lable at ScienceDirect

The Breast 23 (2014) 170e174

Contents lists avai

The Breast

journal homepage: www.elsevier .com/brst

Original article

Decision-making from multidisciplinary team meetings to thebedside: Factors influencing the recruitment of breast cancer patientsinto clinical trials

Chafika Mazouni a,*, Jacqueline Deneuve a, Monica Arnedos a,b, Fanny Prenois a,Mahasti Saghatchian a, Fabrice André a,b,c, Céline Bourgier a, Suzette Delaloge a

aBreast Cancer Group, Institut Gustave Roussy, Villejuif, Franceb INSERM U981, Institut Gustave Roussy, Villejuif, FrancecUniversité Paris Sud, Le Kremlin Bicêtre, France

a r t i c l e i n f o

Article history:Received 11 September 2013Received in revised form5 December 2013Accepted 14 December 2013

Keywords:Breast cancerClinical trialMultidisciplinary team

* Corresponding author. Department of Breast Surg114 Rue Edouard Vaillant, 94805 Villejuif, France. Tel.:42 115256.

E-mail address: chafika.mazouni@gustaveroussy.fr

0960-9776/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.breast.2013.12.008

a b s t r a c t

Aim of the study: Our aim was to determine factors influencing physicians and breast cancer patients torespectively propose or accept participation in a clinical trial following proposals made during amultidisciplinary team meeting (MTM) in a Comprehensive Cancer Centre.Patients and methods: Consecutive patients considered eligible for a clinical trial by a breast cancer-specific MTM were included. A detailed analysis of factors predictive of the physician proposing thetrial and the patient’s acceptance and final inclusion was conducted.Results: MTM proposed 547 inclusions in 25 clinical trials for 397 patients between March andSeptember 2011. The physician proposed the scheduled clinical trial in only 39% of the cases. The patientsaccepted the proposal in 74% of the cases, and finally 29% were included. The main reason for non-inclusion was the physician’s failure to propose the trial in 45e81%, depending on the type of study.The only factor predictive of both the physician proposing the trial and final inclusion was the type ofstudy (both p < 0.001). Diagnostic/prognostic studies were the most frequently proposed trials. Theprofessional status (of the subject) was predictive of acceptance (p ¼ 0.03) with higher rates amongretired patients and executives (84 and 76% respectively).Conclusion: The major reason for non-inclusion in clinical trials was the physician’s failure to propose thetrial, while the patient’s professional status and the type of study influenced both physicians and pa-tients. Educative measures mostly directed at physicians could be implemented to overcome such poorcompliance.

� 2013 Elsevier Ltd. All rights reserved.

Introduction

In the molecular oncology and personalized medicine era, de-cision making for breast cancer (BC) patients at any stage of thedisease relies on decisions emanating from multidisciplinary teammeetings [1,2]. Besides the standard surgical, medical and radiationtherapies, there is considerable leeway for clinical trials in thissetting in all areas of uncertainty. Clinical trials may currentlyenable patients to benefit from treatment de-escalation (surgery,

ery, Institut Gustave Roussy,þ33 1 42 114383; fax: þ33 1

(C. Mazouni).

All rights reserved.

radiation therapy, chemotherapy) or from specific personalizedtargeted therapies in the localized and advanced setting.

However, the rate of patients included in clinical trials inwestern countries has been reported to be less than 10% in mostcountries. This rate varies according to the type of clinical trial andthe disease stage [3,4]. Previous meta-analyses evaluated obstaclesto patient recruitment as well as the possibility of improving pa-tient compliance with medical research [5,6]. The arguments fornon-inclusion concerned both physicians and patients, and weremultifactorial including social, racial, educational and demographicreasons [6]. However, those reports mostly concerned medicalresearch in general or oncology but there is a lack of studies spe-cifically devoted to the BC setting. Moreover, research in BC en-compasses several fields including medical and surgical oncologyas well as radiology and radiotherapy where compliance may notbe documented to the same extent.

Table 1Study population characteristics.

