Barriers to Australian physicians’ and paediatricians’ involvement in randomised controlled trials

The two-part, self-administered questionnaire was based on the “Physician Orientation Profile” (POP).8,12,15,16 We combined the items from the four previous published studies. Where wording of questions differed between studies, we chose the version that was most clear. We also excluded questions that were specific for oncology practice. Ten physicians and paediatricians piloted the original POP questionnaire for acceptability and content. After feedback, three questions were reworded for the Australian context, and two questions were added.

The first part of the questionnaire consisted of 11 items that served as predictors of doctors’ treatment philosophies. These were demographic (age, sex) and practice details (practice specialty, professional appointment, clinical setting, location of practice, and number of patients seen per week) and research experience (previous experience with RCT participation, published articles and designated research time). The second part consisted of 33 items, which assessed the five indices of doctors’ attitudes to treatment and RCTs (see Box 2). The questionnaire is available from the authors upon request.

Binary-option questions were used where possible in the questionnaire to ensure that the final score had high discriminative value. When participants wrote their own answer instead of choosing a given option, the response was matched to the closest option given. A small number of answers were excluded because they did not match a given option. The POP scores for doctors’ attitudes to treatment and RCTs were calculated by a similar method to that used in previous studies, where each item was assigned a score from 0 to 1, with pure “researcher or scientist” responses scoring 1 and pure “clinician or therapist” responses scoring 0.8,12,15 The POP score was the mean of individual scores, and was derived by dividing the sum of individual scores by the number of questions answered and expressed as a number between 0 and 1 to ensure that the number of questions answered did not influence the mean score. In previous studies, doctors with higher POP scores were more research-oriented, reported higher past patient accrual to trials and showed greater intention to enter patients into trials in the future.12

A sample size of 500 was chosen (with 250 doctors in each group), which would allow the detection of statistically significant differences over a range of plausible proportions of responses to each question.

Demographic and practice variables and response to individual questions were compared between physicians and paediatricians using χ2 for differences in proportions. POP scores were compared across demographic and practice variables using analysis of variance (ANOVA) for differences in means. Backwards stepwise regression of the variables found to be significant (P < 0.05) on univariate analysis was performed to identify independent predictors of POP score. The SPSS statistical package19 was used for data analysis.

This project was approved by the ethics committee of the Children’s Hospital at Westmead and was supported by the Royal Australasian College of Physicians.

Three hundred doctors returned completed questionnaires (60% response rate). Significantly more paediatricians responded (55% v 45%; P = 0.004). Most respondents were male (72.8%) and practised in an urban setting (87.2%). The median age was 47 years (range, 32–94 years). Doctors worked in private practice (28.9%), teaching hospitals (53.7%) or both (6.4%), with 11.1% working in other settings. Appointments were in private practice (33.1%), as salaried hospital staff specialists (44.8%), academics (11.4%) or other (10.7%, all non-clinical). Those contacted by email were significantly more likely to participate in the questionnaire (65% v 39%; P = 0.0001).

Many respondents (26.9%) were not currently involved in any research (Box 1), and 78.9% spent less than 30% of their time on research activities; 56% had had an article published in the past 12 months. Nearly a third of respondents (31.5%) reported no experience with enrolling a patient in an RCT, and 44.9% had not enrolled a patient within the previous year.

There were few differences between the paediatricians and physicians. There were more hospital staff specialists (51.5% v 36.2%; P < 0.01) and females (34.1% v 17.9%; P < 0.004) among the paediatricians. Although research time and publication experience were similar in the two groups, fewer paediatricians had experience with enrolling patients in trials (62.2% v 76.3%; P < 0.009).

Primary allegiance (Box 2): The primary allegiance of most doctors was caring for individual patients. Relationships with patients were valued over research accomplishments and most felt that patients’ right to select treatments took precedence over advancing knowledge. A small group of respondents (5%; question [Q] 25) thought that the overall quality of patient care would decrease for trial participants and 27% (Q 27) thought that RCTs restricted their ability to individualise patient care.

