A systematic review of reasons for nonentry of eligible patients into surgical randomized controlled trials

Clinical review A systematic review of reasons for nonentry of eligible patients into surgical randomized controlled trials Ned S. Abraham, MBBS (Hons), MM (Syd), FRACS, FRCS (Engl),a,b Jane M. Young, MBBS, MPH, PhD, FAFPHM,b,c and Michael J. Solomon, MBBCh (Hons), MSc (ClinEpid), FRACS,b,c Coffs Harbour and Sydney, Australia

Background. The low recruitment rates into surgical randomized controlled trials (RCTs) threaten the validity of their findings. We reviewed the reasons for nonentry of eligible patients into surgical RCTs that would form the basis for future prospective research. Methods. A systematic review of the English language literature for studies reporting reasons for nonentry of eligible patients into surgical RCTs and of recommendations made to improve the low recruitment rates. Results. We reviewed 401 articles, including 94 articles presenting the results of 62 studies: 23 reports of recruitment into real surgical RCTs, 11 surveys of patients regarding hypothetical surgical RCTs, 10 surveys of clinicians and 18 literature reviews. The most frequently reported patient-related reasons for nonentry into surgical RCTs were preference for one form of treatment, dislike of the idea of randomization, and the potential for increased demands. Distrust of clinicians caused by a struggle to understand, explicit refusal of a no-treatment (placebo) arm, and the mere inability to make a decision were frequently reported in studies of real RCTs and patient surveys, but were not emphasized in surveys of clinicians and review articles. Difficulties with informed consent, the complexity of study protocols, and the clinicians’ loss of motivation attributable to lack of recognition were the most commonly reported clinician-related reasons. Conclusions. There seems to be a discrepancy between real reasons for nonentry of eligible patients into surgical RCTS and those perceived by the clinicians, which require further prospective research. A summary and discussion of main recommendations sighted in the literature is presented. (Surgery 2006;139:469-83.) From the Coffs Harbour Health Campus, Faculty of Medicine, The University of New South Wales,a Coffs Harbour; the Surgical Outcomes Research Centre (SOuRCe), Sydney South West Area Health Service, The University of Sydneyb; and the Royal Prince Alfred Hospital c

INABILITY TO RECRUIT an adequate number of participants is arguably the greatest threat to the success of surgical randomized controlled trials (RCTs).1,2 The external validity of the findings of such RCTs may be compromised severely by selective participation if patients who agree to participate differ systematically from the underlying patient population.3 Accepted for publication August 25, 2005. Reprint requests: Dr Ned Abraham, Senior Lecturer (Surgery), Coffs Harbour Health Campus, Faculty of Medicine, The University of New South Wales, Locked Bag 812, Coffs Harbour, NSW Australia 2450. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2005.08.014

Others have argued that patient selection in surgical RCTs is unpredictable,4 which would make it difficult to foresee difficulties. In 1991, Benson5 reported that less than 3% of all eligible patients entered RCTs. Ten years later, Haidich and Ioannidis6,7 reported that in excess of 17% of medical RCTs never reached 50% of their target recruitment. Although not reported in most published articles,8 recruitment rates in surgical RCTs are thought generally to be less than 50% of eligible patients.9 Although there is no consensus as to what would constitute an acceptable threshold for the recruitment percentage of eligible patients for RCTs, such a figure probably would differ for different trials and would depend on sample size and difference SURGERY 469

470 Abraham, Young, and Solomon

in magnitude of measured effects between the 2 assessed treatment modalities. The use of tables of misclassification and systematic errors could help illustrate that. For these reasons, a need exists currently for a thorough review of reported reasons for nonentry of eligible patients into established and hypothetical surgical RCTs. This review would form the basis for future prospective research into studying true rather than perceived reasons for nonentry of eligible patients into such trials and for exploring methods that could improve the low recruitment rates. METHODS The aim of the current study was to systematically review the literature for reported reasons for nonentry of some eligible patients or potential patients into true or hypothetical surgical RCTs to ascertain characteristics of eligible patients who are most likely to decline entry into a trial and recommendations made to improve the low recruitment rates. A surgical RCT for the purpose of the current study was defined as one that dealt with the prevention, detection, or management of a surgical condition. The English language literature was searched by using CINAHL, EBM, EMBASE, and MEDLINE databases for articles reporting actual or perceived reasons for nonentry of eligible patients into surgical RCTs, which were published by the end of 2003. These articles were in the form of reports of real surgical RCTs; surveys of patients, potential patients, or clinicians about entry into or willingness to participate in real or hypothetical surgical RCTs; or reviews of reasons for nonentry of eligible patients into such RCTs. Studies also were included if they reported recommendations to improve accrual rates into surgical RCTs. Reference lists of identified articles were searched for other relevant publications (pearling, the ancestry approach or citation chasing).10 Reasons for nonentry of eligible patients in surgical RCTs were grouped into patient-related, clinician (practice)-related, and other.11 Pooled analysis and quantitative ranking were not possible because of the wide variability of the methods used for data collection and presentation. Reasons for nonentry of eligible patients in real surgical RCTs were ranked first, followed by those reported in surveys of real or potential patients about their willingness to participate in hypothetical surgical RCTs, then by those perceived by clinicians and reported in surveys or review articles. Retrieved articles also were reviewed for

Surgery April 2006

strategies adopted or recommended by their authors to improve recruitment rates in surgical RCTS. RESULTS Search results. The literature search detected 6348 potentially relevant publications. All abstracts were reviewed. Reference was made to nonentry of eligible patients in RCTs in 378 abstracts; these were selected for article retrieval. Pearling helped identify a further 23 articles. In total, 401 articles were reviewed. Ninety-four articles presenting the results of 62 studies fulfilled the inclusion criteria: 51 articles presented the results of 23 studies of nonentry of eligible patients into real surgical RCTs; 11 articles, the results of a similar number of surveys of patients and potential patients about willingness to participate in hypothetical surgical RCTs; 14 articles, the results of 10 surveys of clinicians about their attitudes towards recruitment into surgical RCTs and reasons some eligible patients are not enrolled into such trials; and 18 review articles (Figure). To avoid unnecessary repetition, we presented the results as studies performed rather than articles published; only the most relevant articles were cited whenever a study was reported in more than 1 article. Studies of the effect of specific interventions on recruitment rates. Five studies reported the effect of specific interventions on recruitment of eligible patients into established surgical RCTs.1,12-15 The protocol of an RCT of arthroscopy for osteoarthrosis of the knee threatened by a very low recruitment rate was changed to incorporate Zelen’s prerandomization technique, in which patient eligibility is ascertained, followed by random treatment allocation, then the patient’s consent for the trial and the allocated treatment is sought.1 The incorporation of the technique was associated with a 6-fold increase in the accrual rate. In another study, patients enrolled in an RCT of the role of carotene and retinol in the chemoprevention of lung cancer were randomized to receiving or not receiving 1 of 2 incentives or both to remain in the trial.12 This strategy had no statistically significant effect on the retention of patients in the RCT at the end of 2 years of follow-up. In the third study,13 44 institutions were assigned randomly to provide an easy-to-read consent statement or a standard form to 207 patients eligible for selected parent RCTs of management of various malignant conditions. The easy-to-read consent statement was associated with significantly less anxiety and higher satisfaction rates, but there was no statistically significant difference between the 2 groups in accrual

