| | The influences of age and co-morbidities on treatment decisions for patients with HER2-positive early breast cancerAccepted 6 January 2010. published online 25 January 2010.
Corrected Proof Abstract ObjectiveTo investigate the influences of age and co-morbidities on the use of adjuvant chemotherapy and trastuzumab in patients with HER2-positive early breast cancer. MethodsThirty surgeons and 101 oncologists reviewed the profiles of 16 hypothetical patients which included details of age, tumour size/grade, nodal/ER status, and co-morbidities. Respondents viewed different patient profiles. Oncologists were asked how likely they would be to prescribe chemotherapy ±trastuzumab. Surgeons were asked whether they would refer to an oncologist. ResultsOncologists’ treatment decisions were most affected by age and co-morbidities: 81% would prescribe chemotherapy for a high-risk patient aged 68 years, but only 47% for an otherwise identical patient aged 73 years. The majority of surgeons (84%) would still refer older patients. ConclusionsNational variation in the use of adjuvant chemotherapy in women aged ≥70 years with high-risk breast cancer is substantial. Practice audits or clinical trials addressing the outcomes of systemic adjuvant therapy are needed for this ever-increasing population of patients. 1. Introduction  Life expectancy is improving, and it is anticipated that the number of people aged 65 years or over in the UK will increase from 9.2 million in 1996 to 14.5 million in 2061 [1]. The incidence of breast cancer increases with age, and in the USA between 2000 and 2003, 43% of women were aged 65 years or over at the time of diagnosis of breast cancer [2]; with a mean age at diagnosis of 61 years. Over the forthcoming decades, this ageing of our population will lead to a significant increase in the number of older women diagnosed with breast cancer. The benefits of post-operative adjuvant chemotherapy in older women diagnosed with breast cancer are uncertain. It is clear from the Oxford overview published in 2005 that the benefits of chemotherapy diminish with age, but the absolute benefits in older patients are unknown: only 4% of the women accrued to the clinical trials analysed were aged 70 years or over [3]. The International Society of Geriatric Oncology (SIOG) recommends that treatment with adjuvant chemotherapy should not be an age-based decision, but instead, should take into account individual patient's estimated absolute benefit, life expectancy, treatment tolerance, and preference [4]. This issue was due to be addressed by the National Cancer Research Network ACTION trial, randomising women aged 70 years or over with high risk of relapse to adjuvant chemotherapy or observation. Unfortunately, the trial closed owing to poor recruitment. Epidemiological analyses of the SEER (Surveillance Epidemiology and End Results) database in women aged 65 years or over with breast cancer suggest a benefit from adjuvant chemotherapy, but that these benefits appear to be restricted to women with poor prognostic features such as ER-negative, lymph-node-positive tumours [5]. The over-expression of the HER2 oncogene is also associated with a poor prognosis; one might therefore expect older patients with this tumour characteristic also to benefit from adjuvant chemotherapy [6]. The issue of adjuvant chemotherapy in this older HER2-positive high-risk population assumes greater significance because adjuvant trastuzumab has only been shown to be of benefit in those who have received adjuvant (or neoadjuvant) chemotherapy. This survey's purpose was to investigate what factors influence surgeons and oncologists when making treatment decisions about the use of chemotherapy (with or without trastuzumab) in patients with HER2-positive early breast cancer. 2. Methods 2.1. Participants Breast cancer specialists from a representative geographical spread of UK cancer networks were contacted to participate in a survey of treatment practice. Oncologists treating at least 5 patients with HER2-positive disease per month and who are responsible for treatment decisions were asked to participate in an online exercise about how they reach treatment decisions for specific patient cases; all who agreed to participate were e-mailed a link to a secure online questionnaire (30-min duration). Surgeons performed the same exercise, but by face-to-face interview. All participants were remunerated appropriately for their time. 2.2. Survey Six risk factors were investigated in this study: age of the patient, tumour size, tumour grade, nodal status, ER status, and any co-morbidities (none to