N ¼ 547

Age, median (minemax) 55 (21e88)�50 years 203>50 years 344Marital statusMarried 363 (66.4)Single 179 (32.7)Not assessable 5 (0.9)Professional categoryFarmers 1 (0.2)Chief executives, managers, professionals and self-employed 17 (3.1)Executives and intellectual professions 53 (9.7)Intermediate non-manual workers 92 (16.8)Lower non-manual workers 153 (27.9)Manual workers 6 (1.1)Retired 142 (26)Unemployed 66 (12.1)Not assessable 17 (3.1)BC diagnosisInvasive 499 (91.2)Intraductal 48 (8.8)StageEarly 513 (93.8)Advanced 34 (6.2)Type of studyCognitive 211 (38.6)Studies on interventional innovative therapeutics 115 (21)Diagnostic/prognostic biology 136 (24.9)Imaging 13 (2.4)Radiotherapy 72 (13.1)

C. Mazouni et al. / The Breast 23 (2014) 170e174 171

To determine the factors that might influence the recruitment ofpatients into clinical trials of all-stage breast cancer, we retro-spectively analysed the incidence of accruals and reasons for non-inclusion among patients identified as potentially eligible for anytype of clinical trial by a dedicated BC multidisciplinary team.

Patients and methods

Patient population

Patients whose clinical file had been examined at the multi-disciplinary BC meeting of a single institution, the Institut GustaveRoussy (IGR) Cancer Centre, in Villejuif, France, betweenMarch andSeptember 2011 and who had been considered potentially eligiblefor a breast cancer-specific clinical trial were selected. Eligibilitycriteria included: 1) subjects with a primary BC, 2) on-going follow-up of patients at the IGR, 3) patients had been screened and foundeligible for one of the available clinical trials.

Survey and data collection

The medical records of the patients were retrospectivelyreviewed and complete data were retrieved regarding: 1) patientdemographics (age, marital status, professional category, and livinglocation), 2) information on the disease and health status, and 3)clinical trial pre-screening details (date of screening, proposal andinclusion, type of study).

A questionnaire was given to physicians for the purposes of thisstudy, inquiring about and grading the importance of the reasonsbehind failure to propose a clinical trial to patients in general. Itemsincluded i) frequency: how often trials were proposed once thepatient had been identified by the MTM (never, rarely, sometimes,and often) and ii) reasons for failure to propose the trial (13 items).Replies were given anonymously.

Statistical analyses

Descriptive methods were used to summarize demographiccharacteristics. The Chi-square test was used to compare the dis-tribution of baseline characteristics among groups for categoricalfactors, whereas the Student’s t test was used for continuousvariables.

A 5% significance level was used and all p values were two sided.All analyses were performed in R, an open source statistical package(http://www.r-project.org/) [7].

Results

Patient and study characteristics

Five hundred and forty-seven patients with BC initially eligiblefor a clinical trial were identified by theMTM at the Institut GustaveRoussy (IGR) betweenMarch and September 2011.Within the sameperiod, 1650 files were discussed within the MTM. For the patientsselected and finally not eligible, the reasons traced by physicianswere standard and multiple, therefore we did not consider inter-esting to report them: In 8%, it was related to abnormal biologictests, in 16% to outlier delays, in 3%, to previous medical history, in2% to metastatic work-up, in 20% to geographical reasons, in 8% to atrial closed between screening and visit, in 44%, to other various ornot defined causes.

Patient characteristics are shown in Table 1. The median age ofthe entire cohort was 55 years (range 21e88 years). Five hundredand thirteen patients (93.8%) had early-stage disease, and 34 (6.2%)

had advanced disease. The two diagnoses investigated were:invasive BC (n ¼ 499 [91.2%]) and 48 intraductal carcinoma [8.8%].

A total of 252 patients had been considered eligible for a trial:101 for 2 trials, 24 for 3, 4 for 4 and 1 for 5 trials. The classification ofthe different types of clinical trials and patients proposed for eachtype were as follows: 211 (38.6%) patients for cognitive studies, 136(24.9%) for diagnostic/prognostic biology, 116 (21.2%) for interven-tional therapeutic studies, 71 (13%) for radiation therapy and 13(2.4%) for imaging.