Surgery Volume 139, Number 4

Abraham, Young, and Solomon 471

Figure. Summary of the literature search results for studies of reasons for nonentry of eligible patients in surgical RCTs.

rates or levels of patient comprehension. In the fourth study,14 100 patients eligible for enrolment into an RCT of operation versus radiotherapy versus monitoring for localized prostate cancer were randomized to being consented by a nurse or a surgeon. The proportion of eligible patients recruited was 69%. There was a small but statistically insignificant difference in the recruitment rates by nurses versus surgeons. Nurses spent longer periods of time with patients and at a lower cost than surgeons. The fifth and final study15 was a comparison between a direct and culturally specific approach through ethnic community centers and gatherings (religious or otherwise) including using the mother tongue with an indirect approach through the use of newspaper announcements to recruit elderly males into an RCT of methods to prevent surgical tooth extraction in 9 community centers. The direct approach was more successful in recruiting elders from ethnic minority groups and much less expensive than the indirect approach, which was adequate only for recruiting elders from the dominant culture. Surveys of patients invited to participate in real surgical RCTs. Four studies reported the results of surveys of eligible patients who had been invited to participate in real surgical RCTs (Table I).15-19 It was reported that although most patients understood clinical equipoise (ie, the situation in which even experts in a field are uncertain as to the best answer to the clinical question), they did not

accept it necessarily.19 Patients also recalled experiencing difficulty understanding the concepts of a randomized trial and having to rely on their clinicians to resolve the conflict.16,17 Under these circumstances, it was suggested that a great deal of trust placed in the clinician was required for a patient to consent to randomization. Concern about being randomized to an inferior treatment, no treatment, or a placebo arm was reported as a very significant reason for declining trial entry among all 4 studies followed by having a preference for one form of treatment. In 1 study,20 no reason was given for declining to participate by more than 40% of patients, which was assumed to reflect the uncertainty felt by the patient. Altruism and having developed a trust in the doctor were the reasons quoted by most patients for agreeing to participate in 3 RCTs.16-18 Surveys of clinicians invited to participate in real surgical RCTs. In contrast, 4 studies were surveys of clinicians who had participated in surgical RCTs about the various barriers and issues that they might have encountered in recruiting eligible patients in these trials (Table II).20-23 Difficulty in obtaining informed consent and concern about the doctor-patient relationship were reported as being the most common reasons for failing to enroll eligible patients in surgical RCTs.22,24 It was also reported that clinicians often feel challenged by the task of having to integrate the traditionally separate and sometimes competing roles of

472 Abraham, Young, and Solomon

Surgery April 2006

Table I. Surveys of patients who were invited to participate in real surgical RCTs (n = 4) Study

Participants

Main findings

Featherstone (16,17) 1998 & 2002 UK

Of 33 eligible patients, 22 had agreed to participate in an RCT of the management of benign prostatic hyperplasia.

Jenkins (18) 2000 UK

Of 204 patients eligible for multiple RCTs of cancer management (mainly breast cancer), 147 (72%) had agreed to participate.

Mills (19) 2003 UK

Of 21 eligible patients, 12 had agreed to participate in an RCT of the management of localized prostate cancer.

Spiro (20) 2000 UK

Follow-up of 253 lung cancer patients was performed to see if they were entered in an RCT of primary treatment ± cisplatinum.

Most patients—regardless of their knowledge and the decision they made—reported a struggle to understand the concept of randomization. The decision made was determined by whether trust or mistrust was placed in the clinician. The most common reason for declining was concern about randomization. For agreeing to participate, altruism followed by trusting the doctor was the most common reason. RCTs providing active treatment in every arm had a higher acceptance rate than those with a no treatment or a placebo arm. Most patients recalled and understood clinical equipoise. Those who found it acceptable agreed to randomization, while those who did not accept or understand it tended to refuse. The recruitment rate was 104 (69%) of 150. One hundred ninety-eight (75%) patients declined enrollment. More than 40% gave no reason for declining, and more than a third declined because they had a preference for one form of therapy (mostly no cisplatinum).

researchers and clinical care providers.22,23 Many clinicians reported a dislike for having open discussions involving uncertainty with their patients.22,24 Other reported reasons for nonentry of eligible patients into surgical RCTs were difficulties in following study protocols and concern about individual patient’s outcome.22-24 One study reported that many clinicians exercised excessive flexibility in following consent procedures that probably amounted to inappropriateness; such flexibility included bridging the rules of medical ethics by not informing patients that their treatment allocation was random or entering them into a trial without informed consent.24 Organizational and institutional characteristics (eg, supportive staff, financial support, office space) were found to play a more important role in successful recruitment than recruitment strategies (eg, incentives for the patients, method of randomization, and consent procedures).21 However, only a minority of organizational characteristics were found to be evidence-based.21 Comparisons between recruits and eligible nonrecruits in surgical RCTs. Four studies reported detailed comparisons between eligible patients recruited and those not recruited into surgical RCTs (Table III).24-27 There was no strict agreement between these studies, but, compared with enrolled patients, patients who refused to

participate generally tended to be older, of racial minority groups, unmarried, less well-educated, and less likely to reside near study centers. However, in 1 study, tertiary education impacted negatively on recruitment into the RCT.26,27 Some studies also found that patients who refused to participate were likely to be more ill, or have more pain or more extensive disease, or, compared with enrolled patients, their condition interfered with their daily life more,25-27,29 while self-reported physical functioning and mental health status, employment status, cancer history, treatment plan, or previous research participation did not have a significant impact on recruitment.28 In contrast, 6 other studies were nonquantitative reports of the difficulties encountered by clinicians in recruiting eligible patients into real surgical RCTs (Table IV).1,28-32 These difficulties could be summarized in patients’ preferences, unfavourable treatment or intervention option, restricted eligibility criteria, difficulty in obtaining informed consent, complexity of the study design, differences in commitment and resources between study centers, and competing trials. Surveys of patients or potential patients about willingness to participate in hypothetical surgical RCTs. Eleven studies presented the results of surveys of close to 4500 patients or potential patients

Abraham, Young, and Solomon 473

Surgery Volume 139, Number 4

Table II. Surveys of clinicians who participated in real surgical RCTs (n = 4) Study

Participants

Main findings

Foy (21) 2003 UK

Investigators in 7 study centers aimed to recruit 6070 patients into a gastrointestinal RCT.

Taylor (22) 1984 Canada

Ninety-four clinicians participated in a multicenter international RCT of the management of early breast cancer.

Taylor (23) 1992 Canada

One hundred one clinicians were considering participation in an RCT of the management of ocular melanoma.

Williams (24) 1994 UK

Sixty-six clinicians involved in 12 RCTs of the management of different malignant conditions were surveyed anonymously.