New York Heart Association I–IV [7], Table 1). (All of the patients were defined as HER2 positive.) Based on these six risk factors and their subcategories there were 144 patient profiles generated. These were arranged as 6 blocks of 24 profiles. The method for selecting profiles for a given respondent was to randomly pick a block and then show 16 profiles from that block determined by a Partial Latin Square, a method which will ensure balance across the whole sample. Each respondent was therefore presented with the profiles of 16 patients with HER2-positive breast cancer. Based on the six risk factors and their subcategories, a minimum of 1600 observations would be required from a survey of 100 oncologists assessing 16 patient profiles to ensure robust results. For each of the 16 patient profiles oncologists were asked how likely they would be to prescribe chemotherapy, and how likely they would subsequently be to prescribe trastuzumab. Surgeons were asked how likely they would be to refer the patient to the oncologist, and in their opinion, how likely the patient would be to benefit from chemotherapy followed by trastuzumab. All respondents were instructed to assume that the patients had no objections to treatment and for trastuzumab, that there had been no excessive adverse reactions to any prior chemotherapy given. Answers were calibrated according to the Juster Scale [8] This scale consists of an 11-point numerical scale ranging from 0 to 10, each point associated with verbal (‘no chance’ to ‘certain’) and numerical (‘1/100’ to ‘99/100’) probability statements. Those giving a score of 8 or more have previously been found to be likely to carry out the behaviour being studied, and in this analysis the measure was used as a proxy for prescribing [9]. 2.3. Treatment simulator model Results from the online (oncologists) and face-to-face surveys (surgeons) were collected and processed into a ‘treatment simulator’ model. The simulator is a fully interactive model (MS Excel-based), in which different combinations of patient parameters can be selected according to the pre-defined statistical design. Based on the patient characteristics entered the model displays the proportion of oncologists who would intend to prescribe chemotherapy ±trastuzumab for that specific case (as described above). The experimental model design was developed by an independent market research company (Double Helix). Formal confidence intervals cannot be attributed to the model, but as only statistically significant variables with levels of confidence ranging from 95% to 99% were included, overall confidence in the accuracy of the model outputs is high (multinomial logit). 2.4. Statistical methods (ConJoint analysis) The simulator model uses ConJoint or ‘Consider Jointly’ analyses. The objective of ConJoint analyses is to determine what combination of a limited number of attributes is most influential on respondent choice or decision making. A controlled set of explanatory variables (in this case patient characteristics) is shown to respondents and by analyzing what decisions they make when considering different sets of explanatory variables, the implicit valuation (or utility) of the individual variables can be determined. In this study ConJoint analyses were used to determine the relative importance to oncologists of different salient patient characteristics when selecting management strategies for patients with HER2-positive early breast cancer [10]. The estimation of utilities for each of the risk factor levels also permits the estimation of overall risk factor (attribute) importance. 3. Results 3.1. Respondent demographics and characteristics The survey took place from 29 September 2008 to 24 October 2008. A total of 101 practising oncologists from 31 cancer networks across the UK were contacted by telephone and agreed to participate. Of these, 70 were consultants and 31 were specialist registrars (40 medical oncologists and 61 clinical oncologists). A total of 30 surgeons (15 consultants and 15 specialist registrars) were contacted by telephone and agreed to participate in the study. Oncologists and surgeons had been practising for a range of 4–30 years. 