The use of email is poised to revolutionize the delivery of health care with improved efficiency, convenience, satisfaction, or access to care [1-4]. In the clinical setting, email has the potential to be a tool of efficiency for physicians and convenience for patients. It offers yet another means of communication for physician and patient, and has even been used by some as a substitute for clinic visits when appropriate [5,6]. Despite the opportunities offered by this communication technology, physicians’ adoption of email with patients remains low [6-10].

Among the current literature, relatively little attention has been given to how physician-patient interaction through email has changed over time. We do know that email usage with health care providers among patients in the general US population continues to increase, albeit slowly [11]. Although early research reported a reluctance by patients to use email as a communication medium with their physicians [10,12], more recent studies have shown patients to be mostly willing to embrace the idea [13-17]. In examining the barriers and facilitators to physician-patient email communication, recent studies have suggested patient age [18,19], patient race [18,19], patient health status [18], physicians’ satisfaction with their work [20], physician specialty [21], and physician workload [22] to be correlated with email usage. Despite the presence of a few scholarly explorations within this niche, more studies are necessary to determine specific aspects of physician adoption of email as a communication medium.

A 2005 study from our group [9] identified correlates of physicians’ adoption of email use with patients and evaluated physician compliance with best-practice recommendations established by the American Medical Association (AMA) and American Medical Informatics Association (AMIA) [23]. We found that certain physician characteristics were associated with increased likelihood of email use with patients. Moreover, we found that the best-practice guidelines designed to protect physicians from liability issues, as well as protecting the privacy of the patient, were being used very infrequently. In the current study, we made ready use of updated physician data that we collected in 2008 from Florida using similar survey techniques. We examined whether email use with patients has changed over time (ie, 2005-2008). Moreover, given that adherence to best practices was low in 2005, we were interested in determining whether physicians more readily embraced the AMA/AMIA guidelines in 2008. Lastly, we explored the current physician and practice characteristics associated with email use with patients; and determined whether these factors changed relative to 2005.

We used a pooled time-series design that took advantage of two large-scale surveys of physician use of health information technologies in Florida. The two surveys used similar sampling strategies but did not necessarily target the same physicians in both 2005 and 2008. Data and methods from the first survey, conducted in 2005 (N = 14,921), have been previously reported [24-27]. The second survey, conducted in 2008 (N = 7003), had many identical questions. In the current analysis, we focused upon the questions pertaining to email use that were identical in both surveys and analyzed the combined data. The 2008 survey is attached (see Multimedia Appendix 1).

A total of 2057/7003 responses were returned in the 2008 survey, representing a 29.4% participation rate. Demographic and practice characteristics of respondents from both 2005 and 2008 are shown in Table 1. Overall, the 2008 sample included a greater proportion of female physicians and a higher proportion of family physicians and general internists. The 2008 sample also included a smaller proportion of surgical and medical specialists. Lastly, respondents in 2008 indicated having greater access to the Internet via high-speed connections, and fewer respondents indicated having dial-up access only.

The benefits of email communication between physician and patient have been espoused by many researchers [1,5,6]. It has been reported that email between physician and patient can improve efficiency and workflow within a medical practice, and improve access to care and convenience to patients [1,5,6,28]. Despite the improvements this communication medium can provide, Florida physicians in 2005 were infrequently using emailing with patients [9]. In the current study, we made use of newly collected data from Florida to examine trends in email use by physicians over time.

The main finding of our analysis suggests that, while a higher percentage of physicians reported having tried emailing with patients in 2008 than in 2005, the proportion of physicians who are actively doing so on a regular basis did not change significantly during this time frame. Furthermore, physicians who had not yet tried emailing with patients had a waning future interest in doing so. On the contrary, relative to 2005, physician use of email from their practices with entities other than patients remained high. In fact, in 2008, physicians reported an increase in email use with individuals at hospitals and pharmaceutical companies, suggesting that physicians did see value in this communication medium with selected stakeholders.

