A method of classifying children as chronically ill demonstrated high specificity.

In the U.S., a significant proportion of children are chronically ill. There are many different chronic illnesses; some are rare and each has a different trajectory–some are temporary, while others can be life-long. However, there is no standard definition of a chronic illness; and, there are numerous methods for classifying children as chronically ill.

This article reports on a study that used the trajectory of childrens' illnesses to identify them as chronically ill. The primary measure was potential life long chronic conditions (LLCC). The study described a method that used hospital discharge data to identify children with LLCC; the study employed Clinical Risk Groups (CRG), a method of identifying patients based on conditions and the severity of those conditions; the CRGs covers a range of chronic conditions in categories such as: acute, significant acute, minor chronic, and moderate and dominant chronic.

The author's selected patients whose primary care from 2006 to 2007 occurred at Odessa Brown Children's Clinic, a community clinic affiliated with Seattle Children's Hospital. The study examined hospital discharge data from both the hospital and the clinic.

Key Findings:

• In a blinded verification against data for 200 children, the CRG generated seven false-positive LLCC identifications.
• The CRG generated 14 false-negative identifications.

Children with LLCC utilize high levels of health care resources. The authors indicate this is the first study to employ hospital discharge data processed through software coding algorithms, to identify patients for collaborative care management. A limitation to the study was its sensitivity to mental illness.

[This research was funded in part by the National Association of Children's Hospitals and Related Institutions.]