Our study showed that the coverage of influenza vaccination is high among residents of nursing homes for elderly people in France, whereas HCWs influenza vaccination coverage is insufficient. HCWs represent various occupational categories employed in nursing homes, including medical and non-medical personnel with various levels of contact with the residents. Our study shows that the staff involved in direct patient care are not necessarily better immunised than those with no such direct contact. Although no differences were observed between medical and non-medical staff, influenza vaccination uptake varied by category within the medical staff. Consistent with the results of other studies, our analysis indicates that influenza vaccination coverage was higher among physicians than among other HCWs occupations [15–19].
The high influenza vaccination uptake observed among residents indicates that recommendations for residents are well followed. In France, influenza vaccination is provided free of charge for all persons over 64 years in France. They receive a personal voucher for a free vaccine from the national health insurance fund. However, despite this high vaccination coverage among residents, influenza outbreaks still occur in nursing homes even when the adequacy between the circulating strain and the vaccine is good [20, 21]. The introduction of the virus, its dissemination through insufficiently vaccinated HCWs and intensive contacts with incompletely protected residents all contribute to transmitting influenza in nursing homes. Our study showed that despite the existing recommendations, influenza vaccine uptake among HCWs is insufficient in France, an observation shared by many countries. Influenza vaccination coverage of HCWs working in institutions varied usually from less than 10% [22, 10], to around 40% - 50% [16, 17], and rarely exceeded 50% [15, 23]. Even at relatively low level (43% - 51%), influenza vaccination of HCWs have an impact on morbidity and mortality among the residents in these nursing homes [9, 10].
Our study documents several pieces of information that will be useful to improve influenza vaccination uptake of HCWs working in nursing homes. Vaccination campaigns are relatively well followed by physicians, but seem to insufficiently overcome reluctance among nurses, nurse assistants and non-medical staff. One pragmatic approach would be to target and adapt the information given to the staff involved in direct patient care first: nurses and nurse assistants. As our study shows that information and education as key factors associated with influenza vaccination coverage, specific training and staff meetings targeting these categories (i.e. using adapted material) should be widely pushed forward.
According to our study, only half of nursing homes propose these types of measures in France. Educational campaigns could include staff in-service sessions, the use of posters or leaflets, mailings, and the organization of conferences. Furthermore, specific information should be given about the lack of proven efficacy of homeopathy. Homeopathic medications are widely used in France, and a study conducted in a French geriatric hospital reported that almost 60% of unvaccinated HCWs believed that "homeopathic medications are more effective than vaccination in preventing influenza" .
Our study also highlighted the role of easy access to free vaccination. Only 20% of HCWs were vaccinated when free vaccination was not offered. However, because 90% of the nursing homes already propose free vaccination, improvements which could result from this measure are likely to be low. Specific information should be given in LTCFs attended by residents with a high degree of dependency. In these settings, the vaccination policy for residents is particularly well followed, but a low vaccination coverage is observed among HCWs despite the fact that they are offered free vaccination in almost all LTCFs.
International studies have shown that a significant increase of vaccination coverage can be observed among all occupational groups when free vaccination is combined with a communication strategy [10, 16, 17, 24]. But, influenza vaccination coverage among HCWs hardly reached more than 50%, and improving this rate will be challenging . Comparatively, influenza vaccination coverage in the general population was 24% in 2005/2006 .
The role of the private status of nursing homes on influenza coverage of HCW needs to be explored further. Private nursing homes may have more encouraging policies for vaccinating their staff than public ones so that absenteeism is decreased during the influenza season. The role of the size of the nursing homes is unclear, but several reasons may explain these findings. One possible reason may be that nursing homes are more committed to their staff's health, especially when staff members are not numerous. Acquisition of influenza by HCWs may cause absenteeism, and the possibilities of a small team compensating for absenteeism are limited. Furthermore, one study showed that the fact of "believing that most colleagues had been vaccinated" was a main factor associated with complying with vaccination .
Our study has some limitations. First, we only collected aggregated data. Some individual information such as demographic characteristics, factors influencing the acceptance of the vaccine, knowledge on the influenza vaccine could not be collected. On the other hand, because the questionnaire was short and easy to complete, a good response rate was achieved. Secondly the questionnaire was self administered by senior staff in the institution and the accuracy could not be checked by the investigator, which could have lead to errors in responses. Thirdly, vaccination rates only included data reported by the nursing homes. That may possibly underestimate influenza coverage, especially among physicians who frequently do not work full-time in the nursing home, and could already be vaccinated outside the nursing home. Lastly, a recall bias can not be excluded, even if it was expected to be limited, because of the short time interval between the period of vaccination and the study. We believe, however, that the impact of these possible biases is likely to be limited. For instance, the estimates of influenza vaccination coverage obtained through this study are close to those observed in another study . In this study, among 64 LRTI outbreaks that occurred in nursing homes and that were reported to the French institute for Public Health Surveillance (InVS) during the 2006-2007 season, the average influenza vaccine uptake was 38% among the staff and 91% among the residents. In another study, the influenza vaccine uptake observed among patients in geriatric health care settings was 88% during the 2002-2003 season .