A troubling trend involves the lack of adherence to professionally developed best practices designed to protect physicians who choose to email with patients. In 2005, we found that very few doctors abided by most of the recommended best practices developed by the AMA and AMIA. Our updated data from 2008 suggest that even fewer physicians who email with patients were adhering to these best practices. Specifically, even though a greater number of physicians reported abiding by at least one of the 13 guidelines, overall, fewer physicians reported adherence to all 13 guidelines with significant reductions in 4 guidelines. It is possible that physicians were not aware that these guidelines exist. Furthermore, it is possible that, despite their knowledge of these guidelines, physicians found it difficult to incorporate these best practices into their routine workflows. It is also possible that these guidelines may be perceived as outdated given that they were published in 1998 when email usage was much more infrequent between doctors and patients. Our belief is that the guidelines are still relevant and thus, given our findings, physicians are still exposing themselves to potentially unnecessary liability when they fail to heed the recommendations of the AMA and AMIA with respect to email use. Efforts should be made to draw attention to these guidelines, as well as simultaneously demonstrating how these guidelines can be adopted by physicians and integrated into their practice’s workflow.

In 2008, several physician or practice characteristics were associated with email use with patients. In multivariate analyses, female physicians, specialty physicians (as opposed to primary care physicians), and those in a multispecialty practice were all more likely than their counterparts to use email with their patients. Given the increasing time demands on primary care physicians in terms of providing recommended services and preventative care [29], and the increasing length and volume of primary care visits [30], these physicians may have less time available to devote to emailing patients. This is particularly important in light of the national trend toward improving patient-centered medical homes in which primary care physicians are empowered to increase services, including through electronic means, to their patients.

On the other hand, physicians in multispecialty practices can gain economies of scale that help with certain administrative processes [31], which may provide more free time to use email with patients. Lastly, it is not clear why female physicians were more likely than their male counterparts to use email with patients. However, previous researchers have found that female doctors were more likely to earn continuing medical education (CME) credits online [32] and be responsive to email invitations to CME programs [33], both activities that may increase their proclivity to use information technology within their practice. More research is needed to more fully understand this trend.

While, in 2005, physicians in the largest practices emailed with patients most frequently, practice size was no longer a significant predictor of email use in 2008. Furthermore, while age was negatively associated in univariate analysis with patient email use in both 2005 and 2008, multivariate analysis of 2008 data that included the newly available measure of self-reported computer sophistication made differences by age disappear. Our data show that computer sophistication may be a better predictor than age of technology adoption among physicians. Those who were very sophisticated computer users were significantly more likely to report emailing with patients than those who were sophisticated, neutral, unsophisticated, or very unsophisticated in increasingly higher proportions.

There are several limitations of this study worth mentioning. First, given that our data relied on self-reported information, we realize that our data are limited by participants’ ability to recall information accurately and their willingness to do so. Second, we recognize that response rates to both surveys were suboptimal. However, several researchers, including our group using the current data, have found that response bias in studies of health information technology are not likely, given the noncontroversial nature of questions on such surveys [34,35]. Third, given the pooled time-series nature of our analyses, we cannot be certain that the same physicians responded to our survey in 2005 and 2008. Even though we used very similar sampling methodologies, the characteristics of the 2005 and 2008 samples were different in some ways, including gender and specialty. Although these differences may be true changes in demographic characteristics among Florida physicians over time, we recognize that these differences may be a weakness of the study. Lastly, our study was conducted in a single US state where demographic, socioeconomic, and medicolegal characteristics affecting physicians may not generalize well to the rest of the country. Thus, we recommend caution when interpreting our results as reflective of physicians outside of Florida.

If physician-patient email communication is indeed valued as a patient-centered approach to improving health care quality, more effort will be needed to alleviate physician reluctance to engage in this activity. In Florida, the proportion of physicians who regularly email with patients rose only slightly between 2005 and 2008. This was a period when other health information technology applications such as electronic medical records and e-prescribing increased significantly in adoption nationally and in Florida [36,37]. Policy makers will need to seriously consider ways in which to encourage this activity if the potential benefits from physician-patient emailing are to be realized. Physicians who are using or are considering using email with patients are urged to become knowledgeable of best practices, which they can employ in their organizations.