Clinical trials proposed and reasons for non-inclusion

A clinical trial was proposed to 215 (39.3%) patients. Amongthem, 159 patients accepted, which represents 74% of the patientsproposed a trial but only 29% of those initially identified during theMTM.

The main reasons for non-inclusion in clinical trials were: theclinician’s failure to propose the trial in 65.7% (n¼ 255), ineligibilityin 12.1% (n¼ 47), patient refusal in 4.9% (n¼ 19), a study problem in1.3% (n ¼ 5) and other reasons in 16% (n ¼ 62).

The analysis of patient characteristics which could have moti-vated the physicians’ decision to propose the clinical trial is shownin Table 2. The patient’s professional status was not a predictivefactor (p ¼ 0.11). In the univariate analysis, none of the patientcharacteristics were found to have a statistically significant influ-ence on physician triage decision making. Patient age was ofborderline significance, since patients in the “proposed-group”were younger than in the “non-proposed” group (53.9 years vs 55.7years, respectively; p ¼ 0.06). No difference was observed amongthe primary cancer care providers (p ¼ 0.58).

The type of clinical trial significantly influenced the physician’sdecision to propose the MTM-earmarked trials (p < 0.001). Thusstudies on cognitive and interventional innovative therapeuticswere more represented in the “non-proposed” group.

Table 3Univariate analysis: patient and study characteristics according to inclusion or non-inclusion in a trial.

InclusionN ¼ 159

No inclusionN ¼ 388

p Value

Age (mean) 54.6 55.2 0.62Professional status (%) 0.06Farmers 1 (0.6) 0 (e)Chief executives, managers,

professionals and self-employed1 (0.6) 16 (4.1)

Executives and intellectual professions 16 (10.1) 37 (9.6)Intermediate non-manual workers 22 (13.8) 70 (18)Lower non-manual workers 52 (32.7) 101 (26)Manual workers 3 (1.9) 3 (0.8)Retired 45 (28.3) 97 (25)Unemployed 14 (8.9) 52 (13.4)Not assessable 5 (3.1) 12 (3.1)Performance status 0.730 105 (66) 185 (47.7)1 3 (1.9) 9 (2.3)2 1 (0.6) 2 (0.5)Not assessable 50 (31.4) 192 (49.5)Primary care providers 0.25Oncologists 131 (82.4) 298 (76.8)Surgeons 9 (5.7) 29 (7.5)Radiologists 4 (2.5) 9 (2.3)Others 14 (8.8) 52 (13.4)Not assessable 1 (0.6) 0 (e)Interval from surgery to proposal

(days)19.4 20.7 0.57

Early stage (%) 0.81Yes 148 (93.1) 365 (94)No 11 (6.9) 23 (6)Type of study (%) <0.001Cognitive 36 (22.6) 175 (45.1)Studies on Interventional

innovative therapeutics20 (12.6) 95 (24.5)

Diagnostic/prognostic biology 83 (52.2) 53 (13.7)Imaging 4 (2.5) 9 (2.3)Radiotherapy 16 (10.1) 56 (14.4)

Table 2Univariate analysis: patient and study characteristics according to a proposal or noproposal to participate in a trial.

ProposalN ¼ 215

No proposalN ¼ 332

p Value

Age (mean) 53.9 55.7 0.06Performance status 0.730 105 (48.8) 185 (55.7)1 3 (1.4) 9 (2.7)2 1 (0.5) 2 (0.6)Not assessable (%) 106 (49.3) 136 (41)Primary care providers 0.58Oncologists 171 (79.5) 258 (77.8)Surgeons 12 (5.6) 26 (7.8)Radiologists 6 (2.8) 7 (2.1)Others 25 (11.6) 41 (12.3)Not assessable 1 (0.5) 0 (e)Interval from surgery to proposal

(days)19 () 21.2 () 0.32

Diagnosis 0.841 195 (90.7) 304 (91.6)2 20 (9.3) 28 (8.4)Early stage 0.81Yes 205 (95.3) 308 (92.8)No 10 (4.1) 24 (7.2)Type of study <0.001Cognitive 42 (19.5) 169 (50.9)Studies on interventional

innovative therapeutics42 (19.5) 73 (22)

