Development of the Children With Disabilities Algorithm

A major impediment to understanding quality of care for children with disabilities (CWD) is the lack of a method for identifying this group in claims databases. We developed the CWD algorithm (CWDA), which uses International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes to identify CWD.

We conducted a cross-sectional study that (1) ensured each of the 14,567 codes within the 2012 ICD-9-CM codebook was independently classified by 3 to 9 pediatricians based on the code's likelihood of indicating CWD and (2) triangulated the resulting CWDA against parent and physician assessment of children's disability status by using survey and chart abstraction, respectively. Eight fellowship-trained general pediatricians and 42 subspecialists from across the United States participated in the code classification. Parents of 128 children from a large, free-standing children's hospital participated in the parent survey; charts of 336 children from the same hospital were included in the abstraction study.

CWDA contains 669 ICD-9-CM codes classified as having a ≥75% likelihood of indicating CWD. Examples include 318.2 Profound intellectual disabilities and 780.72 Functional quadriplegia. CWDA sensitivity was 0.75 (95% confidence interval 0.63-0.84) compared with parent report and 0.98 (0.95-0.99) compared with physician assessment; its specificity was 0.86 (0.72-0.95) and 0.50 (0.41-0.59), respectively.

ICD-9-CM codes can be classified by their likelihood of indicating CWD. CWDA triangulates well with parent report and physician assessment of child disability status. CWDA is a new tool that can be used to assess care quality for CWD.