for 100 common diseases in large or small be population groups. These populations may inpatients or outpatients or may be ostensibly healthy people, selected on the basis of occupation, geographic distribution, etc.

Limitations of the MDS Method

As has been stated, the MDS method carries out routine tasks of matching patients' symptoms with those found in 100 common diseases and so is essentially a laboratory procedure. This screening method does not rate the clinical importance for the patient of each complex identified, nor does it rate the risk of not recognizing other unidentified disease complexes. Value judgments like these can be trusted only to a physician; they cannot be assigned safely to any computer system now in existence.

Final responsibility for all diagnostic decisions belong to the physician. For these and other reasons, clinical application of the MDS method is as an adjunct to the medical profession in its care of patients.

Because the MDS method analyzes symptoms as they are observed by the patient, it does not readily detect evidences of diseases which can be identified with confidence only through physical examination or special laboratory procedures. Further, the MDS cannot identify symptom complexes of a disease when the patient does not report the symptoms, is asymptomatic, or reports symptoms markedly different from those usually found in the disease. For these reasons, the MDS does not identify every disease that every patient has.

A disease may be falsely identified if a patient has a related disease which is producing symptoms resembling those usually found in the disease identified. No information is available to the MDS method as to which is the presenting symptom, and the MDS does not identify which of the patient's symptom complexes is associated with the present illness.

Summary

The MDS method is essentially a laboratory procedure that brings to the medical profession a kind and volume of information about the whole patient which is not readily available otherwise. Some of the more obvious advantages of the method include the wide applicability of the comprehensive information it develops; the relevance of this information to the medical and psychiatric status of individuals, teenaged through elderly; the high validity of the information; the ease with which it is obtained; and the speed and low cost of the method.

APPENDIX

Statistics for the MDS Method

Because the MDS method analyzes only complaints made by the patient, it permits only presumptive identifications of diseases and does not yield proven diagnoses. Basically, the method gives answers for 100 disease to the question: "Is there evidence that this disease should be considered diagnostically for this patient?" The answer given is a "Yes" or a "No", and a "Yes" answer simply recommends that the physician consider the disease when making his diagnostic evaluation of the patient.

Several publications describe the statistics of the MDS. The questionnaire used in the early studies was the Cornell Medical Index and in the later studies the Medical Data Index (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20).

As does a physician, the MDS method considers a symptom to be significant for a disease if it occurs comparatively frequently in patients with the disease and comparatively infrequently in all patients generally. Similarly, it measures the significance of a whole complex of symptoms for a particular disease after making a correction for age, and then matching the complex reported by the patient with the age-corrected complexes of symptoms characteristically found in patients of the same sex who were diagnosed by physicians as having the disease.

where Pak is the relative frequency of the symptom in the disease and P, is the relative frequency in all patients generally. Each significance value is then corrected for age and all values for a disease are summed, after which the sum is compared to the average sum found with patients who were diagnosed by physicians as having the disease. A sum which equals or surpasses the average sum constitutes an identification of the disease in the patient being tested.

In the computer-aided statistical MDS method, only those patients are identified as having symptoms of a disease whose complexes of symptoms are comparable with those of the average patient. Only typical cases are therefore identified; a physician, on the other hand, by recalling his own and other published experiences with unusual patients, and by using as yet unknown processes of thinking, can recognize atypical cases.

diseases according to the sex and age of the patient, and that the results are consistent with what clinical experience indicates are diseases to be considered in making differential diagnoses for patients claiming the symptoms named.

Description of

the MDI Health Questionnaire

The questionnaire is a four-sided, letter-sized sheet headed Medical Data Index-Health Questionnaire. The heading explains the purpose of the form to the patient. It is self-administe ing with 10 to 15 minutes generally being the time required for completion of the form. The printed directions instructing the patient to circle the "Yes" or the "No" after each question are easily followed.

The questionnaire, a descendant of the Cornell Medical Index (1, 2, 3, 4, 5, 6, 7, 8, 9), consists of medical questions stated in informal language and designed to be understood by people with a reading knowledge of simple English. The questions are, in many respects, similar to those asked by a physician when he is undertaking a detailed review of systems. Questions are identical for men and women except for eight questions referring to the genital system.

Technical terms rarely occur, but when they are necessary an explanation is given in parentheses; so, for example, abdomen is explained as "belly" and mucus as "slime". Questions relate to various categories in the following manner: