Identifying children with special health care needs: development and evaluation of a short screening instrument.

BACKGROUND Public agencies, health care plans, providers, and consumer organizations share the need to monitor the health care needs and quality of care for children with special health care needs (CSHCN). Doing so requires a definition of CSHCN and a precise methodology for operationalizing that definition. RESEARCH OBJECTIVES The purpose of this study was to develop an efficient and flexible consequence-based screening instrument that identifies CSHCN across populations with rates commensurate with other studies of CSHCN. METHODS The CSHCN Screener was developed using the federal Maternal and Child Health Bureau (MCHB) definition of CSHCN and building on the conceptual and empirical properties of the Questionnaire for Identifying Children with Chronic Conditions (QuICCC) and other consequence-based models for identifying CSHCN. The CSHCN Screener was administered to 3 samples: a national sample of households with children (n = 17985), children enrolled in Medicaid managed care health plans (n = 3894), and children receiving Supplemental Security Income (SSI) benefits in Washington State (n = 1550). The efficiency, impact of further item reduction, and flexibility of administration mode were evaluated. Rates and expected variation in rates across demographic groups of children positively identified by one or more of the 5 CSHCN Screener item sequences in each sample were examined and multinomial logistic regression analysis were conducted to evaluate the effect of child characteristics in predicting positive identification. RESULTS The CSHCN Screener took approximately 1 minute per child to administer by telephone and 2.1 minutes per household. During self-administration, over 98% of respondents completed each of the 5 CSHCN Screener item sequences, and respondents accurately followed each of the item skip patterns 94% of the time. Mailed surveys and telephone-administered surveys led to similar rates of positive identification in the same sample. Two Screener items would have identified 80%-90% of children positively identified as CSHCN across the study samples, although using only 2 items eliminates some children with more complex health needs. Rates of children identified by the CSHCN Screener varied according to age, sex, race/ethnicity, health status, and utilization of health services. CONCLUSIONS Results of this study indicate that the CSHCN Screener requires minimal time to administer, is acceptable for use as both an interview-based and self-administered survey, and that rates of children positively identified by the CSHCN Screener vary according to child demographic, health, and health care-need characteristics. The CSHCN Screener provides a comprehensive yet parsimonious and flexible method for identifying CSHCN, making it more feasible than existing measures for standardized use across public agencies, health care plans, and other users.