National governmental agencies including the Institute of Medicine have called for the widespread adoption of electronic health records (EHRs) [1, 2]. Evidence suggests that use of EHRs can increase delivery of preventive care, enhance monitoring of drug therapy, and improve adherence to evidence based guidelines [3–6]. In response to this call to integrate information technology into patient care, US academic medical centers are increasingly incorporating electronic health records into teaching settings [3, 7, 8]. As more universities adopt electronic information systems, medical students are increasingly learning to conduct and document ambulatory visits using the EHR. Medical educators are now trying to assess the impact of EHRs on medical students and post-graduate trainees. Keenan and colleagues, in a recent review article, described that residents were satisfied with electronic medical records for a number of reasons: easy access of clinical data, legibility of notes, improved problem lists and medication lists, better preventive care documentation, and reduced medical errors [9]. The authors also found that the EHR-related education included point-of-care knowledge delivery, computerized decision support systems, profiling ACGME competencies, and daily work-flow management. As for the medical students, educators are also experimenting computerized order entry, templated care notes, and virtual patients, mostly during busy clinical clerkships [9].

Despite the increasing contribution of information technology to clinical care, the literature is sparse on the impact of electronic health records on the learning process for medical students in the ambulatory setting. Of note, electronic text books, hand-held computers and personal digital assistants (PDA) are becoming commonplace in undergraduate medical education [10, 11]. Medical students are using PDAs for convenient access of drug references and clinical calculators [12]. Most of them are satisfied with the use of hand-held computers and find them helpful for medical education and patient care, both in inpatient and outpatient setting [12, 13]. At least one RCT has shown that the use of PDAs by medical students resulted in improved learning in evidence-based medicine through the use of PDA-based decision support software [14]. Medical educators are now developing PDA-based mechanisms to track student learning in clinical clerkships [12, 15–17].

Studies with practicing physicians using the EHR have revealed concerns about adverse impacts on doctor-patient communication and patient care [18, 19]. Also, observational studies of videotaped clinic encounters have raised concerns that using the EHR may adversely impact history-taking skills and doctor-patient communication [20, 21]. Existing publications about medical students have largely advocated for the need for more exposure to information resources, including EHRs and computers [7, 22–25]. Despite the issue's importance, a 2007 Medline search revealed no studies exploring medical students' attitudes and concerns about using EHRs in ambulatory settings for clinical learning and patient care. In our study we report third year medical students' attitudes towards clinical learning using the electronic health record in two university-based ambulatory primary care clinics.

Our study provides interesting perspectives on how medical students in our institution viewed the impact of the EHR on their learning in an ambulatory clinical setting. While most students preferred the EHR as an organizational aid for asking more history questions and for better documentation of visit notes, it was clear that most were not utilizing key features of the EHR to augment their learning. Our students reported not using online resources (e.g., "Up-to-Date", patient education materials) and medication interactions more often.

A number of factors may explain these findings. Most students are of a generation that is familiar with computers and are able to type well. Therefore, it is not surprising that they prefer computer-based documentation, and overall, preferred EHR to a paper chart. However, despite having an adequate training, twelve weeks might have been insufficient for many students to learn how to use the EHR to its full potential, including linking to on-line resources and using preventive care templates and medication interactions. Arguably, the low reported use of such resources may simply reflect a developmental learning process. Perhaps, a number of factors, such as the clinical work-flows, design of and training with an EHR influence how users perceive the utility and benefits of an EHR. Of note, the relevant literature reveals that there is no correlation between the length of EHR experience and use of it's features. Arguably, student learning and perceptions could have been different if we used a different electronic record. We believe our EHR, like others out in the market, offered a similar set of features for outpatient clinical practice. Furthermore, we do not know if students' perceptions were influenced by the attendings' EHR use. We do not have any knowledge about the extent to which attendings were using the EHR and encouraging students to use those features during clinical encounters. We were interested to explore the extent to which medical students, who are typically expected to do detailed history and examination, view EHR-related tools as facilitating to learning and patient care.

Students reacted differently to different prompts from the medical record. Most reported that the prompts made them ask more history questions, but we do not know if this increased questioning resulted in a more effective clinical encounter. Although only around 40% reported ordering more preventive services when prompted by the EHR, that 40% translates into a great many preventive services that students might not otherwise have ordered. Similarly, only 30% reported learning about medication interactions from the EHR prompts, and it is not clear whether students already knew about these drug interactions or if they ignored the prompts. Also, since the nursing staff in our clinic updates a patient's medication list before he/she sees the provider, students may not have viewed as many medication interaction prompts as if they had updated the medication lists personally.

