Identifying Children with Lifelong Chronic Conditions for Care Coordination by Using Hospital Discharge Data

Background en with lifelong chronic conditions (LLCC) are costly, of low prevalence, and a high proportion of patients at children’s hospitals. Few methods identify these patients. ives ght to identify children with LLCC in hospital discharge data for care coordination by using clinical risk groups (CRGs), to evaluate the accuracy of this methodology compared with a chart review and to investigate accuracy according to condition groups. s ftware identified LLCC children who receive care at a primary care clinic, Odessa Brown Children’s Clinic, by using Seattle Children’s Hospital discharge data. s were 5356 active Odessa Brown Children’s Clinic patients with at least 1 clinic encounter in 2006–2007. Six hundred two (11.2%) patients were admitted to Seattle Children’s Hospital, and 1703 (31.8%) were seen only in the emergency department over 7 years (2001–2007). One hundred sixty-four (7%) were identified to have a LLCC. In a blind review of 200 (33.2%) children with inpatient encounters, the specificity of the CRG designation to LLCC was 95.0% (95% confidence interval [CI], 90.0%–98.0%), sensitivity 76.3% (95% CI, 63.4%–86.4%). Mental health conditions formed the largest group that was chart-review positive and CRG negative (7 of 14). Children hospitalized before 13 months of age were the second largest group (3 of 14). Clinical review placed the 164 patients in these condition groups: sickle cell disease, 43 (26.2%), neurological, 37 (22.6%), mental health, 22 (13.4%), malignancies, 4 (2.4%), other 52 (31.7%), and no chronic condition 6 (3.7%). sion tudy demonstrates a unique way to identify children with LLCC for care coordination by using hospital administrative data.