An obstacle in comparisons of already existing qualitative interview data from different studies is that the context and content of the interviews vary extensively and there is usually inadequate information about the context to interpret how the findings compare. Still, comparative qualitative interview studies are uncommon.

In this comparative study we started out with a focus group study about GPs' attitudes to national clinical guidelines in Norway and subsequently repeated the study among GPs in Denmark. The Norwegian study confirmed international findings of central concerns about guidelines [8]; e.g. practitioners were concerned about the potential loss of clinical discretion and the holistic approach of general practice, which they feared increased use of clinical guidelines might lead to. However, even though only a fraction of Norwegian clinical guidelines are restrictive, the Norwegian GPs seemed especially sceptical to rationing motives in clinical guidelines issued by the health authorities, and claimed they frequently took the patient's side in opposition to the health authorities.

Norway and Denmark are neighbouring countries with many common characteristics regarding population and welfare and with similar health care systems. However, there are some relevant differences between the organisation of Norwegian and Danish general practice that we expected could be reflected in the GPs' attitudes. In general Norwegian GPs are subject to less government regulation and control than their Danish counterparts, although stricter regulations have been signalled [9]. In many ways Norwegian health authorities have followed the Danish: For example, a patient list capitation system and government issued clinical guidelines and regulations have been common in Danish general practices since the 1950s. A patient list system with partial capitation payment based on the Danish model was introduced in Norway in 2001.

An overview of the total volume of Norwegian guidelines relevant to general practice is lacking, but we know that government guidelines are fewer and of a more recent date in Norway. While some government guidelines are purely based on clinical assessments or giving advice about novel technology, government guidelines are increasingly used as a tool for rationing health care and thus incorporating economic considerations, e.g. since 2002 all new drugs are exposed to a cost-benefit analysis before being accepted for patient reimbursement [10]. Different government led institutions, such as the Directorate for Health and Social Affairs, the Norwegian Medicines Agency, the Knowledge Centre for the Health Services, the Norwegian Board of Health, and the Ministry of Health and Care Services, as well as some independent and private organisations; the Norwegian Medical Association, the Norwegian College of General Practitioners, and pharmaceutical companies, have all produced clinical guidelines for general practice in Norway in recent years. Comparatively fewer guidelines are issued by the government led institutions. The Directorate for Health and Social Affairs has only recently (after these interviews were conducted) set up a web-site - "The Health Library" - where guidelines and educational materials issued from both governmental and other sources since 2000 are accessible http://www.shdir.no/helsebiblioteket.

In Denmark clinical guidelines are primarily issued by The Danish College of General Practitioners (DSAM), The National Board of Health and the Institute for Rational Pharmacotherapy, a partly independent institute under the Danish Medicines Agency. DSAM and The National Board of Health have issued guidelines since the late 1980s and they issue both guidelines regarding innovative technology and restrictive guidelines to secure priority setting. DSAM started developing clinical guidelines to prevent health authorities from monopolising the agenda in guideline development [11]. Their strategy was to issue DSAM guidelines as well as to participate in developing government guidelines of relevance to general practice. Since 1999 DSAM has issued 15 guidelines and participate today in the development of guidelines issued by The National Board of Health, which has issued 8 guidelines since 1999.

The Institute for Rational Pharmacotherapy was established in 1999 "to promote the most rational use of current and future medicinal products with respect to both pharmacological and economical aspects" and is "directed towards both primary and hospital care" [12]. Since 1999 this institute has issued 8 reference programmes and 183 drug assessments.

The main views that emerged in both the Norwegian and the Danish interviews were:

• The GPs generally felt unable to keep updated on new treatments and research evidence. Many expressed that they regret this situation. They therefore underlined the need for more short and practical clinical guidelines.

• In addition the participants claimed that they were sometimes sceptical to the evidence clinical guidelines were based on. Still, there was consensus about trust in guidelines developed by the GP organisations.

• The interviews also show that GPs experience a dilemma between standardising practice (guidelines are part of this) and individual treatment of the consulting patient. Patients can also appear demanding, which sometimes lead to negotiations with patients to comply with clinical guidelines.

• The interviews indicate that format, accessibility and implementation strategy influence the use of clinical guidelines.

The GPs claimed that they mainly use guidelines that make work easier, not those who complicate practice.

They found some of the guidelines difficult to access.

Some said they miss one complete source for clinical information and guidelines.

An interesting national difference was evident from an early stage in the analysis. This was a difference of degree, not of substance, but it was striking nonetheless: The Danish group discussions did not reflect the Norwegian concern with economic considerations in guidelines. The Danish GPs demonstrated a much more positive attitude to government guidelines, and they appear more appreciative of government efforts to implement priority setting through guidelines. Although they conveyed some displeasure with obligatory regulations from the The National Board of Health, the Danish GPs appear to have internalised the health authorities' goals for resource allocation (as well as the clinical advantages of standardising practice), accepted the need for rationing and thus see their practice as part of a greater scheme of priority setting. While the Norwegian GPs present themselves as allied with their patients against the authorities' attempt to ration health care, the Danish GPs see a need for rationing. The Danish GPs also claimed they sometimes explain their view to patients when they ration health care to follow government guidelines. Thus the Danish GPs do not convey the same one-sided confidence in and alliance with the patients as we registered in the Norwegian study [8]. Additionally, the Danish GPs' seemed to have less respect for patients' demands than their Norwegian colleagues, even though the Danish GPs also admitted difficulties with rationing healthcare when patients are demanding.

The comparative analysis also suggests that the Norwegian doctors fear a coming development towards more standardisation of practice, less clinical freedom and a strengthening of economic considerations at the expense of clinical concerns in government guidelines. In contrast, the Danish GPs remember that they used to be sceptical to government issued clinical guidelines but acknowledge that their attitudes have changed. Hence they now underline the need to standardise care and secure fair allocation of resources.