Organizational characteristics play a more important role in recruitment than recruitment strategies: 63% of 56 recruitment strategies and 17% of 29 organizational characteristics were found to be evidence-based. The most commonly reported reasons for failing to enroll eligible patients were concern about the doctor-patient relationship (73%), difficulty with informed consent (38%), dislike of discussions involving uncertainty (22%), conflict between the roles of a scientist and a clinician (18%) and practical difficulties with protocol (9%). The main challenge for clinicians considering recruitment of patients into RCTs is the need to integrate the formerly separate and sometimes competing roles of a researcher and a clinician (2 articles). Consent procedures were the most common reason for failure to enrol. Only 32% of those surveyed obtained written consent: 21% did not obtain a signature; 42% obtained verbal consent only; and 5% did not obtain consent at all. Twelve percent did not tell patients about the RCT before randomization; 42% provided patients with information about the allocated treatment only, and 38% did not reveal the trial was randomized.

Table III. Recruits versus eligible nonrecruits in real surgical RCTs (n = 4) Study

Outline

Main findings

CASS (25) 1984 USA

Retrospective comparison between recruits and eligible nonrecruits in an RCT of surgery for ischemic heart disease: 780 (59%) of 1315 eligible patients were recruited.

COMS (26,27) 2001 USA

Retrospective comparison between recruits and eligible nonrecruits in an RCT of the management of ocular melanoma: 1326 (46%) of 2882 eligible patients were recruited.

Sears (28) 2003 USA

A report on potential factors influencing recruitment in an RCT of psychoeducational intervention in the management of early breast cancer: of 1323 eligible patients, 42% were successfully recruited.

Yuasa (29) 2003 Japan

Retrospective comparison between recruits and eligible nonrecruits in an RCT of the management of painful conditions of the mandible.

Nonrecruits treated medically had less extensive disease than those treated surgically. The situation was reversed among recruits. It is uncertain if this was indicative of certain biases in the recruitment pattern (8 articles). Eligible nonrecruits had worse visual acuity in the other eye, were less likely to reside near study centers, and more likely to have tertiary education than recruits. The 2 groups were similar otherwise (8 articles). Eligible nonrecruits were more likely to be unmarried, less educated, older, and of racial minority groups than recruits, but there was no difference in physical functioning, mental health status, employment, cancer history, treatment plan, or previous research participation. Patients refusing to participate had more pain, and their condition interfered with their daily life more seriously than those who agreed.

concerning entry into hypothetical surgical RCTs (Tables V and VI).35-45 In an Australian survey,38 women attending for screening or diagnostic assessment were at least twice as likely to agree to

participate in a hypothetical RCT of the management of breast cancer than patients with the disease. However, in another survey conducted in Japan, patients thought to suffer from malignant

474 Abraham, Young, and Solomon

Surgery April 2006

Table IV. Reports of real surgical RCTs and factors impacting on recruitment (n = 6) Study Coombs (30) 1993 USA Eisinger (31) 2001 France

Outline An RCT of home-based burn treatment modalities that was not completed Review of recruitment rates in an RCT of screening for colorectal cancer at 37 study centers

Goodwin (32) 2000 Canada

An RCT of methods of counselling breast cancer patients

Jack (4) 1990 Scotland

An RCT of conservative management of early breast cancer

Plaisier (33) 1994 Holland

An RCT of lithotripsy vs cholecystectomy. Of 448 eligible patients, 37 (8%) were recruited.

van der Windt (34) 2000 Holland

A report of the difficulties associated with an RCT of the management of stiff shoulder syndrome

head and neck disease did not have a lower acceptance rate than those thought to suffer from benign disease.44,45 Dislike of the idea of randomization and loss of ability to make decisions by the patients were reported as the most common reasons for declining to participate. Many patients reportedly did not understand the concepts of RCTs, were confused about the use of placebo, and were wary about medical research. It also was reported that less knowledgeable or less educated older patients were more likely to decline participation than young, well-educated patients. In 1 survey, patients scored less than 50% on average in a 7-point questionnaire about knowledge of RCTs.37 However, in another Australian survey, women who received an educational booklet were significantly less likely to consider participation in an RCT of the management of breast cancer than those who did not.36 In a hypothetical RCT of the choice of anaesthesia for surgical procedures,42 patients were randomized to 5 methods of consent randomization and consent methods: 1-sided informed consent (standard), prerandomized consent to experimental treatment, prerandomized consent to standard treatment, 1-sided physician-modified informed consent (the clinician has a preference for innovation), or 1-sided patient-modified informed consent (patient is allowed to have a preference for

Main findings The complexity of the study design made it difficult for clinicians to adhere to the protocol and compromised recruitment. Of 850.000 eligible individuals, 0.14% (one sixth of expected) were recruited. The significant variability between centers was arguably attributable to differences in commitment and resources. Of 654 eligible patients, 43% were recruited. Enrollment in competing trials was the most frequent reason for failure to enroll. The most common reason for declining participation reportedly was having a preference for a specific form of therapy. The authors argue that the results of surgical RCTs may be distorted by unpredictable patient selection. Strong patients’ preference for one form of therapy, consent procedures, restricted eligibility criteria, and the innovation of laparoscopic cholecystectomy were the most frequent reasons for failure to recruit (3 articles). Recruitment rate was significantly lower than anticipated. Patients’ preference was quoted as being the most frequent reason for nonentry in the trial.

innovation). This approach had no impact on recruitment rates. In another study, when the results of adjuvant therapy for breast cancer were described verbally, patients were significantly more likely to agree to participate in a hypothetical RCT than when they were described numerically.40 Surveys of clinicians and review articles about participation in surgical RCTs. Fourteen articles reported the results of 10 surveys of in excess of 2000 clinicians about their attitudes toward surgical RCTs and their reasons for not entering some eligible patients into them (Table VII).5,43,46-53 Although the results of these surveys could not be pooled into 1 data set, there was agreement between most of the surveys on reasons perceived by clinicians for nonentry of eligible patients into surgical RCTs and the difficulties encountered in embarking on recruiting patients into such trials. The surveys suggested that many clinicians probably never pursue RCTs and that some actually discourage their patients from entering trials.5 Radiation oncologists, community-based clinicians, and general surgeons reportedly were less likely to recruit patients into RCTs than medical oncologists, hospital-based clinicians, and subspecialized surgeons, respectively.5,47 Others also suggested that the clinician’s reluctance to enter patients in RCTs stems from the difficulty

Abraham, Young, and Solomon 475

Surgery Volume 139, Number 4

Table V. Surveys of patients (real or potential) about entry into hypothetical surgical RCTs (n = 11) Study

Participants

Main findings

Caldwell (35) 2003 Australia

33 parents of pediatric patients

Ellis (36) 2002 Australia

83 breast cancer patients randomized to receiving vs not receiving an educational booklet

Ellis (37) 1999 Australia

60 patients with breast and other malignancies

Ellis (38) 2001 Australia

545 women attending a breast clinic for diagnostic assessment/screening (two thirds) or for the management of breast cancer (one third)

