Article Text

Download PDFPDF

Clinical assessment of the reliability of the examination (CARE)
  1. Sharon E Straus, MD1,
  2. Finlay A Mcalister, MD2,
  3. David L Sackett, MD3
  1. 1University of Toronto Toronto, Ontario, Canada estraus{at}
  2. 2University of Alberta Edmonton, Alberta, Canada finlay.mcalister{at}
  3. 3Trout Education and Research Center at Irish Lake Markdale, Ontario, Canada sackett{at} on behalf of the CARE group

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Diagnostic tests are of crucial importance in clinical care and can help to determine the most appropriate treatment for individual patients, to monitor and modify ongoing treatment, and to determine prognosis. The critical details in deciding whether or how to use a diagnostic test are its precision and accuracy. Unfortunately, evidence to support these data is often difficult to find and frequently is not available at all. Further, although we are all taught that “a good history and physical examination” have considerable value, little evidence exists to support the diagnostic utility of the clinical findings.

    In a primary care setting, 88% of all diagnoses were established by the end of the initial history and physical examination,1 and similar results have been observed in a general medicine clinic.2 However, despite the importance of the history and physical examination to the clinical setting, their accuracy and precision have rarely been subjected to rigorous evaluation. Using the definition of a high quality diagnostic study applied by such evidence-based journals as Evidence- Based Medicine and ACP Journal Club (independent blinded comparison with the reference standard among an appropriate spectrum of patients), few such studies of the clinical examination can be identified. JAMA began the Rational Clinical Examination series with the objectives of summarising the literature on elements of the clinical examination and drawing attention to those areas in which there was an evidence gap. Only half of the reviews originally commissioned have been completed because the investigators were unable to locate high quality evidence relevant to their research question.3 More than half of the published reviews identified major gaps in the literature within their subject area,3, 4 and the studies that were found were often small and inconclusive.4

    Why aren't there better studies of the clinical examination? Several reasons have been offered.4, 5 Firstly, these studies are challenging to design and difficult to execute. Secondly, the specificities of symptoms and signs are often low in the usual study sites (tertiary care academic centres). Thirdly, diagnosis seldom arises from 1 symptom or sign, and significant methodological skills are needed to successfully carry out multivariate analyses that look simultaneously at several signs and symptoms. Fourthly, with the pressure to see more patients in less time in many situations, it is more efficient for the clinician (but not for the healthcare system) to order the “definitive investigation.”

    One way to progress is to conduct large, methodologically robust studies on the precision and accuracy of the clinical examination. These studies should be done among patients in primary, secondary, and tertiary care settings to document any changes in accuracy as patients move along the referral pathway. Furthermore, the studies must involve the broad spectrum of real world clinicians, with various levels of training and experience, who are the targets for what is learnt.

    Introducing the CARE group

    In response to this challenge, and building on the success of practice-based research networks in primary care,68 we have helped to form an international collaborative group to design and execute large, simple studies of the accuracy and precision of the clinical examination. We have borrowed from the terminology and expertise of clinical trialists in mounting simple studies by enrolling unselected consecutive patients and collecting a minimum of data.9 We hope that by keeping these studies quick and simple, participation will not be too onerous for either busy clinicians or patients.

    The CARE (Clinical Assessment of the Reliability of the Examination) group was initiated 2 years ago, and since then >600 clinicians from 35 countries have joined. The group is open to healthcare professionals at any stage of training or experience who are working in any practice setting. Any member can suggest symptoms or signs for possible evaluation and distribute them to the rest of the group by email. Members who share an interest in any of these topics come together electronically to develop the protocol, enroll patients in the study, and report their results through the internet to the study coordinating centre ( All participating investigators are required to adhere to local clinical and ethical practice. Use of the internet for data entry allows instantaneous data checking and editing. The electronic network also allows the study results to be analysed and disseminated to the investigators within hours of study completion.

    Testing CARE's feasibility

    To test the feasibility of this idea, we completed 2 studies of the accuracy of a few clinical items used in diagnosing chronic obstructive airways disease (COAD) validated against independent blind spirometry (10; unpublished data—to be reviewed in the November/December issue of Evidence-Based Medicine). In our first study, 332 patients were recruited in just 5 weeks, and 177 more patients were recruited in a second COAD study done 2 months later. Multivariate regression was used to develop a simple 4 variable model that will be tested in an independent set of patients in a third COAD study, which is scheduled for this autumn.

    Now that the feasibility of this network has been established, other studies of the clinical examination have been developed and are currently under way, such as assessment of the accuracy of a simple tool for predicting falls in elderly people. Bandolier highlighted an observation by Lundin-Olsson et al that patients living in a long term care facility who stopped walking when a conversation was initiated were more likely to fall within the next 6 months (likelihood ratio of 10) than those who were able to continue walking and talking.11 This screening test is simple, fast, inexpensive, and requires no special equipment. Furthermore, this simple manoeuvre, if validated in larger studies in various settings, could allow clinicians either to identify the people who would benefit most from interventions proved to decrease the risk for injury from falls or to find participants for randomised trials of interventions that might decrease that risk.

    Next steps

    Although several studies have been done to derive risk predictions for perioperative cardiac risk,1214 few have rigorously evaluated the prediction of pulmonary risk.15, 16 Indeed, we are unaware of any published prediction rules for assessing the risk for perioperative pulmonary complications. To meet this need, CARE investigators are currently enrolling patients in a study designed to determine the accuracy of selected elements of the history, the physical examination, and the routine preoperative tests for predicting postoperative pulmonary complications in patients having non-thoracic surgical procedures.

    It is our goal to expand CARE to a size and scope that would facilitate large (>100 clinicians enrolling >1000 patients), simple (<2 min per patient and <15 patients per participating clinician), and fast (<2 wk) studies of the clinical examination. Please join us in this fun and exciting project dedicated to improving our knowledge and performance of the clinical examination.


    View Abstract