As the need for high quality health services delivered at minimal cost intensifies, reliable and valid measures of disease severity are needed to insure appropriate case-mix adjustment in comparisons of patient and provider groups. Based on a widely tested adult severity measure, we developed and tested a new measure of disease severity in children ages 5-12 with cardiac disease (n=79), asthma (n=50), bone marrow transplant (n=48), or diabetes mellitus (n=73). We established disease-specific severity scales using symptoms, signs, and laboratory features. Treatments, utilization of services, and unusual or expensive procedures were excluded from severity measures. Data for each patient were obtained from hospital records and physicians' questionnaires. For example, for cardiac disease, we developed a 5-level clinical severity scale (CSS) combining dyspnea and fatigue (D&F), and similar scales using chest xray findings of congestive heart failure, arrhythmia effects, and cyanosis. Scales were tested against physician's global assessment of disease severity (range was 0-100, least to most severe), parent's report of missed school days, child's self-report of physical function (range was 0-100, worst to best), and percent of patients who used the ER and who were hospitalized for cardiac problems during the preceding year. The table displays each validity variable by CSS category for D&F:

Table 1
Full size table

Similar validity levels were observed for the other cardiac CSS dimensions and for the asthma, diabetes and bone marrow transplant scales. Although so far based on small numbers, we conclude that the CSS approach is valid for arraying children by clinical severity and can be used not only to assess intervention effects but also to adjust the performance of physicians and health care institutions for appropriate interpretations of information on cost and quality.