancer registration data. Clin Oncol 2019;31:444–52. URL: https://eprints.whiterose.ac.uk/146368/4/Clin%20Oncology%202019_chemo_%20authors%20accepted%20%20version_with%20tables%20and%20figs.pdf (accessed September 2021).

Shrestha A, Martin C, Burton M, Walters S, Collins K, Wyld L. Quality of life versus length of life considerations in cancer patients: a systematic literature review. Psycho-Oncology 2019;28:1367–80. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC6619389/pdf/PON-28-1367.pdf

Martin C, Shrestha A, Burton M, Collins K, Wyld L. How are caregivers involved in treatment decision-making for older people with dementia and a new diagnosis of cancer? Psycho-Oncology 2019;28:1197–206. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC6563536/pdf/PON-28-1197.pdf (accessed September 2021).

Ward SE, Richards PD, Morgan J, Holmes GR, Broggio J, Collins K, et al. Omission of surgery in older women with early breast cancer has an adverse impact on breast cancer-specific survival. Br J Surg 2018;105:1454–63. URL: https://eprints.whiterose.ac.uk/132928/4/ETALONE%20SURVIVAL%20paper_BJS_accepted%20_with%20tables%20%2526%20figures_changes%20accepted.pdf (accessed September 2021).

Collins K, Reed M, Lifford K, Burton M, Edwards A, Ring A, et al. Bridging the age gap in breast cancer: evaluation of decision support interventions for older women with operable breast cancer: protocol for a cluster randomised controlled trial. BMJ Open 2017;7:e015133. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC5642653/pdf/bmjopen-2016-015133.pdf (accessed September 2021).

Morgan JL, Walters SJ, Collins K, Robinson TG, Cheung KL, Audisio R, et al. What influences healthcare professionals’ treatment preferences for older women with operable breast cancer? An application of the discrete choice experiment. Eur J Surg Oncol 2017;43:1282–7. URL: https://eprints.whiterose.ac.uk/113923/1/Morgan%20et%20al%20FINAL%20word%20version%20of%20accepted%20paper%20EJSO%202017.pdf (accessed September 2021).

Burton M, Kilner K, Wyld L, Lifford KJ, Gordon F, Allison A, et al. Information needs and decision-making preferences of older women offered a choice between surgery and primary endocrine therapy for early breast cancer. Psycho-Oncology 2017;26:2094–100. URL: https://eprints.whiterose.ac.uk/115105/3/Information%20needs%20and%20DM%20of%20older%20women%20with%20breast%20cancer%20good%20final%20draft.pdf (accessed September 2021).

Richards P, Ward S, Morgan J, Lagord C, Reed M, Collins K, Wyld L. The use of surgery in the treatment of ER+ early stage breast cancer in England: variation by time, age and patient characteristics. Eur J Surg Oncol 2016;42:489–96. URL: https://eprints.whiterose.ac.uk/96118/3/WRRO_96118.pdf (accessed September 2021).

Morgan JL, Richards P, Zaman O, Ward S, Collins K, Robinson T, et al. The decision-making process for senior cancer patients: treatment allocation of older women with operable breast cancer in the UK. Cancer Biol Med 2015;12:308–15. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC4706524/pdf/cbm-12-04-308.pdf (accessed September 2021).

Morgan JL, Collins K, Robinson TG, Cheung KL, Audisio R, Reed MW, Wyld L. Healthcare professionals’ preferences for surgery or primary endocrine therapy to treat older women with operable breast cancer. Eur J Surg Oncol 2015;41:1234–42. URL: https://shura.shu.ac.uk/10090/1/Collins_Healthcare_professionals%E2%80%99_preferences_for_surgery_.pdf (accessed September 2021).

Morgan J, Richards P, Ward S, Francis M, Lawrence G, Collins K, et al. Case-mix analysis and variation in rates of non-surgical treatment of older women with operable breast cancer. Br J Surg 2015;102:1056–63. URL: https://eprints.whiterose.ac.uk/87495/12/registry_variation_in_surgery_rates_paper_submitted_version.pdf (accessed September 2021).

Morgan JL, Burton M, Collins K, Lifford KJ, Robinson TG, Cheung KL, et al. The balance of clinician and patient input into treatment decision-making in older women with operable breast cancer. Psycho-Oncology 2015;24:1761–6. URL: https://eprints.whiterose.ac.uk/169849/3/PON%20Manuscript%20balance%20of%20clinician%20and%20patient%20input%20psychoonoclogy%20accepted%20version.pdf (accessed September 2021).

Lifford KJ, Witt J, Burton M, Collins K, Caldon L, Edwards A, et al. Understanding older women’s decision making and coping in the context of breast cancer treatment. BMC Med Inform Decis Mak 2015;15:45. URL: www.ncbi.nlm.nih.gov/pmc/articles/PMC4461993/pdf/12911_2015_Article_167.pdf (accessed September 2021).

Morgan J, Wyld L, Collins KA, Reed MW. Surgery versus primary endocrine therapy for operable primary breast cancer in elderly women (70 years plus). Cochrane Database Syst Rev 2014;5:CD004272. URL: www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD004272.pub3/epdf/full (accessed September 2021).

Morgan JL, Reed MW, Wyld L. Primary endocrine therapy as a treatment for older women with operable breast cancer – a comparison of randomised controlled trial and cohort study findings. Eur J Surg Oncol 2014;40:676–84. URL: https://eprints.whiterose.ac.uk/169876/3/Manuscript%20for%20submission%20to%20EJSO%20%28Morgan%29%20plus%20tables%20submitted%20version.pdf (accessed September 2021).

