Reports from several constituencies have identified inadequacies in the teaching of MSK PE clinical skills [15-22]. A survey of medical students at Harvard University revealed that although the students considered MSK medicine as the third most important topic to their future medical career, only 26% of fourth year students passed a nationally validated written MSK exam [23]. To improve the consistency of MSK teaching, several national and multinational groups have developed MSK undergraduate medical school training standards [15,17,19,22,24-27].

Patient educators have been successful in teaching many aspects of general and system specific physical examination [28-34]. The Pelvic Exam Program is a patient educator initiative that has improved male and female genital examination clinical skills teaching [35-39].

The features of patient educator led teaching are well aligned with several educational learning theories. Patient educator programs were initially grounded in Bandura's social learning theory and bridge both Lave's Situated Learning Theory and Vygotsky's Social Development Theory [40-42]. Lave's Situated Learning Theory emphasizes that knowledge must be presented in authentic contexts in order to help learners move from the periphery to the centre of a "community of practice". As preclinical medical students are just entering the periphery of the medical community of practice, patient educators present a safe yet contextually authentic way to motivate them and impart knowledge. In terms of Vygotsky's Social Development learning theory, patient educators can act as the "more knowledgeable other" to help students gain more and more independence in their performance of accurate MSK physical examination.

The MSK specific patient educator program, Patient Partners^® in Arthritis (PP^®IA), is a centrally coordinated, standardized national program that trains volunteer patient educators with arthritis to teach and evaluate MSK clinical examination skills. Patient educators in this program undergo 100 hours of training. Though originally developed in 1989, this program was introduced in Canada in 1995 and is now active at thirteen of the seventeen Canadian medical schools at the undergraduate, postgraduate and/or continuing medical education levels. This program was introduced at the pre-clerkship level at McGill University in 2004 as an integral part of the teaching program and student evaluations of this teaching have been uniformly strongly positive.

Medical students report that first patient encounters are strong emotional experiences and first patient physical examinations are often described as anxiety provoking and confusing [43]. They also report feeling conflicted around the issue of using patients for their own learning [43]. Others have shown that the early introduction of clinical skills improves medical students' comfort when they start clerkship [44]. We hypothesize that PP^®IA patient educator led sessions could provide a transition experience and potentially lower the stress related to first patient encounters. These sessions may also alleviate the guilt related to the students' perception that they are using patients for their own learning as the patient educators are acting as true partners in the students' education. These sessions may also improve interest in MSK as an area of future study as previous studies have shown that, in general, the early introduction of clinical skills teaching and improved student self confidence with the subject area may also enhance learning interest in the subject [45,46].

There have been several descriptive studies of the arthritis patient educator programs' effectiveness [47-49] and six previously reported controlled trials of the impact of arthritis patient educator interventions [50-55]. Quantitative assessments of the efficacy of patient educator versus physician MSK PE teaching have yielded mixed results. However, four of six controlled studies used student self-report of clinical skills or single station objective structured clinical exams (OSCEs) to assess student performance [50,51,53,54]. These studies were characterized by unbalanced controls, small sample size or incomplete description of methods. Further, the varying teaching doses provided (anywhere from one to nine hours) and the differing outcomes measured, may have contributed to the lack of uniformity of the findings of even the more rigorous studies.

One study with very rigorous methodology compared a nine hour intervention of PP^®IA patient educator led versus rheumatologist-led small group clinical examination teaching [52]. This study found near equivalency of the two teaching methods. However, in the current Canadian climate, medical schools are rarely able to provide such extensive rheumatologist-led small group sessions at the undergraduate level [56]. The most recent study used a well designed two station MSK OSCE and showed equivalency in skills scores but slightly higher student instructor ratings for final year medical students in the UK [55]. This is the only study to date that included power calculations, allowing the authors to conclude equivalency between the 2 groups for the two OSCE stations based on scores within 10%.

This brief teaching session was also successful in improving students' retrospective pre and post ratings of comfort levels with many different aspects of MSK examination. It is reassuring that both groups' perceptions of comfort improved after the teaching and tOSCE sessions. It is also interesting that where significant differences did exist between the groups, PP^®IA patient educator led students consistently reported more improvement in their comfort levels than did physician tutor led students. With such a short teaching intervention, it is probably not surprising that significant differences in stated interest in further study in MSK were not seen.

One of the two rigorously conducted previous studies also showed PP^®IA patient educator teaching can result in similar OSCE scores for a nine hour teaching intervention compared to a similar duration session given by rheumatologists [52]. However, this study cannot comment on actual equivalence, as power calculations were not presented. Previously published surveys show that this amount of either patient educator led small group time or specialist-led small group time far exceeds what the vast majority of schools currently provide [3,56,64,65]. The current study's findings extend the results of Humphrey-Murto's study to a shorter teaching intervention of two hours total duration, which is more realistic in the preclinical setting. This study also extends these findings from MSK specialist physician tutors to general physician tutors. This is very relevant as survey data documents that non-MSK specialist clinical teachers teach MSK clinical skills at 60% of Canadian medical schools [56].

Although the other rigorously conducted study did present power calculations and did find equivalence between patient educators for a 2 hours intervention, it also used rheumatologists as its comparison, which again is not representative usual teachers of MSK clinical skills in most universities. In addition, this study did not use the existing standardized patient educator training program (PP^®IA), but rather employed its own unique training strategy to prepare patient educators which further limits the ability to generalize findings to other centres [55].