Although most students reported that their documentation was better and more complete with the EHR, only 24% reported that their oral presentations were better organized. The reasons for this are unclear. Our cohort were mid year third year students, so many may have considered that their baseline presentations were already well-organized. Also, these are self-reported data, and many students may not have been able to accurately judge the extent of any improvement on their presentations.

Regarding the impact of EHR use on patient follow up, over 50% of students reported accessing patients' test results electronically more often than from a patient's paper chart. This evidence of increased patient follow up is encouraging, but prompts questions about how frequently (or infrequently) do students routinely follow up on ambulatory labs they have ordered from paper charts? In a related point, only 15% of students used a very simple electronic prompt to remind themselves to follow up on patients. Either students were using other prompts (e.g. a notebook), simply remembering, or not following up at all.

In concordance with the existing literature, the students in our study raised concerns about how an EHR can impact patient, student and physician communication. While 64% reported overall satisfaction with doctor-patient communication with the EHR, many felt that the EHR might be a barrier for relationship building tasks (talking to, looking at, and building rapport with patients) during clinical encounters. Almost half (48%) reported spending less time looking at patients because of the EHR, and 34% reported spending less time talking to patients. Only 24% agreed or strongly agreed that using the EHR improved their rapport with the patient, and only 21% agreed or strongly agreed that their patients liked them using the EHR. More advanced learners (e.g., internal medicine residents) raised similar concerns in a study done at a VA (Veterans Affairs) primary care clinic. In that study, resident physicians and their patients were more concerned about the interpersonal aspects of care in the presence of an exam room computer, compared to faculty physicians and their patients [26]. It is possible that increasing experience with EHRs along with overall clinical experience may positively impact a provider's attitudes to the effect of the EHR on the doctor-patient relationship.

Students reported some impact of the EHR on the preceptor-learner relationship. Fortunately, only 9% of students considered that precepting with the EHR adversely impacted communication with their teachers. In contrast, almost 40% reported receiving more feedback on their electronic charts than on their paper charts. Considering the well-documented paucity of feedback in the ambulatory setting, we consider 40% of students reporting increased feedback to be an extremely positive finding.

Our study has several limitations. We had a small sample size – one group of third year medical students from a single institution, and our response rate of 62% is lower than the 70% generally accepted for generalizability. In addition, we distributed our questionnaire electronically, which may have biased our respondents towards those who are more comfortable with electronic technology. If that bias did exist, we might postulate that non-responding students would be less comfortable with the EHR, use its features less, and be more concerned about its impact on the doctor-patient relationship. We did not use an existing survey instrument, but we believe it had good content and face validity, as it was assembled by an expert panel from combined personal experience, recent literature findings and student-generated themes. However, we conducted no further formal validation on the instrument. Although, we added two open-ended questions in the survey instrument, we could have asked more specific questions to elicit finer details of responses not captured in the questionnaire (for example, "what features of an EHR would you want in your clinical practice?"). Our results are based on students' self-report, and as such, the data need to be interpreted carefully. We did not videotape any clinical encounters of our students nor did we query the EHR to obtain actual usage data, both of which could have yielded different and meaningful findings. We also wonder if some technological enhancements in our EHR, such as direct order-entry, pop-up reminders, easy-to use decision support tools, and use of tablet PCs could provide different results. Last but not the least, our lack of knowledge about attending physicians' behaviours and role modelling in an EHR-enabled setting is worth mentioning. Students' opinions in this regard may largely be dictated by whether or not they worked with technologically-savvy physicians. Despite these limitations, we believe our study should be viewed as a small but important pilot experiment. It adds to the paucity of literature on this extremely important subject, and raises several important questions to be addressed in future studies.

In our setting, third year medical students reported generally positive attitudes towards using the EHR in the ambulatory learning environment, although they consistently underused many technical features of the record (e.g. the ability to check medication interactions). They reported performing more complete histories and documentation. They reported receiving significantly more feedback from their preceptors on their electronic charts than on paper charts. However, students reported significant concerns about the potential impact of the EHR on their ability to conduct the doctor-patient encounter.

Since all current medical students are highly likely to use EHRs in their future practices, medical educators should attend to the 'red flags' raised by our study. As a community of educational scholars, we must conduct more research into how to best integrate the EHR into students' clinical learning. We must investigate why students consistently under-use important features such as medication interactions, health maintenance prompts and on-line resources such as clinical guidelines. We must also study the impact of the EHR on the teacher-learner relationship. There will likely be other benefits or drawbacks to precepting using the EHR apart from a reported increase in feedback. Finally, we must address learners' concerns about the impact of the EHR on the doctor-patient relationship. Perhaps Deans and clinical skills course directors should consider incorporating computers into basic communications skills classes for first and second year students.