In the following two sections we present some characteristic extracts from the Norwegian and Danish focus groups that illustrate the core differences in attitudes to rationing and government guidelines.

Nevertheless, we also noted a clear cross-national difference: While the Norwegian GPs in this study appear rather sceptical to economic evaluations incorporated in clinical guidelines, and do not see such clinical guidelines as part of a greater scheme for priority setting, the Danish GPs seem able to combine the need for rationing with their professional identity as holistic generalists. Some of the Danish participants explicitly state that economic evaluations need to be included for the benefit of the public. GPs' concerns about economic vs. clinical standards in guidelines have been noted in a few earlier qualitative studies [17,19].

On the one hand, the similarities are important findings to report because they support the knowledge base, and they are easier to explain than the diverging findings. It is more difficult to explain the dissimilarities, which could for example be results of atypical study samples. However, striking differences merit attention and are interesting because they may lead to new hypotheses about explanations that inspire further studies.

If we for example presume that GPs' attitudes reflect organisational features in the national health care systems, it is relevant to ask whether the difference in attitudes between the two groups of GPs in this study represents stages in an international development of general practice. Whereas Norwegian health care has recently started a process of incorporating economic considerations in government guidelines, Danish GPs are accustomed to applying priority setting. Hence, the Norwegian GPs could be voicing preoccupations about a new situations and how it will form clinical practice, while the Danish GPs have internalised the rationale behind using clinical guidelines for rationing. Thus while the Danish GPs do have to negotiate with patients to follow clinical guidelines and still sometimes end up with acceding to patients' requests, they are not themselves opposed to the motivation implicit in the guidelines.

It follows from this line of argument that as rationing through clinical guidelines is increasingly imposed on GPs, they gradually become more positive to such guidelines as they internalise the ideas behind rationing of health care, or perhaps adapt to a situation they cannot change. The development of primary care in UK would be an interesting case for testing this theory. The British primary care system has in many ways inspired reforms and development of the Danish and Norwegian primary care sectors: The British system is well advanced with a central authoritative guideline source (The National Institute for Health and Clinical Excellence), which produces a steady stream of clinical guidelines. In contrast, Norwegian doctors seem to relate to no more than a dozen clinical guidelines and seem confused about which sources to trust [8,20]. Also with regard to government regulation vs. professional autonomy, the countries differ; British GPs are more closely monitored, and have long experiences with rationing health services because of budget responsibility, especially after the new GP contract of 2004 which implies that GPs' payment partly depends on whether evidence based guidelines are followed. Denmark seems to be in a middle position in this alleged development, both with respect to the range of government guidelines and in terms of regulation of professional discretion. If the theory holds, British GPs should be even more positive to rationing than Danish GPs, but this remains to be tested.

A team of British researchers who have studied how British GPs' manage their clinical practices and professional identity explains the development of British general practice thus [21]: From the late 19^th Century the biomedical model based on hospital practice became prevalent and GPs gradually attempted to associate themselves with this model to maintain status. From the 1960s an anti-biomedical wave grew, partly spurred by patient empowerment, and led to the development of a holistic and patient centred approach in general practice. However, in later years, health authorities have increased the evidence based standardisation and regulation of general practice. Interestingly, the empirical results from this qualitative study indicate that even though British GPs currently are practicing according to the biomedical model, the GPs for the time being seem to be upholding their holistic identity i.e. the participants in this study claimed to be practicing holistically in spite of the apparent incongruity with their actual practice. Earlier UK studies suggest the opposite, e.g. in an interview study preceding the 2004 contract Charles-Jones et al found that British GPs were re-adopting a biomedical view on medicine, specialising to increase efficiency and that the holistic approach was being marginalised [22]. Summing up, it is fair to say that, according to current knowledge in the field, there is no unambiguous evidence that British GPs are welcoming rationing and standardisation through clinical guidelines.

Still, scepticism and concerns about future changes in work situation is a well-known psychological phenomenon [23] and fits well with current discourse in the medical literature, e.g. voiced in a comment in Current Surgery which recognises the controversies concerning guidelines for cost-control but still argues pro-guidelines as long as guidelines do not become the "demanded standards" of practice enforced by health authorities [24]. However, given that there is a relationship between increased economic considerations and GPs' attitudes to clinical guidelines, it need not be linear. Also, there may be alternative explanations for why different interview studies convey different attitudes among GPs. In depth studies are normally not suitable for generalising because the context in terms of interview questions, terminology, setting, relationship between interviewer and interviewees etc vary between studies. Hence, the differences noted across studies may be related to a range of other factors besides the organisation of the health care systems. Adding to the inevitable contextual differences between interview studies, a review of studies applying focus interviews in primary care research warns that the variation in how the methodology is conducted is tremendous [25].

The great advantage of focus groups is the rich and comprehensive material they yield, which in this case indicates that the differences and similarities noted here are well-founded. Adding to the validity of the study is the researchers' impression that the interviewees seemed relaxed and honest during the group interviews; humour and laughs were frequent, as were enthusiastic discussion. Also, we registered no comments indicating scepticism towards the researchers' or their role in this project.

On the other hand it is difficult to assess the representativeness of the participating GPs, i.e. the study sample. According to the limited data we have, the samples are similar to the average of the GP populations in both countries according to a few background characteristics, but bias according to unobservable factors such as attitudes and personality types remains unknown. Thus we cannot ignore the possibility that the difference in attitudes between the Norwegian and the Danish groups could be the result of peculiarities in the samples. However, the group by group analysis did not indicate marked differences between groups within each country, which indicates that unless there have been very different mechanisms influencing the group leaders' decision to participate in the two countries, the groups at least can be expected to represent the same type of GPs in both countries.