Unlike the Humphrey-Murto study, there were no tOSCE stations showing significant differences between the patient educators and physician tutor groups in the current study. One potential consideration that may account for this is the fact that the two stations where they found significant differences (ankle examination and sciatica examination) were not included in our tOSCE. The objectives for our teaching sessions were clear in their inclusion only of major peripheral joints. When considering the short time available, the authors did not feel that both the spine examination and the complete peripheral joint examination could be meaningfully covered in a two hour teaching session. However, the differences that they found were both in favour of the rheumatologist taught students. In our study, the hip station was the only one that came even close to significance and unlike the previous study, this was in favour of the patient educator taught students. It must be noted, however, that the current study focuses on peripheral joint exam and these finding may not be extrapolated to other aspects of MSK examination such as spine, soft tissue, power or gait examinations.

There were only five tOSCE stations used in our study compared to nine stations in Humphrey-Murto's study. When determining the numbers of stations to include in our tOSCE, it was important to consider teaching "dose" versus examination duration. Since only two hours of teaching were provided in this study, it was difficult to justify a nine station tOSCE. In attempts to resolve this dilemma, the author consulted two international experts in clinical skills evaluation (CV and LG). On considering their advice, it was decided that a five station OSCE would give the best balance of reliability through number of stations and sampling validity through an appropriate teaching to examination ratio.

There are several limitations that need to be considered when interpreting the results of this study. They include potential issues relating to sampling validity, sensitivity of the outcome measures and external validity.

Firstly, one must consider how representative the sample studied is compared to the general cohort of both instructors and students. As part of the criteria for the study, none of the physician tutors were specialists in MSK disciplines such as rheumatology, orthopedic surgery or physiatry. In all cases, the tutor agreed to give this MSK session. This reduced the possibility that participating tutors were more or less skilled at teaching MSK PE as all were included. Some of the participating tutors did express that they had not taught MSK PE when they taught this course in the past. This would indicate that we were not dealing with a group of general physician tutors that were unusually confident in their MSK teaching. The patient educators who participated in the teaching volunteered to give three two-hour teaching sessions over the 21 week IIM course. Although it could be argued that those who volunteered may have been more committed to the program, this type of selection would be parallel to what would occur if patient educators were asked to participate in actual undergraduate teaching sessions of similar duration. In terms of the students, student assignment to hospital site was quasi-random through the central Undergraduate Education Office and the few requests for change in placement occurred prior to students' knowledge of the study. Thus, it is unlikely that these requests would have caused systematic differences between these students and those at the other teaching hospitals.

Another potential question to the validity of the study relates to whether the tOSCE tested material that was representative of what students should learn in an undergraduate MSK PE session. The OSCE was structured around course objectives that were derived from the literature review on undergraduate MSK teaching objectives. One may wonder if a tOSCE is sensitive enough to pick up relevant differences between the intervention groups. However, a similar OSCE did pick up differences in the previous study by Humphrey-Murto and the reliability of 0.67 for the current study's OSCE was similar and high according to Cohen's guidelines [66]. One may also question whether the seven to ten day delay between the teaching and the tOSCE was appropriate. We chose this timing to minimize the influence of confounders on the results but we recognize that future studies with a second delayed OSCE to evaluate long-term retention would be desirable.

One may also wonder if the novelty of this intervention created a Hawthorne effect. However, as there was a change to the normal structure of both groups by way of the presence of a community arthritis patient or a patient educator, and as no differences were found between the groups this is unlikely to have had a major confounding effect. Had it done so, one would expect that the patient educator group would have outperformed the physician led group and this was not the case. One may also question the validity of student self-assessment. However, this study does not draw conclusions on the absolute self-confidence ratings but rather evaluates the comparison between two quasi-randomly selected groups of students. There is not reason to believe that there would be differences in the reliability of this data between the two groups.

Finally one may question whether the findings from a study performed at one University in one country can be generalized more widely. Fortunately, when comparing this university's practice to those reported in national Canadian, American and UK surveys, the amount of teaching offered appears more in line with standard MSK teaching practice in that a very brief amount of MSK clinical skills teaching is offered, it is offered primarily by non-MSK specialist physicians and is supplemented by patient educators [3,56,64,65].

There are many interesting points raised by this study. For example, the authors find it somewhat unusual that there were no differences in physical examination skills tOSCE scores despite reports from the qualitative data referred to earlier in this paper that some physician tutors did not manage to cover all of the specified joints in their teaching sessions [63]. These findings may be explained by the fact that students were aware of the upcoming OSCE and may have read independently to bring their knowledge in par with objectives regardless of teaching content or source. Although assessment is known to drive learning, previous studies with non-teaching comparison arms have demonstrated significant differences in MSK PE skills even when students were aware of an upcoming tOSCE [37,38].

In a qualitative analysis that occurred in parallel to this study, the authors found that physician tutors were more likely to emphasize the verbalization of physical exam manoeuvres as they were performed, than patient educators [63]. Differences in the students' abilities to express rather than perform skills could have led to a bias in the results in favour of the physician tutor group. Despite this, no significant differences were found.

The early introduction of patients into the undergraduate medical curriculum is desirable as it may improve the authenticity of the teaching experience by making the context of the learning environment similar to what students will experience in their future clinical roles. Patient educators give students experience on how to conduct themselves in front of real patients and give them more responsibility to develop rapport with patients who can give specific feedback in the absence of supervising physicians. Furthermore, patient educators provide students with explicit opportunities to learn from patients, appreciate patients as knowledgeable partners, and incorporate the patient's perspective as part of the educational messages.