MEDITEL

To the Editor. —  The Journal has, during the past 2 years, brought to the attention of its readers information about two available computer-assisted diagnostic systems: DXplain, described in 1987,¹ and our system, MEDITEL, reviewed this year.²The evaluation of such systems has posed new challenges for medical journals, as articulated by a pioneer in medical informatics, the late Dr Marsden Blois.³ The development of fair, objective, and appropriate methods requires consideration of the special characteristics of computer systems in contrast to books or scientific reports.Relative to the two systems, a recent publication compared the performance of DXplain and MEDITEL in 103 serial admissions to the inpatient internal medicine wards of a major university hospital not involved in the development of either system.⁴ Both ease of use (time for system access and data entry) and accuracy (inclusion of the correct diagnosis in the system output)