Margit Kooijman

50 This leads to the second aspect where general practitioners and physiotherapists appear to diverge from mainstream evidence which, opposite to the above, draws attention since it is as good as absent from their treatment, namely giving attention to the psychosocial aspect of the complaint. In addition, general practitioners do not seem to advise all their patients on home exercises and/or activity modification or refer to a physiotherapist for this reason. All in all, it seems that clinicians feel an urge to do something about their patients̕ suffering and less talk about it, especially when pain is prominent or with persistent complaints. This is interesting since current guidelines emphasize discussing the importance of staying active, the often lengthy and capricious course and the influence of psychological factors 11,14 . A very plausible explanation could be the patient’s expectation for the general practitioner or physiotherapist to solve their problem. And although it is only natural that this appeals to the care giver , without awareness on these sometimes unspoken demands, it easily leads to an increase of all kinds of interventions. While giving more advice will almost inevitably leads to a higher demand on the already pressurized time patient and clinician have together nowadays, in the long run it may save time and costly interventions. And last but not least, may lead to more effective recovery. The main concern in the interpretation of current study results is that stated rather than actual practice was measured. In their systematic review, Hrisos et al. (2009) found overall inconclusive evidence for the validity of clinicians self-report, but several vignette studies suggested that estimates of clinical behaviour were close to actual practice 18 . However, with close-ended lists of response options there is a risk of overestimation of performance since they provide an extensive number of possible actions. By using the same list of response options for every vignette we did an attempt to at least standardize this error. In addition, as is often wise with proxy measures, we focused on the big picture; on global differences between general practitioners and physiotherapists and between patterns of care amongst scenarios. Nevertheless, by its nature vignettes are brief and static and lack the possibility of detecting any non-verbal information, which can influence decision-making. Another disadvantage of this type of research is that it only captures one moment in time, which complicates assessment of appropriate referring. Referrals are an important aspect of treatment and to be able to compare clinician’s thoughts and ideas on this topic more precisely, we specified the reasons for referral from general practitioner to physiotherapist and vice versa. While this works well to measure if and why they refer to the most designated caregiver, it is far more difficult to value their capability of doing so at the right time. It could well be that clinicians chose to indicate that immediate referral was appropriate next to their own treatment whilst in actual practice they would await the effects of their interventions before potentially doing so. Despite this possible overestimation, our results show clear differences between scenarios. So, more detailed investigations into the effects of earlier versus later referral and the capability of primary care clinicians to do so timely is