3.2. Main findings The main focus of this report is the findings of the oncologist survey. For all oncologists surveyed the most important factors affecting treatment decisions for patients with HER2+ early breast cancer were co-morbidities (34%) and age (25%) (Fig. 1). We wished to investigate the effects of age and co-morbidity on prescribing practices and these can best be appreciated when one uses the treatment simulator to establish behaviour in specific scenarios. The first scenario chosen was of an older patient with a lymph-node-positive ER-negative tumour; according to the SEER analysis, such patients may potentially benefit from adjuvant chemotherapy [5]. The scenario data generated were: patient aged 73 years, 2–5 cm, grade 3 tumour, ER-negative, 1–3 lymph nodes involved, no co-morbidities. The second scenario chosen was of an older patient with a high-risk ER-positive tumour. Although at a lower risk of recurrence than the patient with an ER-negative tumour, the issue of adjuvant chemotherapy in such patients remains contentious, and such a patient would have been eligible to enter the ACTION trial specifically addressing the benefit of adjuvant chemotherapy. The scenario data generated were: patient aged 73 years, 2–5cm, grade 3 tumour, ER-positive, 1–3 lymph nodes involved, no co-morbidities. The influence of age on prescribing in these two scenarios was investigated by inputting different ages, but keeping the other variables constant (Fig. 2A and B). It can be seen that as patient age increases, oncologists are less likely to consider chemotherapy (±trastuzumab) as a treatment option. Only 47% of oncologists were likely to prescribe chemotherapy for the patient aged 73 years with an ER-negative tumour. For the same patient scenarios surgeons’ likelihood to refer and how likely they thought the patient would the patient be to benefit from chemotherapy followed by trastuzumab showed similar trends (Fig. 3A and B). Although in the 73-year-old patients with an ER-negative tumour surgeons were still likely to refer for an opinion (84%), oncologists become much less likely to prescribe at this age (47%). The influence of co-morbidities was investigated by keeping the age constant at 73 years with the same prognostic features as above, and inputting different levels of co-morbidities (Fig. 4A and B). As co-morbidities increased, the likelihood of an oncologist prescribing chemotherapy and trastuzumab decreased. 4. Discussion Patient and physician surveys, and statistical models that help predict treatment outcomes, are becoming increasingly useful to healthcare providers. The results of this study show that age and co-morbidity are the factors most likely to affect an oncologist's treatment decisions for patients with HER2-positive breast cancer. This is not surprising: the benefits of adjuvant chemotherapy diminish with age [3], and patients with co-morbidities are likely to be at increased risk from the side-effects of chemotherapy. However what is surprising is the degree of change in prescribing practice once the patient's age exceeds 70 years. Eighty-one percent of oncologists would prescribe chemotherapy for the defined high-risk patient aged 68 years, but only 47% would do so if the patient were aged 73 years with the same prognostic features. (Although it should be recognised that oncologists’ prescribing at a younger age was also somewhat conservative, with 10% not prescribing chemotherapy to the same woman if she was 55.) Relatively few patients aged 70 years or over were enrolled in the studies analysed in the Oxford overviews [3]; many oncologists may therefore think that there is insufficient evidence for chemotherapy in this age group. However, there is no clear consensus among the UK oncologists surveyed: approximately half of the oncologists would prescribe chemotherapy and half would not. In reality, the average life expectancy of a 73-year-old woman in the UK is 13.7 years [11], and her risk of relapse from the tumour described is well over 50% at 10 years [12]. Thus her risk of dying from breast cancer is considerable, particularly if there are minimal co-morbidities. It would certainly be reasonable to consider adjuvant chemotherapy and trastuzumab in such a patient, although the absolute benefits of treatment on survival are unknown. An additional finding is that if oncologists are willing to prescribe chemotherapy, they are likely to prescribe trastuzumab in addition. This presumably reflects the greater perceived toxicity of chemotherapy. Surgeons’ perceptions of the benefit of chemotherapy and trastuzumab essentially mirrored those of the oncologists. However, surgeons remained likely to refer patients to an oncologist even if they perceived there to be little benefit from adjuvant systemic therapy. This study has limitations. It was originally conceived as market research, and thus did not address a prospectively defined hypothesis. In addition, whereas such a survey may be useful in judging oncologists intentions, it does not take into account other