Ellis (39) 1998 Australia

21 healthy mothers and 20 breast cancer patients interviewed through a focus group discussion

associated with the integration of the seemingly conflicting roles of a care provider and that of a research scientist.53 Of the 18 review articles assessed,54-71 there was 1 systematic review57 of barriers to participation in trials by clinicians or patients (rather than of reasons for nonentry of eligible patients in RCTs, surgical or otherwise). Clinician-related reasons for nonentry of eligible patients into surgical RCTs. Significant overlap exists in the literature between reporting clinicians’ participation (or lack of it) in surgical RCTs and the success by those actively involved in established surgical RCTs in recruiting eligible patients. However, several non--patient-related reasons for inadequate recruitment into surgical RCTs could be identified and are listed in Table VIII according to the frequency with which they were quoted as being significant in reports of real surgical RCTs followed by surveys and review articles.1,4,5,11,16-18,20-24,30-35,37,38,40,41,43,46-49,51,54-56,58-68 The most commonly reported clinician-related reasons in the surveys and review articles were the difficult, time-consuming, and uncertain nature

Most would decline because of concerns about side effects, allocation to an ineffective treatment or placebo arm, and the demands of participation. The decision would be based on perceived risks and benefits, meeting others in similar circumstances, altruism, the child’s health status, preference for one form of therapy, and the doctor’s recommendations. The 57% who would decline to participate were more concerned about the negative aspects of trials but less likely to be anxious or to have advanced disease than participants. Adjusting for these variables, the education booklet had a negative impact on willingness to participate. The 66% who would decline to participate perceived the doctor unfavorably and the trial as being experimental, inconvenient, or leading to loss of control. Willingness to participate (33%) correlated with increased knowledge about trials, younger age, the desire to be involved in decision making, and a perceived positive impact of trials. Cancer patients were twice as likely to decline participation as others. Most participants did not appreciate the concept of a trial, were confused about placebo, and would decline because they were wary about medical research and saw it as a gamble with no benefit to the individual patient.

of informed consent,1,11,22,24,33,37,43,51,55,58,62-66 the difficulty or complexity of the RCT protocol,21,22,30,33,49,51,58,63,67 loss of motivation or interest in a trial attributable to lack of recognition or rewards (eg, financial),23,31,34,58 and the clinicians’ attitudes and commitment to research in general.11,22,23,43,49,55,62,63 A summarized list of reported clinician-related and other reasons for nonentry of eligible patients into surgical RCTs. 1. Difficulty in following complex study protocols and follow-up requirements 2. Preference for one form of therapy or specific dislike for another 3. Difficulties associated with consenting a patient for an RCT a. The time-consuming nature of the informed consent procedures b. Difficulty explaining procedures involved c. Potential incompatibility with professional obligations towards patients d. Loss of professional autonomy e. Increased accountability

476 Abraham, Young, and Solomon

Surgery April 2006

Table VI. Surveys of patients (real or potential) about entry into hypothetical surgical RCTs (n = 11) Study

Participants

Fetting (40) 1989 USA

282 breast cancer patients randomized to 2 methods of presenting RCT information

Maslin (41) 1994 UK

300 breast cancer patients

Myles (42) 1999 USA

770 surgical patients randomized to 5 recruitment approaches for a hypothetical RCT of anesthesia for their procedure

Solomon (43) 2003 Australia

200 newly diagnosed colorectal cancer patients

Yuasa (44) 2002 Japan

1079 patients with malignant head and neck disease

Yuasa (45) 1999 Japan

1083 patients with benign head and neck conditions.

4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

f. Effect of the uncertainty of what patients are told on the doctor-patient relationship g. Concern about truly informed consent for non– English-speaking patients The imposition of strict eligibility criteria Time constraints The expense and financial implications for the clinician and the institution Lack of resources, equipment, and administrative and support staff The clinician’s general orientation and attitude towards research Issues related to specific trials (eg, relevance) Concern for the individual patient’s outcome and conflicts of priority Difficulty in identifying relevant RCTs Lack of an adequate process of invitation to participate in RCTs Medicolegal considerations

Patient-related reasons for nonentry of eligible patients into surgical RCTs. As mentioned above, there was not enough similarity in the methodology of the studies or the way data were presented to allow for pooled analysis of outcome data. However, reported patients’ reasons for refusing to enroll in surgical RCTs, whether real or hypothetical, were extracted and summarized in Table VIII. A preference for one form of

Main findings Participation was more likely if the expected results had been described verbally (52%) than numerically (35%). Predictors of participation were young age, college education, previous participation, and high expectations. Dislike of the idea of randomization, loss of ability to make decisions, and reluctance to being a ‘‘guinea pig’’ were the common reasons for lack of willingness to participate. Fifty-four percent would decline participation. The recruitment method did not affect willingness to participate. Predictors of recruitment were patient age >45 years, English-speaking at home, male researcher-male patient interaction, and high expectations of innovation. Fifty-two percent would decline to participate. There was no significant association between patient characteristics and willingness to participate. The important barriers to participation were risk of cancer recurrence, dislike of the idea of randomization, and concern about quality of life issues. Eighty-eight percent would decline to participate. Patients with malignant disease did not have a lower acceptance rate than those with benign disease (see also ref 45). Eighty-nine percent would decline to participate. The difference in the willingness rate between university and general hospital patients was statistically insignificant.

treatment and dislike of the idea of randomization were the 2 most commonly reported reasons for nonentry of eligible patients into surgical RCTs.4,18,20,33,34,38,40,41,43,47,49,54-60,63 The results of comparisons made between enrolled and nonenrolled eligible patients and reported in RCTs and surveys suggest that the patient likely to refuse participation in a surgical RCT seems to be the older, single, male patient who comes from an ethnic minority group, is in need of an interpreter, resides away from study centers, is not in a white collar occupation, is less educated or less knowledgeable about RCTs, is doubtful of the benefits of an intervention or the risks of inaction, and has no previous participation in RCTs (Table IX).15,25,26,28,29,38,40,42-44,48 A summarized list of reported patient-related reasons and barriers to entering surgical RCTs. 1. Having a preference for one form of therapy 2. Difficulty understanding the idea of a trial 3. Anxiety/unhappiness about the concept of randomization 4. Concern about a negative outcome or the risk of disease (cancer) recurrence 5. Ill health 6. Advanced age 7. Excessive travel requirements 8. Inadequate health insurance

Abraham, Young, and Solomon 477

Surgery Volume 139, Number 4

Table VII. Surveys of clinicians about low recruitment rates in surgical RCTs (n = 10) Study

Participants

Benson (5) 1991 USA

244 surgeons, oncologists, and radiation oncologists

Deber (46) 1990 Canada

130 breast cancer clinicians

Ellis (47) 1999 Australia

269 breast cancer clinicians

Foley (48) 1991 USA

210 clinicians treating different malignancies

Langley (49) 2000 UK

Audio-taped and fully transcribed semistructured interviews with a selected sample of 20 clinicians

Martin (50) 2003 USA

201 surgical oncologists and general surgeons who graduated between 1985 and 1999

