Against the backdrop of the likely future increase in numbers of carers because of an ageing population and more people living with improved healthcare so that more people live with disability for longer , it is important that general practice and GPs are in a position to support them. This study has demonstrated that although GPs regard general practice as an appropriate place to do this and think they should be pro-active in their role, they frequently lack confidence and training and sometimes knowledge to do it effectively. Key issues highlighted include the identification of carers - some GPs are aware that they are not identifying all carers in their practices and would like to be guided on how best to do this. Few services are currently being offered by general practices specifically for carers despite GPs' belief that carers want their support. The few carers' registers reported here by the GPs is all the more surprising given the financial incentives to have them.
There are several possible reasons for this. GPs may believe that as carers have no specific medical diagnosis and have social not medical problems, supporting them is perhaps only on the periphery of their role . They may also think identification and support of carers may add to an already demanding workload and have concerns that carers may present problems they are unable to help with. However, the GPs in this study agree that carers deserve their support even if insufficient time and resources and concerns about confidentiality remain. Clearly the fact that carers and the person being cared for may not have the same GP or even attend the same practice complicates both identifying carers and may make communication more difficult.
There are noticeable differences between our findings and those of Simon and Kendrick (2001). Variations in question wording (such as open-ended questions versus checklists in Simon and Kendrick's study) and our sample's attendance at a carers workshop account for some of this, but the differences are still striking. On a positive note, the vast majority of our sample (90%) compared with a quarter in the earlier study said that they should be proactive in supporting carers. Approximately a third of our participants (29%) said they had received some previous training in carers compared with 10% nearly a decade ago. On the other hand, five GPs here (6%) reported keeping a carers' register. In Simon and Kendrick's study, a quarter recorded carer status. Although these activities are not identical, this suggests that identifying carers may remain a low priority. Also nearly half the GPs in Simon and Kendrick's study offered information to carers, compared with a handful here.
There are similarities with our findings and the research from Australia . In both there is an apparent gap between what GPs think carers might like and what they offer them. Despite awareness of carers' emotional needs, GPs in Australia tended to refer carers to practical, rather than emotional support. Similarly here although a quarter of the GPs thought carers would like emotional support, the only service provided that might be directly expected to offer such support was carer support groups available at three practices. GPs here described fewer services for carers than in the Australian study but nearly all those mentioned here were also practical in their approach.
Nearly a decade ago it was suggested that primary care teams could support carers in a number of ways including acknowledging the problems they have and ensuring the general practice team are aware of them, flagging carers' notes so that GPs were aware of their caring role, acknowledging the role they play, treating them as team members and provision of information for the carer relating for example to the condition of the person the carer is looking after and information about benefits and services . The fact that the DH and the RCGPs organised these workshops suggests greater awareness of the importance of carers and the significant role they play but our findings suggest that many of these earlier recommendations still stand. Although most GPs here did agree that carers should be a partner in the care of the recipient of care, little progress seems to have been made with the other recommendations.
However, there is very little recent published literature on the topic and it may be worth speculating why. Perhaps it is a symptom of lack of awareness of the major role played by carers or perhaps an indication of the uncertainly and maybe ambivalence amongst GPs about supporting this group.
Setting up carers registers is an important first step but if GPs are unaware of common problems amongst carers and feel they have insufficient knowledge and time to offer support, raising expectations may be detrimental. A possible approach here is increasing the numbers of carers' leads or champions in primary care teams . Such carers champions are a member of the general practice staff who can recognising the needs and difficulties of carers and be able to offer them information and respond to their enquiries. The impact of this role has not been formally evaluated but they could assume much of the responsibility for supporting and signposting carers.
There remains work to be done. Repeating this survey with GPs who had not opted to attend training on carers would offer a wider perspective although the fact that the participants here had expressed an interest in the area but still lacked some knowledge and awareness of likely problems amongst carers (as suggested by the questions where they failed to answer or were unaware of some facts about carers) can also be seen to further the argument for the value of our findings. Better appreciation of GPs' perspectives and those of other primary care team members is required but research is also needed to identify what support carers need and want from primary care. Possibly adopting qualitative methods would allow a more in-depth understanding of all these perspectives. Research exploring the potential of roles such as carers' champions who might be well-placed to support carers whilst minimising the input from GPs would be beneficial.
Strengths and limitations
A strength of our study is that we achieved a high response rate for this group of participants  and the GP participants came from a range of areas geographically including both rural and urban populations and involved both large and small practices. The response rate was also high. Given the lack of research in the area, in many ways the study is exploratory since it investigated a wide range of issues many of which need further research.
A limitation is that the study participants had mostly chosen to attend a workshop on carers and may have been an atypical group motivated to support carers. It is impossible to know what impact this had on the findings but given the lack of research in the area, this is an important step in understanding GPs perceptions. Importantly these GPs were mostly attending the workshops because they wanted to know more about carers. There were a higher proportion of female GPs and they were slightly younger than might be expected  and nearly half said that they had an interest in carers and may therefore have been better informed compared with other GPs, although the current findings suggest they lack some knowledge in the area.