Burton M, Collins K, Caldon L, Wyld L, Reed M. Information needs of older women faced with a choice of primary endocrine therapy or surgery for early-stage breast cancer: a literature review. Curr Breast Cancer Rep 2014;6:235–44. URL: https://shura.shu.ac.uk/8484/1/Burton_Information_needs_of_older_Women.pdf (accessed September 2021).

Burton M, Collins KA, Lifford KJ, Brain K, Wyld L, Caldon L, et al. The information and decision support needs of older women (> 75 years) facing treatment choices for breast cancer: a qualitative study. Psycho-Oncology 2015;24:878–84. URL: https://eprints.whiterose.ac.uk/115105/3/Information%20needs%20and%20DM%20of%20older%20women%20with%20breast%20cancer%20good%20final%20draft.pdf (accessed September 2021).

Trial sponsor

This trial was sponsored by Doncaster and Bassetlaw Teaching Hospitals NHS Foundation Trust, Clinical Research Office, Doncaster Royal Infirmary (Doncaster, UK).

Data-sharing statement

The Age Gap data set will be used in 2025/26 for 10-year survival outcome analysis, when we plan to make contact with the UK Cancer Registry again to gain access to 10-year survival outcomes for both the ET ALONE versus surgery cohort, and the adjuvant chemotherapy versus no chemotherapy cohort. Once these key final papers have been published, we will make the data set (of fully anonymised data) available for open access in 2027. All data requests should be submitted to the corresponding author.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health and Care Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, CCF, PGfAR or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the PGfAR programme or the Department of Health and Social Care.

Introduction

The Early Breast Cancer Triallists’ Collaborative Group analysis of adjuvant chemotherapy in women aged ≥ 70 years found that adjuvant chemotherapy reduced the risk of recurrence by 12% and death by 13%. 41 The results were not statistically significant owing to the small number of women aged ≥ 70 years; nonetheless, international guidelines recommend that adjuvant chemotherapy should be considered in fitter older women with high-risk breast cancer. 107 Despite this, numerous studies report low rates of adjuvant chemotherapy and trastuzumab in older women. 21,24,108 Adjuvant chemotherapy omission may result from concerns about toxicity, but age-specific data on this are sparse and are reviewed below. This review was used to populate the adjuvant chemotherapy DESI with age-specific outcome data.

Methods

Studies were identified from MEDLINE, EMBASE, Cumulative Index to Nursing and Allied Health Literature (CINAHL) and the Cochrane Library. The following medical subject heading (MeSH) terms were used to search for adjuvant chemotherapy toxicities in elderly patients with early-stage breast cancer: ‘Breast’, ‘Cancer’, ‘Neoplasm’, ‘Malignancy’, ‘Older’, ‘Aged’, ‘Elderly’, ‘Frail’, ‘adjuvant chemotherapy’, ‘toxicity’, ‘side effects’, ‘adverse effects’, ‘adverse events’ and ‘tolerability’. We searched for relevant studies of trastuzumab toxicity in older patients with early-stage breast cancer using the same MeSH terms, but ‘adjuvant chemotherapy’ was omitted and ‘adjuvant’, ‘Trastuzumab’, ‘Herceptin’ and ‘HER-2’ were used instead.

Results

In total, 627 studies relevant to toxicities of adjuvant chemotherapy and 313 relevant to adjuvant trastuzumab in older people were identified. Of these, 20 were selected that reported toxicity from adjuvant chemotherapy. These comprised five RCTs, three pooled analyses of RCTs, four cohort studies and eight retrospective studies.

In total, data from four studies relevant to adjuvant trastuzumab toxicity in older patients were reviewed; this included one cohort study and three retrospective studies.

Rates of toxicity from adjuvant chemotherapy in older women were as follows:

• The febrile neutropenia/neutropenic sepsis rate was between 0% and 29%.

• Anthracycline-based regimens had a reported febrile neutropenia rate of 0–29%, whereas non-anthracycline regimens had reported rates of between 0% and 16.7%.

• Adjuvant chemotherapy toxicity-related death rates ranged from 0% to 1.5%.

• Trastuzumab-related cardiac morbidity in older patients ranged from 0.2% to 9.2%, with arrhythmias occurring in 0.2–1% of patients (grade 1 severity).

Only 16% of the patients entering the HERA (herceptin adjuvant) and NSABP (national surgical adjuvant breast and bowel project) B31/N9831 adjuvant trastuzumab trials109,110 were aged ≥ 60 years, but the benefits of trastuzumab in this subgroup were the same as for the whole-study populations. In the HERA trial, the incidence rates of cardiac adverse events for those aged ≥ 60 years compared with those aged < 60 years were not significantly different (3.75% vs. 3.61%). 111 However, there were too few patients aged ≥ 65 years to analyse this older group and the clinical trial population selected in the HERA trial may not be representative of the general breast cancer population. The data reviewed suggest that the majority of older patients who are treated with adjuvant trastuzumab complete the course of treatment with low risk of short-term cardiac complications, although there are few data regarding longer-term risks.

Discussion

The median reduction in breast cancer-associated mortality between 1989 and 2006 was 37% for women aged < 50 years, compared with a 2% reduction for women aged > 69 years. 112 The reasons for this are multifactorial, but undertreatment is likely to have contributed. Undertreatment is partly due to understandable concerns over life-threatening toxicities of adjuvant chemotherapy. However, given that the median survival of a 70-year-old woman in the UK is 15.9 years, it becomes apparent that conservatism could leave a cohort of fit, older women inadequately treated who are likely to live long enough to experience a recurrence. Adverse event rates in this older age group are acceptable in published series. 113

Options at a glance

Here are questions women with breast cancer like yours often ask about the treatment options, along with the answer to each question.

Introduction

This section gives an overview of the detailed methods and general results of the Age Gap cohort study. These are relevant to various sections in the main report.

General methods