A two-page survey was mailed to staff using an internal mail system in November 2006. The survey was previously piloted among several groups of different health professionals. A postal survey was decided on because feedback suggested it would be more acceptable to staff than an online version. Questions in the survey asked staff how often they consulted textbooks or papers, experts, coworkers, search engines, and library resources to find information. The term “experts” was undefined, and respondent were left to identify colleagues they considered experts. The questions focused on ten key Internet-based resources provided by the library: UpToDate, Clinical Evidence, EBM Reviews, Harrisons, MEDLINE via Ovid MEDLINE or PubMed, Embase, TRIP, CINAHL, Google, and Google Scholar. Staff were asked how often they used these, how valuable they found each one, and what techniques they used when searching MEDLINE. Additional questions asked staff about their method of assessing the validity of information, wish to obtain more teaching in searching and/or appraisal, and preferred method for further tuition.

Professional groups showed some variation in information resource use. Medical and dental staff were more likely to consult search engines at least weekly (82.8%, n=58) or use library resources (62.9%, n=46) than either nursing (38.8% and 16.5%, n=101) or allied health staff (51.4% and 27.3%, n=72). A similar proportion of nurses consulted coworkers at least once a week (85.4%, n=228) as did medical staff (81.3%, n=52), but the proportion of allied health professionals was smaller (54.6%, n=107). All professional groups frequently consulted experts (nurses: 80%, n=216; doctors: 73.2%, n=52; allied health: 80%, n=107).

All professional groups used Google more than any other electronic resource (Table 1). More than half of all professional groups used Google at least once a month, including 86% of medical and dental staff. Ovid/PubMed was used at least monthly by most (81.6%) medical and dental staff but fewer nursing (31%) and allied health (40.6%) staff. By contrast, nursing staff reported higher levels of use of the CINAHL database, particularly compared with medical and dental staff. Differences between professional groups were also evident in use of the UpToDate service, with a higher proportion of medical and dental staff using it at least monthly (34.8%), compared with nursing (15.8%) and allied health (8.9%). Overall, many respondents did not regularly access any information resource: 37% rarely or never consulted Google, and 58% rarely or never consulted Ovid MEDLINE or PubMed.

Those resources for which high levels of use were reported somewhat predictably also received high value ratings from staff. Google received the most “very valuable” ratings of any resource. It topped all other resources in responses from nursing and allied health staff and was ranked second behind Ovid/PubMed by medical and dental staff. Interestingly, some resources that were accessed less often were still given comparatively high value ratings, notably TRIP and EBM Reviews.

In relation to the searching methods clinical staff employed with their use of MEDLINE, almost all responders (98%) used keywords to search the database, but few (7%) used a Clinical Queries or Medical Subject Headings browser.

The most popular option (85.3% of responders) for how staff assessed the reliability and authority of medical information found on the Internet was whether the information had been published in a reliable journal. This option was followed by information published by a professional organization (64.8%) and staff's critical appraisal of the information (48.3%).

All three groups of hospital clinical staff show a clear preference for Google among electronic resources. The high use of Google recorded in this study is probably related to the expansion of the Internet, the growing familiarity of health professionals with electronic resources, and health professionals' realization that the Internet, and particularly Google, is able to yield results quickly in time-pressured clinical environments [5]. Google has been shown to be useful in diagnosing difficult cases [1]; however, concerns have been raised about the quality of information it returns [6]. Google rates information according to the number of hits and may therefore favor older references. While Google Scholar offers better coverage of peer-reviewed medical literature, the only health sciences database indexed is MEDLINE and complex searching is still limited [7]. The fact that staff most commonly judges information to be authoritative if it is published in a reliable journal suggests that, while staff frequently use Google, they might be cautious about the reliability of information they find this way. It appears that with the increasing amount of academic health information being placed on the Internet, staff may be using Google as a quick way to access journals, possibly even using Google as a portal to MEDLINE, instead of using traditional library resources.

The data on the search techniques of those staff who used the MEDLINE database to find information presented very similar results when compared with previous research. Cullen consistently noted that 94% of users searched with keywords [8], and both studies showed very low use of Clinical Queries or Medical Subject Headings browsers and other advanced functions offered by MEDLINE. The absence of these skills suggested that practitioners might not be receiving the best evidence, and they identified a priority for future training, particularly given the phenomenal growth in MEDLINE [9].