Diagnostic/prognostic biology 97 (45.1) 39 (11.7)Imaging 6 (2. 8) 7 (2.1)Radiotherapy 28 (13.1) 44 (13.3)

C. Mazouni et al. / The Breast 23 (2014) 170e174172

Patient inclusion

The results of the univariate analyses concerning factors pre-dictive of non-enrolment in a study are summarized in Table 3.Significant differences between professional categories were foundamong patients enrolled in a trial (p ¼ 0.06). Lower-ranked non-manual workers were more frequently enrolled. Significant differ-ences between enrolled and non-enrolled patients were found forthe type of clinical trial. More biological and fewer interventional orcognitive studies were selected for enrolled patients.

In Table 4, we analysed the characteristics of patients who hadbeen proposed a trial but had finally not been enrolled and thosewho had been enrolled. The professional status (p ¼ 0.03) and typeof clinical trial (p ¼ 0.0007) were significant factors distinguishingthe two groups in the univariate analysis.

Physician motivations

Only 17/40 physicians (42.5%) of the breast group replied to theanonymous questionnaire. The frequency of proposals to partici-pate in a trial once it had been identified by the MTMwas: never in11.1%, rarely in 38.9%, sometimes in 33.3%, and often in 16.7%. Themain reasons for physician failure to propose a trial were: a lack oftime in 76.3%, a lack of knowledge concerning the study protocol in70.6%, the study was considered uninteresting in 70.6%, no help forconsent information in 70.6%, an insufficient benefit for the patientin 70.63%, and “fear that the patient would no longer have confi-dence in them” was the reply given by a small percentage of phy-sicians (23%).

Discussion

During the last decade, research on diagnostic and therapeuticstrategies has led to clinical trials being increasingly proposed forBC management. As a consequence there has been a shift in deci-sion making, from a sole practitioner to a shared decision by a

specialist team of patient primary care providers. Clinical trials aimto determine the efficacy of diagnostic or therapeutic strategies, butalso compete with clinical practice guidelines. This might be areason why medical practitioners do not feel committed to theMTM decision which implies modifying their standard protocols.Here, we determined obstacles to proposing a trial identified by theMTM and to ultimate inclusion of patients in these clinical trials.

In our series we found that after MTM counselling, less than 40%of the pre-screened patients had been recruited for participation ina clinical trial. The inclusion rate we observed is consistent withthose reported in previous studies on oncology trials [3,4]. How-ever, compliancewithMTM decisions varies between countries andspecialties [8,9]. Also, in a recent analysis of 2059 trials, Dogan et al.reported a decrease in incidence of drug trials in favour to symp-toms management studies. Thus BC trial participants representedmerely 3.2% of treated BC patients [10,11]. This low enrolment ratecasts doubts on physician compliance with MTM decisionsregarding treatment selection. This lack of recruitment flies in theface of the recognized beneficial role of MTM. Indeed, previousreports demonstrated that the MTM exerted a positive impact onthe improvement of care and survival [1,2]. Although recent con-tradictory data arise about the relevance of tumour boards on theimprovement of cancer care [12,13].

In our analysis, only the type of trial was found to influence theclinician’s decision to propose a trial to patients, whereas agewas ofborderline significance. Thus, a higher rate of diagnostic/biologystudies was observed in the group of enrolled patients, whereas therate of cognitive trials was lower. The absence of a direct benefit forpatients might be an argument in favour of practitioners and

Table 4Patient and study characteristics.

Proposal eno inclusionN ¼ 65

Proposal &inclusionN ¼ 150

p Value

Age (mean) 54.6 52.3 0.13Professional status 0.03Farmers 0 (e) 1 (0.7)Chief executives, managers,

professionals and self-employed4 (6.2) 1 (0.7)

Executives and intellectual professions 5 (7.7) 16 (10.7)Intermediate non-manual workers 16 (24.6) 20 (13.3)Lower non-manual workers 25 (38.5) 49 (32.7)Manual workers 1 (1.5) 2 (1.3)Retired 8 (12.3) 42 (28)Unemployed 5 (7.7) 14 (9.3)Not assessable 1 (1.5) 5 (3.3)Performance status 0.240 42 (64.6) 100 (66.7)1 4 (6.2) 3 (2)2 0 (e) 1 (0.7)Not assessable (%) 6 (9.2) 46 (30.6)Primary care providers 0.55Oncologists 49 (75.4) 122 (81.3)Surgeons 3 (4.6) 9 (6)Radiologists 2 (3.1) 4 (2.7)Others 11 (16.9) 14 (9.3)Not assessable e 1 (0.7)Interval from surgery to proposal