Penn (51) 1990 UK

Consultants in 36 hospitals about elective caesarean section for the preterm breech infant

Solomon (43) 2003 Australia

219 clinicians as if they themselves were cancer patients

Taylor & Kelner (52) 1987 Canada

170 selected breast cancer clinicians from 8 countries

Taylor (53) 1987 Canada

484 selected breast cancer clinicians from 57 institutions in 5 countries

Main findings The common barriers were rigid study protocols, conflict of priorities in the individual patient’s care, and excessive followup requirements. A third of the 244 clinicians never pursued RCTs, and 14% discouraged their patients from entering them. Surgeons and oncologists were more likely than radiation oncologists to pursue RCTs. Community clinicians were more likely to have concerns about informed consent and financial implications and less likely to recruit patients than were hospital-based clinicians. The clinician’s general orientation towards research, preference for one form of therapy, or dislike for another were the main barriers to recruiting patients into RCTs Medical oncologists and surgeons were about 50% more likely than radiation oncologists to have participated in RCTs. The main reported barriers to participation were lack of resources and specific trial issues (eg, relevance) (2 articles). The commonly reported patient-related barriers were ill health, advanced age, excessive travel requirements, and inadequate health insurance. The commonly reported clinician-related barriers were complexity of the protocol, cost, lack of time, and inadequate administrative support. Difficulty in identifying key relevant RCTs often exists. Recruitment into such trials is hindered by time constraints, difficulty in following protocols, attitude to research, concern for the individual patient, patients’ preference, strict eligibility criteria, and the expense of follow-up. Trained surgical oncologists were more likely than general surgeons to utilize RCTs in clinical decision making and to participate in RCTs. The most commonly reported reasons for nonparticipation in RCTs were absence of an active process of invitation to participate and the lack of time. The time-consuming nature of consent procedures, providing information to non-English speaking mothers, anxiety about the availability and skill of attending personnel, difficulties with protocol details, concern about staff support, and medicolegal considerations were the reasons for reluctance to participate. Oncologists (63%) were more likely than surgeons (44%) to agree to participate in RCTs. Risk of cancer recurrence was the most commonly reported factor that would determine the decision to participate (3 articles). Consenting a patient for an RCT was reported to involve possible incompatibility with professional obligations toward the patient, loss of professional autonomy, increased accountability, and the effect of what trial patients are told on the doctor-patient relationship. Difficulty arises from trying to integrate the physician’s separate and seemingly conflicting roles of a health care provider and a research scientist (2 articles).

478 Abraham, Young, and Solomon

Surgery April 2006

Table VIII. Reasons for nonentry of eligible patients into surgical RCTs Reason Patient-related reasons A preference for one form of treatment Dislike of the idea of randomization Potential for increased demands on patients Distrust of the clinicians often caused by a struggle to understand Explicit refusal of placebo or a no intervention arm Concern about a negative outcome Loss of ability to make decisions or reluctance to become a ‘‘guinea pig’’ No reason given Clinician-related and other reasons Informed consent difficult, time-consuming and surrounded by uncertainty The difficulty in following complex study protocols Lack or loss of motivation or interest due to lack of recognition or rewards The clinician’s attitude and commitment to research in general Lack of resources and administrative support Existing practice commitments and lack of time Concern about negative outcomes and related medicolegal issues Concern about compromising the doctor-patient relationship Loss of autonomy Competing trials Ethics of randomization if the surgeon has a preference for an intervention Special groups (children, ethnic minorities, underdeveloped countries and the mentally ill)

Real RCTs

Hypothetical RCTs Clinicians’ surveys & reviews

4, 20, 33, 34, 54 18 54 16

38, 40,43 38, 41, 56 35 17

18

35

11, 43, 47, 49, 55, 56, 58-60 11, 55, 56, 58, 63 58, 63

35, 38, 43 35, 41

43, 46, 55, 58, 61

1, 22, 24, 33

37, 51, 62

11, 43, 55, 58, 63-66

22, 30, 33

49, 51

21, 58, 63, 67

20

23, 31, 34

58

22, 23

49, 62

11, 43, 55, 63

31 31 22

51, 52, 62 51, 52 48, 51

21, 43, 58, 63, 67 11, 43, 55, 58, 63, 67 5, 49, 58

22

5, 43, 55, 58

22 32

11, 43, 58, 63

Reported recommendations to improve participation in RCTs. The recommendations made to improve recruitment into RCTs could be divided into trial design and conduct issues (including institutional issues) and clinician-related issues. These recommendations are summarized in Table X,1,6,12-18,21,23,31,32,35,42,43,45,48,50,53,55,58,63,67,69-73 but only a minority of them have been tested in an experimental setting. The recommendations most commonly made in relation to design and conduct issues were multidisciplinary collaborations, improved communication between study headquarters and study centers6,21,32,48,55,58,67,69 and adequate and continued support for clinicians (eg, data managers).23,43,55,58,67 As far as recruitment strategies were concerned, the recommendations most commonly made were choosing an appropriate method of invitation, using targeted recruitment policies (eg, for ethnic groups) and evidence-based

43, 49, 51

11, 43, 55, 61, 63 64-66, 68

assessment of recruitment strategies,15,21,67 using prerandomization screening,31,67 Zelen’s prerandomization,1,71,72 providing improved knowledge and information to patients,16,17,35,63,67 and taking measures to gain patients’ trust and relieve their anxiety.16-18,63 The clinician-related recommendations made most often were choosing clinicians who have shown enthusiasm about research, have appropriate orientation profile (experimenter/ therapist), and are well trained,23,50,53,55,58 running regular workshops, distributing newsletters and trial updates at scientific meetings,67 and involving the primary health care providers.67 DISCUSSION This study has shown that the 3 most commonly reported patient-related reasons for nonentry of eligible patients into surgical RCTs were preference

Abraham, Young, and Solomon 479

Surgery Volume 139, Number 4

Table IX. Characteristics of patients, nature and extent of disease, and severity of symptoms that may impact negatively on recruitment into surgical RCTs Characteristic Patient characteristics Single Older Younger Ethnic minority group/need for an interpreter Less educated or less knowledgeable of RCTs, the benefits of intervention or the risks of inaction More educated Non-white collar occupation Residing away from study centres No previous participation in a trial Ill-health Inadequate health insurance No correlation found More extensive disease or more severe symptoms No correlation found Malignant versus benign disease No correlation found

for one form of treatment, dislike of the idea of randomization and the potential for increased demands on the patient. Four other patientrelated reasons for failure to enroll in surgical RCTs were emphasized in reports of real RCTs and in surveys of patients or potential patients, but not in surveys of clinicians and review articles. These 4 reasons were distrust of the clinicians often caused by a struggle to understand, explicit refusal of placebo, discomfort with loss of ability to make decisions (being a ‘‘guinea pig’’), and refusing for no reason other than the inability to make a decision. This last ‘‘reason’’ was reported in 1 study to have influenced 41% of patients refusing to participate in a surgical RCT. These findings imply that clinicians’ surveys and reviews may not necessarily be the most reliable way to determine patientrelated reasons for nonentry into surgical RCTs. In one study, clinicians were surveyed as if they were colorectal cancer patients, and some were 50% more likely to show willingness to participate in a hypothetical RCT than real patients with the disease.43 Our study also has provided a summary of reported non--patient-related reasons for failing to enroll eligible patients in surgical RCTs in order of the frequency and significance with which they have been reported to affect recruitment in real or hypothetical surgical RCTs. These reasons can be summarized as being related to informed consent (difficult, time consuming, and involving uncertainty), the complexity of study protocols, clinician’s loss of motivation attributable to lack of recognition or financial or other rewards, or a change in the

