In this study, 61% of physicians declared that they requested an ID consultation more than once a month. They were physicians working in medical, surgical, and intensive care units, although those working in medical and surgical units declared requesting more consultations than the others. This is concordant with a previous study conducted in the same setting showing that 55% of the 2933 consultations requested of ID specialists over 1 year came from medicine or rehabilitation units, 41% from surgical units and 2% from intensive care units . We could have expected a majority of requests from surgeons as they are less experienced with IDs. However, surgeons were less numerous than physicians working in medical units. Furthermore, physicians working in medical units in French university hospitals are often highly specialized. Although they are used to managing common IDs themselves, most hospitalized patients have complex illnesses that ID specialists are specifically trained to assess [1, 13].
The majority of responders (93%) was satisfied with ID consultations and had a positive opinion on the pertinence of recommendations. This is much higher than a previous study that showed only 53% satisfaction for curbside consultations among primary care physicians and medical subspecialists . The consultation with an easily reached specialist seems to be an appropriate answer to an important need for the hospital community given the frequency of ID encountered among inpatients [13, 15]. Consultations were provided by a dedicated team used to informal and formal consultations for inpatients in the different departments. It can be assumed that this team had developed the basic principles of good communication with the requesting physicians, which has been shown to increase adherence with recommendations [16, 17].
The most frequent reasons for requesting ID specialist consultations were to obtain therapeutic advice and to improve quality of care, consistent with other studies [12, 13]. Another frequent reason, declared by nearly two-thirds of physicians, was the rapidity of access to the consulting ID specialist. Although the rapidity of access by itself may not be the only reason for requesting a consultation, it can obviously facilitate it. Indeed, if an ID specialist can be reached easily and gives an appropriate recommendation, it is likely that non-ID specialists will contact him rather than seeking other sources of information, which is more time-consuming. This result is highlighted by the predominance of reaching specialists via the hotline and much less often via other routes. The importance of communicating directly at the time a consultation is requested has been reported previously .
In the present study, self-reported adherence to therapeutic and diagnostic recommendations was 83% and 79%, respectively. These figures are close to those found in the Sellier et al. study conducted in the same hospital where adherence to ID recommendations was objectively assessed by an independent physician and was measured at 88% for therapeutic recommendations and 72% for diagnostic recommendations . In the free comments, several physicians found that recommendations were sometimes overly standardized because the specialist did not examine the patient. Though they found the consultations useful, they decided not to follow the recommendations precisely because they believed they knew the patient better. Informal consultations make up 46% of ID consultations dispensed in this setting . ID specialists seem to provide informal consultations for less complex cases in comparison with formal consultations . It is likely that the most important recommendations would not change with additional clinical information and that standard recommendation may be justified. Moreover, a previous study showed that adherence to ID specialist recommendations and patient outcomes were comparable between formal and informal consultations . It should be noted that informal consultations seem to require more experience on the part of physicians, as it has been shown that ID residents tended to provide informal consultations less than board-certified specialists.
Another frequent suggestion made by the requesting physicians was to reach only board-certified ID specialists and not residents. Coverage during working hours by both a resident and a board-certified specialist was a choice made in order to reduce the time to obtaining the answer. The substantial workload of this ID consultation requires an operational mode based on ease and rapidity of access. Furthermore, it has been demonstrated that ID trainees provide valuable recommendations during consultations . Three requesting physicians expressed a concern about the traceability of recommendations given by telephone. Shared access to the patient's computerized medical chart can be a way to decrease the risk of liability. Also, in an informal consultation, it could be suggested that the ID specialist write a note in the chart mentioning the reason why the patient did not need to be examined.
The majority of physicians declared using the consultation service to update knowledge. Although the educational impact of specialist consultation has often been described [1, 21, 22], it is worthwhile to consider another aspect underlined in the present study. The rapidity and the guaranteed relevance of the recommendations made by the ID specialist may incite the requesting physician to rely on the consultation and not seek the answer to his request on his own.
Our study has several limitations. Forty-five percent of the physicians having received the survey responded to it. It is likely that those who participated in the study were the most likely to be aware of this consultation service, but we cannot exclude that physicians who answered may have had a better opinion regarding ID consultations than those who did not. Thus, satisfaction might have been overestimated in this study. By comparing the demographic characteristics of responders and nonresponders, we found certain differences between them. They may consequently represent two different populations. Moreover, although few responders answered the open question, the comments made should help specialists of the team to consider how they could improve their practice. Finally, our survey was carried out in one university hospital in France, and we cannot exclude that the findings may be different in other settings and geographic locations.