The challenge of defining chronic illness for health monitoring in children: opening a can of worms

Abstract

Abstract Background A straightforward health monitoring definition of chronic illness in children would have great utility for estimation of health care and social service needs as well as resource allocation. Methods This study analyses various definitions of chronic illness in children and adolescents in cross-sectional national health surveys in Germany based on: the Children with Special Health Care Needs (CSHCN) 14-item-questionnaire, a list of 15 self-reported medical diagnoses in the past 12 months, and a generic single-question parent-reported chronic condition from the Minimum European Health Module (MEHM), also used in EU-SILC and EHIS surveys. Results 13.7% (95% CI 13.0-14.4) of children and adolescents aged 0-17 years had special health care needs in the first National Interview and Examination Survey in Children and Adolescents in Germany KiGGS 2003-2006. However, surprisingly, the prevalence was less than half in children with a migration background and was not associated with social status or income. The list of diagnoses identified a very high proportion of 31.5 (30.7-32.4) of children aged 0-17 years as chronically ill but does not take into account disease severity or impact. The generic MEHM-question in the KiGGS 2009-2012 survey identified 16.2% (15.3-17.1) of 0- to 17-year-olds as chronically ill, however, only one in five was affected in their daily activities. Parents denied chronic illness in 11.2% of children with epilepsy, 18.8% of children with asthma and 62.8% of children with attention deficit hyperactivity disorder. Conclusions None of the three definitions of chronically ill children is entirely satisfactory so far: CSHCN may need an adaptation of the German version to better capture inequalities, the disease list needs further severity differentiation and MEHM has recently shown low test-retest reliability (kappa 0.56) when administered twice within 6 months in a large Belgian sample and may need revisiting as well.