Real RCTs 28 28 15, 28 28 26 26 26

Hypothetical RCTs

38, 40, 48 42 42 38, 40

48 40 48 48 43

25, 26, 29 28 38 44

attitude toward research in general or in the commitment to the RCT in particular. Other reasons included lack of resources and administrative support, practice commitments and lack of time, concern about negative outcomes, a compromise to the doctor-patient relationship, and loss of surgeon’s autonomy and medicolegal and ethical concerns, especially if the surgeon develops a preference for a certain intervention. In the early 1980s, Taylor22 reported that, in more than three quarters of the cases, concern about the doctor-patient relationship was the reason for not entering eligible patients into an RCT designed to assess the role of lumpectomy in the treatment of breast cancer. In more recent articles, this reason has been emphasized more in literature reviews than in reports of real RCTs.5,48,49,51,58 This tendency may be because of the change of patients’ attitudes to and increased awareness of the concepts of RCTs. The difficulties associated with obtaining an informed consent and adhering to the study protocol, compounded by lack of time and supportive staff, appear to have overtaken the concern about the doctor-patient relationship in significantly contributing to the disappointing recruitment rates in many surgical RCTs. There were conflicting reports as to whether the nature of the disease (benign or malignant) has a significant impact on the patient’s decision to participate in a surgical RCT. The finding by Australian researchers38 that breast cancer patients were significantly less likely to show willingness to participate in RCTs than those attending outpatients

480 Abraham, Young, and Solomon

Surgery April 2006

Table X. Recommendations made to improve accrual rates in surgical RCTs Recommendation Design and conduct issues Multidisciplinary collaborations and communication between study headquarters and study centers Adequate continued support for clinicians (eg, data managers) and providing incentives to participate. Evidence-based assessment of organizational characteristics Comparing collected data uniformly and prospectively Recruitment issues Appropriate method of invitation, targeted recruitment policies (eg for ethnic groups), and evidence-based assessment of recruitment strategies Prerandomization screening Zelen’s prerandomization Improved information provided to trial patients Aiming for a high recruitment rate early in the trial Measures to relieve the patient’s anxiety and gain their trust Same gender researcher-patient interaction A patient reimbursement moratorium Well-defined referral routines and inclusion criteria Clinician-related issues The clinician’s enthusiasm, orientation profile (experimenter/ therapist), and level of training Regular workshops, distribution of newsletters, and trial updates at scientific meetings. Involvement of primary health care providers Measures reported to have had no influence on recruitment Providing institutional incentives Surgeon or nurse recruiter An easy-to-read consent form University or district hospitals Method of randomization

clinics for screening or investigation of breast conditions was at conflict with the finding by Japanese researchers44,45 that patients with malignant head and neck conditions were as likely to show willingness to participate in RCTs as those with benign disease. This conflict may be related to the intense emotional concerns probably associated with a diagnosis of breast cancer or to the different cultures and attitudes to disease in general and to cancer in particular between citizens of Australia and Japan. The results of the current study suggest that the extent of the disease and the severity of its symptoms negatively impact on the success of recruitment (2 studies support this argument,26,29 and 1 study reports no correlation28). The characteristics of patients most likely to decline entry into RCTs require further investigation. When supported by reports of nonsurgical RCTs, however, the results of the current study suggest several observations when compared with patients enrolled in RCTs: Eligible patients who

RCTs

Surveys & reviews

6, 32, 48

21, 48, 55, 58, 67, 69

23

43, 55, 58, 67 21, 58 67, 70

15

21, 67

31 1

67 71, 72 16, 17, 35, 63, 67

6 18

23

16, 17, 63 42 73 67 50, 53, 55, 58 67 67

12 14 13 45 42

are most unlikely to be recruited in an RCT are the single, older, male smokers from ethnic minority groups who need the assistance of an interpreter; those in a nonwhite collar occupation; those who are less-educated and less knowledgeable about RCTs and the benefits of intervention or the risks of inaction; those who reside away from study centers, have no previous participation in a trial, and suffer from ill health or malignant versus benign disease; and those who have more extensive disease or more severe symptoms.74-76 A prospective study of the reasons for the low recruitment rates into surgical RCTs using the example of the Australasian Laparoscopic Colon Cancer Study currently is in press (unpublished data, 2006). In this study, data were collected prospectively for all eligible-but-not-enrolled patients seen and treated by the surgeons actively involved in the trial for a period of 6 months. This study should help meet the current need to prospectively assess the reasons for the low recruitment rates into

Surgery Volume 139, Number 4

surgical RCTs. However, the findings of the current systematic review would suggest that ideal recruitment in a well-designed and justified surgical RCT would be best achieved if it involved the following approach: recruitment of altruistic, well-informed patients who do not have a preference for one form of treatment or the other, are capable of understanding the process involved, and believe that the study was drawn up for the study population; and developing a trust with an enthusiastic and motivated clinician supported by multidisciplinary collaborations whose practice commitments can allow him or her to spend an adequate period of time with the patient. Various recommendations made by researchers to improve recruitment also need to be prospectively assessed in an evidence-based fashion. These recommendations have been categorized in this review into trial design issues, recruitment strategies, clinician-related issues, and organizational issues. The fact that the most important part of a study design is deciding on the study question needs to be emphasized, as trials which are of little interest to the clinician or have complex protocols are unlikely to have good recruitment rates.47 As the patient’s attendance by the clinician is the point of entry for any clinical trial, adequate preparation at this level is of utmost importance. An important aspect of this preparation is the availability of support staff such as data managers, nurse practitioners, clerical staff, junior medical officers, and so forth. The recruitment process needs to be tested and practiced before patients are invited to participate in the trial. Various incentives for clinicians, patients, and institutions may need to be considered and tested against the success of recruitment. Zelen’s randomization technique might need further assessment before it could be recommended further, as several ethical concerns have been raised about this approach.71,72 Since this technique’s innovation,77 some studies have shown that it had a positive impact on recruitment,1 while others did not.42 REFERENCES 1. Chang RW, Falconer J, Stulberg SD, et al. Prerandomization: An alternative to classic randomization. The effects on recruitment in a controlled trial of arthroscopy for osteoarthrosis of the knee. J Bone Joint Surg (American) 1990; 72:1451-5. 2. Allen PJ, Stojadinovic A, Shriver CD, et al. Contributions from surgeons to clinical trials and research on the management of soft tissue sarcoma. Ann Surg Oncol 1998;5: 437-41. 3. Britton A, McKee M, Black N, et al. Threats to applicability of randomised trials: exclusions and selective participation. J Health Serv Res Policy 1999;4:112-21.