(days)18.4 19.3 0.75

Early stage (%) 0.71Yes 63 (96.9) 142 (94.7)No 2 (3.1) 8 (5.3)Type of study (%) 0.0007Cognitive 12 (18.5) 30 (20)Studies on interventional

innovative therapeutics22 (33.8) 20 (13.3)

Diagnostic/prognostic biology 17 (26.1) 80 (53.3)Imaging 2 (3.1) 4 (2.7)Radiotherapy 12 (18.5) 16 (10.7)

C. Mazouni et al. / The Breast 23 (2014) 170e174 173

patients refusing enrolment. The reason for physician failure topropose a trial did not appear in our records because this type ofinformation was obtained retrospectively for the present study,through a self-administered questionnaire. Previous studies re-ported patient-related factors, health care professional-relatedfactors, research- and investigator-related factors as determinantsfor inclusion in clinical trials [14].

Interestingly, our analysis on factors concerning inclusion weremore informative. Thus factors predictive of non-inclusion byphysicians were the patient’s professional status (p ¼ 0.03) and thetype of study (p ¼ 0.0007). The highest inclusion rates were amongretired persons and those in the executive category (84 and 76%respectively). Thus, diagnostic/prognostic biology studies appearedto be more attractive to physicians and/or patients, while accrual incognitive trials was lower. Previous reports have underlined thediscrepancy between the low rate of participation in trials and thelimited inconvenience or risk to patients [15,16].

We acknowledge that our study has several limitations. Inparticular, the retrospective design limits the interpretation ofreasons for physician failure to propose a trial or reasons for patientrefusal to participate. However, the results of the anonymousquestionnaire among MTM physicians in our department showedthat the reasons for failure to propose a trial were time constraints,the clinician’s opinion regarding the interest of the study in itself aswell as in terms of patient benefit. Familiarity with the proposedprotocols might have influenced the physician’s decision since inour institution some of medical oncologists and radiotherapists arepart-time practitioners. In previous papers, research physicians, aswell as physicians in cancer centres or university hospitals were

more likely to refer patients for research trials [17]. Probably in thefuture the training of physicians could be proposed to improveaccrual. However, the benefit of such training is still to be defined.In a recent randomized trial the effect of additional communicationtraining is not encouraging since no did not appear to benefit theirpatients [18]. In addition, underrepresentation of surgical or radi-ation trials limits the extrapolation of our findings to all oncologysettings, and might bias our results. Despite these limitations, thehigh rates of failure to propose trials may indicate a need forintervention among primary care providers to increase compliancewith MTM decisions and enrolment in clinical trials. Educationalprograms for patients, as well as the implication of nurses could bepotential solutions. Cost-related reasons were not an issue in thisstudy because clinical trials are covered by health insuranceschemes in France. A comprehensive approach to clinical researchinvolving all practitioners involved in BC management might leadto greater interest in research and improve commitment to trials.Greater participation in tumour board meetings might alsocontribute to higher accrual in trials [19]. In addition, expertise inclinical research and the habit of enrolling the patients of theclinical investigator proposing the trial is a non-negligiblecontributing factor to high accrual rates [16].

Conclusions

In conclusion, we found low commitment to MTM proposalsconcerning patient inclusion in BC clinical trials. The proportion ofpatients proposed and ultimately included was related to the typeof clinical trials. The patient’s professional status also influencedthe decision to propose a trial. Future directions should involvegreater recognition of non-clinicians (research nurses) to improvepatient accrual. Educative measures mostly directed at physicianscould also be implemented to overcome such poor compliance.

Conflict of interest statement

There are no financial disclosures from any authors and noconflict of interest.

Acknowledgement

The authors thank Lorna Saint Ange for editing.