influences on prescribing, such as local funding arrangements, and an individual patient's circumstances and wishes. These limitations aside, we believe that the findings have significant implications for current practice and ongoing clinical trial design. They reinforce previous data documenting uncertainty regarding the use of adjuvant chemotherapy in women aged over 70 years with high-risk breast cancer, with approximately half of oncologists being willing to prescribe chemotherapy and half not. This presumably reflects uncertainty either over the benefits of treatment or its toxicity in this patient age group. In the absence of a viable study defining treatment benefit, one potential future study design would be to prospectively audit chemotherapy toxicity in this patient age group. This would, of course, be prone to selection bias, as the physician has already made the decision to offer chemotherapy. However it would at least provide some objective data addressing the needs of this ever-increasing population of patients. Reviewers Professor Malcolm Reed, Royal Hallamshire Hospital, Academic Surgical Oncology Unit, K Floor, Sheffield S10 2JF, United Kingdom. Professor Hans Wildiers, University Hospital Gasthuisberg, Dept. of Medical Oncology, Herestraat 49, B-3000 Leuven, Belgium. Funding The market research and administrative support were funded by Roche Products Ltd. (UK). The interpretation, discussion, and publication by the author are independent of the funding organisation, which sought no control over the content of the subsequent publication. Conflict of interest Dr. Alistair Ring has acted as a consultant for Roche Products Ltd. (UK). Acknowledgements Barry Crook and Rebekah Turner of Double Helix Development were responsible for design and implementation of the survey and Gary Bennett was responsible for statistical modelling. Sophie Berry and David Hallett are acknowledged by the author for their impartial administrative support during the preparation of this manuscript. References [1]. [1]Khaw KT. How many, how old, how soon?. BMJ. 1999;319:1350–1352. [2]. [2]Ries LAG, Harkins D, Krapcho M, et al. National Cancer Institute. SEER cancer statistics review 2000–2003. Available at: http://seer.cancer.gov/csr/1975_2003/results_merged/topic_age_dist.pdf. Accessed 27 March 2009. [3]. [3]Early Breast Cancer Trialists’ Collaborative Group . Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15 year survival: an overview of the randomised trials. Lancet. 2005;365:1687–1717. Abstract | Full Text |
Full-Text PDF (423 KB)
|
CrossRef
[4]. [4]Wildiers H, Kunkler I, Biganzoli L, et al. Management of breast cancer in elderly individuals: recommendations of the International Society of Geriatric Oncology. Lancet Oncol. 2007;8:1101–1115. Abstract | Full Text |
Full-Text PDF (269 KB)
|
CrossRef
[5]. [5]Giordano SH, Duan Z, Kuo YF, Hortobagyi GN, Goodwin JS. Use and outcomes of adjuvant chemotherapy in older women with breast cancer. J Clin Oncol. 2006;24:2750–2756.
CrossRef
[6]. [6]Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science. 1987;235:177–182. MEDLINE [7]. [7]Hurst JW, Morris DC, Alexander RW. The use of the New York Heart Association's classification of cardiovascular disease as part of the patient's complete Problem List. Clin Cardiol. 1999;22:385–390. MEDLINE |
CrossRef
[8]. [8]Juster FT. Consumer buying intentions and purchase probability: an experiment in survey design. J. Am. Stat. Assoc. 1966;61:658–696. [9]. [9]Brennan M, Esselmont D. The accuracy of the Juster Scale for predicting purchase rates of branded, fast-moving consumer goods. Market Bull. 1994;5:47–52. [10]. [10]Green PE, Wind J, Rao VR. Conjoint analysis: methods and applications. The technology management handbook. CRC Press; 1999;. [11]. [11]Current interim life tables: United Kingdom (2005–2007). Office for National Statistics. Available at: http://www.statistics.gov.uk/downloads/theme_population/Interim_Life/ILTGB0507Reg.xls. Accessed 27 March 2009 [12]. [12]Adjuvant Online; Accessed 8 April 2009. http://www.adjuvantonline.com. . Alistair Ring, M.A. MRCP M.D. is senior lecturer and honorary consultant in oncology at Brighton and Sussex Medical School, UK. He has specific interests in breast cancer and the treatment of cancer in the elderly. He is a member of SIOG and the EORTC Elderly Task Force. Sussex Cancer Centre, Royal Sussex County Hospital, Eastern Road, Brighton, East Sussex BN2 5BE, United Kingdom  Tel.: +44 0 1273 696955; fax: +44 0 1273 623312.
PII: S1040-8428(10)00003-X doi:10.1016/j.critrevonc.2010.01.002 © 2010 Published by Elsevier Inc. | |
|
FULL TEXT