Abraham, Young, and Solomon 481

4. Jack WJ, Chetty U, Rodger A. Recruitment to a prospective breast conservation trial: why are so few patients randomised? BMJ 1990;301:83-5. 5. Benson AB 3rd, Pregler JP, Bean JA, et al. Oncologists’ reluctance to accrue patients onto clinical trials: An Illinois cancer centre study. J Clin Oncol 1991;9:2067-75. 6. Haidich AB, Ioannidis JP. Effect of early patient enrollment on the time to completion and publication of randomized controlled trials. Am J Epidemiol 2001;154:873-80. 7. Haidich AB, Ioannidis JP. Patterns of patient enrollment in randomized controlled trials. J Clin Epidemiol 2001;54: 877-83. 8. Abraham NS, Young JM, Solomon MJ. Meta-analysis of short-term outcomes after laparoscopic resection for colorectal cancer. Br J Surg 2004;91:1111-24. 9. Allen PJ, Stojadinovic A, Shriver CD, et al. Contributions from surgeons to clinical trials and research on the management of soft tissue sarcoma. Ann Surg Oncol 1998;5: 437-41. 10. Stevens KR. Systematic reviews: The heart of evidence-based practice. AACN Clin Issues Adv Pract Acute Crit Care 2001 Nov;529-38. 11. Fung EK, Lore LM. Randomized controlled trials for evaluating surgical questions. Arch Otolarygol Head Neck Surg 2002;128:631-4. 12. Bowen D, Thornquist M, Goodman G, et al. Effects of incentive items on participation in a randomized chemoprevention trial. J Health Psychol 2000;5:109-15. 13. Coyne CA, Xu R, Raich P, et al. Randomized, controlled trial of an easy-to-read informed consent statement for clinical trial participation: a study of the Eastern Cooperative Oncology Group. J Clin Oncol 2003;21:836-42. 14. Donovan JL, Peters TJ, Noble S, et al. Who can best recruit to randomized trials? Randomized trial comparing surgeons and nurses recruiting patients to a trial of treatments for localized prostate cancer (the ProtecT study). J Clin Epidemiol 2003;56:605-9. 15. MacEntee MI, Wyatt C, Kiyak HA, et al. Response to direct and indirect recruitment for a randomised dental clinical trial in a multicultural population of elders. Community Dent Oral Epidemiol 2002;30:377-81. 16. Featherstone K, Donovan JL. ‘‘Why don’t they just tell me straight, why allocate it?’’ The struggle to make sense of participating in a randomised controlled trial. Soc Sci Med 2002;55:709-19. 17. Featherstone K, Donovan JL. Random allocation or allocation at random? Patients’ perspectives of participation in a randomised controlled trial. Br Med J 1998;317:1177-80. 18. Jenkins V, Fallowfield L. Reasons for accepting or declining to participate in randomized clinical trials for cancer therapy. Br J Cancer 2000;82:1783-8. 19. Mills N, Donovan JL, Smith M, et al. Perceptions of equipoise are crucial to trial participation: a qualitative study of men in the ProtecT study. Control Clin Trials 2003;24:272-82. 20. Spiro SG, Gower NH, Evans MT, et al. Recruitment of patients with lung cancer into a randomised clinical trial: experience at two centres. On behalf of the Big Lung Trial Steering Committee. Thorax 2000;55:463-5. 21. Foy R, Parry J, Duggan A, et al. How evidence based are recruitment strategies to randomized controlled trials in primary care? Experience from seven studies. Fam Pract 2003;20:83-92. 22. Taylor KM. Physicians’ reasons for not entering eligible patients in a randomized clinical trial of surgery for breast cancer. N Engl J Med 1984;310:1363-7.

482 Abraham, Young, and Solomon

23. Taylor KM. Physician participation in a randomized clinical trial for ocular melanoma. Ann Ophthalmol 1992;24: 337-44. 24. Williams CJ, Zwitter M. Informed consent in European multi-centre randomised controlled trials. Are patients really informed? Eur J Cancer 1994;30A:907-10. 25. Anonymous. Coronary artery surgery study (CASS): A randomized trial of coronary artery bypass surgery. Comparability of entry characteristics and survival in randomized patients and nonrandomized patients meeting randomization criteria. J Am Coll Cardiol 1984;3:114-28. 26. Hawkins BS, Vine AK, Willis J, et al. The Collaborative Ocular Melanoma Study (COMS) randomized trial of preenucleation radiation of large choroidal melanoma I: Characteristics of patients enrolled and not enrolled COMS report no. 9. Am J Ophthalmol 1998;125:767-78. 27. Hawkins BA, Vine AK, Willis JM, et al. The COMS randomized trial of iodine 125 brachytherapy for choroidal melanoma, II: Characteristics of patients enrolled and not enrolled: COMS report no. 17. Arch Ophthalmol 2001;119:951-65. 28. Sears SR, Stanton AL, Kwan L, et al. Recruitment and retention challenges in breast cancer survivorship research: results from a multisite, randomized intervention trial in women with early stage breast cancer. Cancer Epidemiol Biomarkers Prev 2003;12:1087-90. 29. Yuasa H, Kurita K, Westesson PL. External validity of a randomised clinical trial of temporomandibular disorders: analysis of the patients who refused to participate in research. Br J Oral Maxillofac Surg 2003;41:129-31. 30. Coombs DW, Dimick A, Bronstein JM, et al. Conceptual and methodologic problems in the evaluation of a new burn treatment modality. J Burn Care Rehabil 1993;14: 568-71. 31. Eisinger F, Giordanella JP, Brigand A, et al. Cancer prone persons. A randomized screening trial based on colonoscopy: background, design and recruitment. Familial Cancer 2001;1:175-9. 32. Goodwin PJ, Leszcz M, Quirt G, et al. Lessons learned from enrollment in the BEST study—a multicenter, randomized trial of group psychosocial support in metastatic breast cancer. J Clin Epidemiol 2000;53:47-55. 33. Plaisier PW, Berger MY, van der Hul RL, et al. Unexpected difficulties in randomizing patients in a surgical trial: a prospective study comparing extracorporeal shock wave lithotripsy with open cholecystectomy. World J Surg 1994;18: 769-72; discussion 773. 34. van der Windt DA, Koes BW, van Aarst M, et al. Practical aspects of conducting a pragmatic randomised trial in primary care: patient recruitment and outcome assessment. Br J Gen Pract 2000;50:371-4. 35. Caldwell PH, Butow PN, Craig JC. Parents’ attitudes to children’s participation in randomized controlled trials. J Ped 2003;142:554-9. 36. Ellis PM, Butow PN, Tattersall MH. Informing breast cancer patients about clinical trials: a randomized clinical trial of an educational booklet. Ann Oncol 2002;13: 1414-23. 37. Ellis PM, Dowsett SM, Butow PN, et al. Attitudes to randomized clinical trials amongst out-patients attending a medical oncology clinic. Health Expect 1999;2:33-43. 38. Ellis PM, Butow PN, Tattersall MH, et al. Randomized clinical trials in oncology: understanding and attitudes predict willingness to participate. J Clin Oncol 2001;19:3554-61. 39. Ellis PM. Focus group interviews examining attitudes to randomised trials among breast cancer patients and the

Surgery April 2006

40.