Decision-making under conditions of uncertainty (Box 2): When published data and clinical judgement conflicted, 54% of respondents claimed they would rely on published data (Q 26). When there was controversy in the literature regarding the best treatment, most respondents preferred to personally select a treatment, with only 35% being willing to refer a patient for trial participation (Q 13). However, 54% were willing to refer a patient if they were personally uncertain about treatments (Q 40). Most would seek major input from the patient when making critical or controversial decisions (Q 17). More respondents were reluctant to participate in a trial involving a protocol that included a treatment that was more aggressive than their usual treatment (Q 33) compared with one involving a less aggressive treatment (Q 34); 28% were reluctant to participate in a trial involving a placebo (Q 35).

Professional activities (Box 3): Research plays a small role in most respondents’ professional activities (Q 19, Q 32, Q 42). Although most were willing to enter some patients into RCTs, only 25% would enter most, and 12% would enter all, of their potentially eligible patients (Q 18).

Perceived rewards (Box 3): Respondents perceived more rewards for clinical skills than contribution to scientific knowledge (Q 15) and thought that trial participation made no difference to their patient population (Q 44), although some (45%; Q 39) thought trial involvement might enhance their reputation. The laborious recruitment process (Q16, Q37), detailed monitoring of their management of trial patients (Q 29), increased paperwork (Q 30) and lack of financial rewards (Q 38, Q 41) were significant deterrents for trial participation for some participants.

Peer-group influence (Box 3): Most respondents (63%; Q 36) reported not being affected by the opinions of referring doctors about trial participation, and most perceived little pressure from their institution to participate in trials (Q 28).

These comparisons are shown in Box 4. The mean POP score for respondents was 0.44 (95% CI, 0.43–0.46). There was no significant difference in POP scores between physicians and paediatricians, or between males and females. Univariate analysis showed that doctors who were younger, in salaried positions or in urban settings and saw fewer patients had higher POP scores (ie, were more research-oriented) than doctors who were older, or in private or rural practice. Research involvement (ie, experience with enrolling patients in trials, time assigned to research activities and number of articles published in the past year) also correlated with higher POP scores. There were linear relationships between number of patients who participated in an RCT in the past year, amount of research time, number of articles published in the past year, and number of patients seen per week (after deleting doctors who did not see any patients) and the POP score — lower number of patients seen per week, higher patient participation in trials and increased publications correlated with higher POP scores. On multivariate analysis, three independent variables were found to significantly predict POP scores. These were allocated research time (0.050; 95% CI, 0.036–0.064; P < 0.0001), history of enrolling a patient in an RCT in the past year (0.022; 95% CI, 0.007–0.037; P = 0.005), and articles published in the past 12 months (0.012; 95% CI, 0.003–0.022; P = 0.012). The overall correlation coefficient (R2) for the model was 0.351 (ie, 35.1% of the variability in respondent’s POP score can be explained by these three variables).

We were unable to obtain practice and demographic data for non-respondents. Comparison with Australian workforce data obtained from 2001 suggests that our sample under-represented doctors in private practice. More respondents in our study reported involvement in research compared with the Australian clinical workforce (46.5% v 26.9%), although the questions by which these data were elicited were somewhat different.17 Despite efforts to minimise non-response bias, it is likely that our results overestimate the research orientation of Australian specialist-physicians, who are probably even more clinician-oriented than our findings suggest.

As the RACP database did not provide information on status of clinical practice, our sample included doctors who were not in active clinical practice. Response from practising clinicians may differ significantly from those not in clinical practice, as was shown in our study.

The crisis in clinical research is recognised worldwide and has been highlighted by the Clinical Research Roundtable in the US26 and by the National Health and Medical Research Council in Australia.27 Taskforces have been formed to address this issue.

We identified lack of research experience and opportunities for research as factors affecting doctors’ attitudes, and this may explain why research has a low priority for clinicians. Australian doctors are clinician-oriented rather than research-oriented in their attitudes to RCT participation, highlighted by their personal preference for selecting treatment rather than referring for trial participation in the face of treatment uncertainty. Research plays only a small role in their professional activities, and the importance of research participation is undervalued.

Future research is needed to further explore Australian specialist-physicians’ views, attitudes and barriers to participation in RCTs and to promote a positive attitude to trials and trial participation, particularly among doctors in private or rural practice. Strategies for providing the necessary time, resources and opportunities for research while reducing the workload for trial participation are needed. It is unclear whether the mandatory research projects of the RACP for Fellowship accreditation, which are designed to foster research interest, have been effective. The problems in clinical research are unlikely to be resolved unless the role of research is restructured in a primarily clinically oriented setting.