How the Internet Can Help Clinicians Improve Their Clinical Skills

David L. Sackett, MD, FRSC, Director, Trout Research & Education Centre, Irish Lake, Ontario, Canada


Since no computer can match the accuracy of a seasoned clinician's history taking and clinical examination, why (on earth) mention the Internet and clinical skills in the same sentence? The reason is because 500 clinicians around the world are using the Internet to narrow down the huge number of signs and symptoms to just those that are really valuable in making accurate and efficient diagnoses (and, where possible, eliminating the need for expensive laboratory testing). They are inviting their Medscape-using colleagues to join this exciting and rewarding program of studies.

By the end of the initial history taking and physical examination, 88% of all diagnoses achieved in primary care[1] and 73% in general medicine clinics[2] are established. Despite the central importance of the history and physical examination to the clinical process, their accuracy (how well they predict the results of a biopsy or other definitive "gold standard") and precision (how well 2 clinicians would agree that a given sign was present or absent) have rarely been assessed rigorously, and never tested in the real world where 99% of medical care takes place. As a result, frontline clinicians have to decide for themselves which symptoms and signs are most relevant. Rather than work in isolation, an international group of clinicians has come together to carry out large (our target is over 1000 patients per study) and easy (no more than a couple of minutes added to the examination of just a handful of patients per clinician) studies of the clinical examination, using the Internet to nominate, discuss, design, and carry out individual studies and to disseminate their results immediately. They have formed the CARE (Clinical Assessment of the Reliability of the Examination) interest group, and invite their Medscape-using colleagues to join this free enterprise at A full description of the CARE program appeared in the Lancet[8] and the first CARE study was recently reported in JAMA.[4]

The CARE group is open to clinicians at any stage of training or experience and in any setting, and is structured so that any member can nominate symptoms or signs for assessment and circulate this to the rest of the group by email. Led by the CARE member who nominated the study, others who share that interest will design and debug the protocol, enroll patients in the study (members are expected to adhere to local clinical and ethical practice), and report their results via the Internet (with instantaneous data checking and editing) to the study-coordinating center. Patient names and other identifiers are never entered into the database, so that strict confidentiality is maintained. Results are analyzed, disseminated to the investigators, and synthesized into reports with multiple authors.

To take chronic obstructive airways disease (COAD) as an example, a scan of 21 textbooks of physical examination published between 1957 and 1998 yielded 40 different physical signs recommended for use in diagnosing COAD, but none of them described more than half these signs, and no textbook reported their precision or accuracy. Moreover, the only high-quality study of these signs and symptoms involved 2 university-based physicians, and it took them a year to examine 164 patients and to compare their examinations with "blinded" spirometry.[5]

To test the practicality of the CARE approach, its members nominated a study of the accuracy of 6 clinical items in the diagnosis of COAD, validated against independent blind spirometry performed during the same visit. A total of 26 pairs or groups of clinicians from 14 countries quickly joined the pilot study, and after deciding on the protocol they reported the clinical and spirometric findings from 309 patients in just 5 weeks (that's 20 times the rate achieved in the only other high-quality study of the disorder). After the study, a random sample of clinicians was asked to send their original data-collection sheets by fax. Checks with the database showed that the error rate in their entries was 1.6%. The CARE clinicians showed that just 4 findings (a positive history, a "low-riding" larynx, smoking more than 40 pack-years, and age over 44 years) were important in diagnosing COAD (defined as forced expiratory volume in 1 second [FEV1] and FEV1: forced vital capacity [FVC] ratios below the fifth percentile, adjusted for age, sex, and height). A total of 99% of patients with all 4 findings had positive spirometry, in contrast to only 12% of patients who had none of them. On the other hand, wheezes and laryngeal descent on inspiration were redundant, and simply took time to perform while adding no further diagnostic information.[4]

A second Internet-based study of the clinical examination for COAD has been completed. CARE members are now investigating the ability of the preoperative exam to predict postoperative respiratory complications. There is another CARE study in which nurses and physiotherapists are joining physicians as frontline investigators over the next 6 months to determine whether the elderly ambulatory patient who stops walking in order to start talking is at high risk for falls and fractures. All clinicians who are Medscape users and want to improve their clinical skills are invited to join this Internet-based real-world research group at