41.

42.

43.

44.

45.

46.

47.

48. 49.

50.

51.

52.

53.

54. 55. 56.

57.

58.

59.

general community. Aust N Z J Public Health 22:528-31, 1998 Aug. Fetting JH, Siminoff LA, Piantadosi S, et al. Effect of patients’ expectations of benefit with standard breast cancer adjuvant chemotherapy on participation in a randomized clinical trial: a clinical vignette study. J Clin Oncol 1990;8:1476-82. Maslin A. A survey of the opinions on Ôinformed consentÕ of women currently involved in clinical trials within a breast unit. Eur J Cancer Care 1994;3:153-62. Myles PS, Fletcher HE, Cairo S, et al. Randomized trial of informed consent and recruitment for clinical trials in the immediate preoperative period. Anesthesiology 1999;91: 969-78. Solomon MJ, Pager CK, Young JM, et al. Patient entry into randomized controlled trials of colorectal cancer treatment: factors influencing participation. Surgery (Australian) 2003; 133:608-13. Yuasa H, Yoda T, Kawai T, et al. Consent for enrolment in randomised controlled trials: A questionnaire study. Asian J Oral Maxillofac Surg 2002;14:125-31. Yuasa H, Yoda T, Kawai T, et al. Consent to enrollment in randomized clinical trials. Gan to Kagaku Ryoho Jpn J Cancer Chemother 1999;26:1459-64. Deber RB, Thompson GG. Variations in breast cancer treatment decisions and their impact in mounting trials. Control Clin Trials 1990;11:353-73. Ellis PM, Butow PN, Simes RJ, et al. Barriers to participation in randomized clinical trials for early breast cancer among Australian cancer specialists. Aust N Z J Surg 1999;69:486-91. Foley JF, Moertel CG. Improving accrual into cancer clinical trials. J Cancer Educ 1991;6:165-73. Langley C, Gray S, Selley S, et al. Clinicians’ attitudes to recruitment to randomised trials in cancer care: a qualitative study. J Health Serv Res Policy 2000;5:164-9. Martin RC 2nd, Polk HC Jr, Jaques DP. Does additional surgical training increase participation in randomized controlled trials? Am J Surg 2003;185:239-43. Penn ZJ, Steer PJ. Reasons for declining participation in a prospective randomized trial to determine the optimum mode of delivery of the preterm breech. Control Clin Trials 1990;11:226-31. Taylor KM, Kelner M. Interpreting physician participation in randomized clinical trials: the physician orientation profile. J Health Soc Behav 1987;28:389-400. Taylor KM, Shapiro M, Soskolne CL, et al. Physician response to informed consent regulations for randomized clinical trials. Cancer 1987;60:1415-22. Morrison W. Trial and tribulations: Patients prespectives of the BETASERON study. AXON 1994;16:51-5. McLeod RS. Issues in surgical randomized controlled trials. World J Surg 1999;23:1210-4. Solomon MJ, McLeod RS. Surgery and the randomised controlled trial: past, present and future. Med J Aust 1998;169: 380TA-3TA. Ross S, Grant A, Counsell C, et al. Barriers to participation in randomised controlled trials: a systematic review. J Clin Epidemiol 1999;52:1143-56. Kaluzny AD, Lacey LM, Warnecke R, et al. Accrual of patients to randomized clinical trials. Factors affecting cancer prevention and control research. Int J Technol Assess Health Care 1994;10:506-16. Angell M. Patients’ preferences in randomized clinical trials. N Engl J Med 1984;310:1385-8.

Surgery Volume 139, Number 4

60. Silverman WA, Altman DG. Patients’ preferences and randomised trials. Lancet 1996;347:171-4. 61. Levine RJ. Referral of patients with cancer for participation in randomized clinical trials: ethical considerations. CA Cancer J Clin 1986;36:95-9. 62. Gattellari M, Young JM, Solomon MJ. Randomized, controlled trials in surgery: Perceived barriers and attitudes of Australian colorectal surgeons. Dis Colon Rectum 2001;44: 1413-20. 63. Ellis PM. Attitudes towards and participation in randomised clinical trials in oncology: a review of the literature. Ann Oncol 2000;11:939-45. 64. Taylor KM, Kelner M. Informed consent: the physicians’ perspective. Soc Sci Med 1987;24:135-43. 65. Apateerapong W, Ananworanich J. Participation in randomized trials in developing countries. J Ped 2004;144: 140. 66. Prentice ED, Antonson DL, Jameton A, et al. Can children be enrolled in a placebo-controlled randomized clinical trial of synthetic growth hormone? Irb: A review of human subjects research 1989;11:6-10. 67. Jordhoy MS, Kaasa S, Fayers P, et al. Challenges in palliative care research; recruitment, attrition and compliance: experience from a randomized controlled trial. Palliat Med 1999; 13:299-310. 68. Lester H, Wilson S. Practical problems in recruiting patients with schizophrenia into randomised controlled trials. BMJ 1999;318:1075.

Abraham, Young, and Solomon 483

69. Sawka CA, Pritchard KI. Can improved communication increase patient participation in randomised clinical trials? Eur J Cancer 2001;37:297-9. 70. Probstfield JL, Wittes JT, Hunninghake DB. Recruitment in NHLBI population-based studies and randomized clinical trials: data analysis and survey results. Control Clin Trials 1987;8(4 Suppl):141S-9S. 71. Braunholtz DA. A note on Zelen randomization: attitudes of parents participating in a neonatal clinical trial. Control Clin Trials 1999;20:569-72. 72. Snowdon C, Elbourne D, Garcia J. Zelen randomization: Attitudes of parents participating in a neonatal clinical trial. Control Clin Trials 1999;20:149-71. 73. Jonas S. A proposed method for using a reimbursement moratorium to encourage recruitment for a randomized study of carotid endarterectomy. Stroke 1986;17:1335-6. 74. Hollis JF, Satterfield S, Smith F, et al. Recruitment for phase II of the Trials of Hypertension Prevention: Effective strategies and predictors of randomization. Ann Epidemiol 1995;5:140-8. 75. Israni AK, Halpern SD, McFadden C, et al. Willingness of dialysis patients to participate in a randomized controlled trial of daily dialysis. Kidney Int 2004;65:990-8. 76. Welsh JL, Adam P, Fontaine P, et al. Recruiting for a randomized controlled trial from an ethnically diverse population: lessons from the Maternal Infection and Preterm Labor Study. J Fam Pract 2002;51:760. 77. Zelen M. A new design for randomised clinical trials. N Engl J Med 1979;300:1242-5.