The potential and limitations of personalised medicine in primary care.

been developed for screening and for clinical prediction leaves open a temptation to confuse the two, and to misapply them in circumstances for which they were not designed. Tools designed to be used for screening are not necessarily effective in cases of individual decision making. This is not to say that there can never be circumstances in which screening tools are appropriate for individual assessment, but the onus must be upon those who wish to use tools in novel ways to justify such use in a systematic fashion. Rosendal and colleagues address an important issue. Improved diagnosis of somatisation will help clinicians to provide appropriate support and understanding for patients with this disorder. Better diagnosis will help to avoid unnecessary investigations, referrals, and invasive procedures with the attendant problems of cost, delay, and distress for all concerned. The challenge for studies of somatisation that seek to address these important issues is to find definitions of the disorder that can be explicitly related to decisions about therapy, investigation, and referral. Diagnostic categories that can be related to these pragmatic questions will make the greatest difference to patients and clinicians in primary care. TOM LOVE Research Fellow