References

[1] Patkar V, Acosta D, Davidson T, Jones A, Fox J, Keshtgar M. Cancer multidis-ciplinary team meetings: evidence, challenges, and the role of clinical decisionsupport technology. Int J Breast Cancer 2011;2011:831605.

[2] Kesson EM, Allardice GM, George WD, Burns HJ, Morrison DS. Effects ofmultidisciplinary team working on breast cancer survival: retrospective,comparative, interventional cohort study of 13 722 women. BMJ 2012;344:e2718.

[3] Go RS, Frisby KA, Lee JA, Mathiason MA, Meyer CM, Ostern JL, et al. Clinicaltrial accrual among new cancer patients at a community-based cancer center.Cancer 2006;106(2):426e33.

[4] Lara Jr PN, Higdon R, Lim N, Kwan K, Tanaka M, Lau DH, et al. Prospectiveevaluation of cancer clinical trial accrual patterns: identifying potential bar-riers to enrollment. J Clin Oncol 2001;19:1728e33.

[5] Mills EJ, Seely D, Rachlis B, Griffith L, Wu P, Wilson K, et al. Barriers toparticipation in clinical trials of cancer: a meta-analysis and systematic reviewof patient-reported factors. Lancet Oncol 2006;7:141e8.

[6] Ford JG, Howerton MW, Lai GY, Gary TL, Bolen S, Gibbons MC, et al. Barriers torecruiting underrepresented populations to cancer clinical trials: a systematicreview. Cancer 2008;112:228e42.

[7] R Development Core Team. R: a language and environment for statisticalcomputing. Vienna, Austria: R Foundation for Statistical Computing; 2010,ISBN 3-900051-07-0; 2010. http://www.R-project.org/.

[8] Saini KS, Taylor C, Ramirez AJ, Palmieri C, Gunnarsson U, Schmoll HJ, et al. Roleof the multidisciplinary team in breast cancer management: results from alarge international survey involving 39 countries. Ann Oncol 2012;23:853e9.

C. Mazouni et al. / The Breast 23 (2014) 170e174174

[9] Stewart JH, Bertoni AG, Staten JL, Levine EA, Gross CP. Participation in surgicaloncology clinical trials: gender-, race/ethnicity-, and age-based disparities.Ann Surg Oncol 2007;14:3328e34.

[10] Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials:race-, sex-, and age-based disparities. J Am Med Assoc 2004;291:2720e6.

[11] Chow CJ, Habermann EB, Abraham A, Zhu Y, Vickers SM, Rothenberger DA,et al. Does enrollment in cancer trials improve survival? J Am Coll Surg2013;216:774e80.

[12] Blayney DW. Tumor boards (team huddles) aren’t enough to reach the goal.J Natl Cancer Inst 2013;105:82e4.

[13] KeatingNL, LandrumMB, LamontEB, BozemanSR, ShulmanLN,McNeil BJ. Tumorboards and the quality of cancer care. J Natl Cancer Inst 2013;105:113e21.

[14] Baquet CR, Henderson K, Commiskey P, Morrow JN. Clinical trials: the art ofenrollment. Semin Oncol Nurs 2008;24:262e9.

[15] Pentz RD, Young LN, Amos CI, Hess KR, Wei Q, Anderlik MR. Informed consentfor tissue research. J Am Med Assoc 1999;282:1625.

[16] Rasco DW, Xie Y, Yan J, Sayne JR, Skinner CS, Dowell JE, et al. The impact ofconsenter characteristics and experience on patient interest in clinicalresearch. Oncologist 2009;14:468e75.

[17] Klabunde CN, Keating NL, Potosky AL, Ambs A, He Y, Hornbrook MC, et al.A population-based assessment of specialty physician involvement in cancerclinical trials. J Natl Cancer Inst 2011;103:384e97.

[18] Bernhard J, Butow P, Aldridge J, Juraskova I, Ribi K, Brown R. Communicationabout standard treatment options and clinical trials: can we teach doctorsnew skills to improve patient outcomes? Psychooncology 2012;21:1265e74.

[19] Minasian LM, O’Mara AM. Accrual to clinical trials: let’s look at the physicians.J Natl Cancer Inst 2011;103:357e8.