Hand assessment for infants : development, internal scale validity, reliability and normative reference values

Abstract

Infants use their hands to explore the world and neural event in the developing brain can affect the infant’s ability to use the hands. An infant that has suffered a brain lesion may show clinical signs of unilateral cerebral palsy as early as at three months of age, however, to quantify an asymmetric hand use and evaluate the results of early interventions there is a need for assessment tools producing valid and reliable measures. The aims of the Studies in this thesis were to construct an assessment tool measuring bimanual goal oriented hand use in infants 3-12 months of age, to investigate its psychometric properties for infants with clinical signs of unilateral cerebral palsy as well as to create normative reference values.

Study I describes the development of the Hand Assessment for Infants (HAI). The HAI is an assessment tool that constitutes of 17 observable items all measuring different aspects of goal oriented hand use. Twelve of the items are unimanual, scored for the right and the left hand separately, and five items are bimanual where one score is given. All items are scored on a three-point rating scale (0-2) and the scores of the unimanual items are summed to the Each Hand Sum score (EaHS) with a range of 0-24 raw scores. The EaHS of the better functioning hand and the lesser functioning hand is used to calculate an asymmetry index where a higher percentage indicates a larger asymmetry. The results of all the items are added to the Both Hands Sum score (0-58 raw scores) that has been converted to logits through the Rasch measurement model analysis to become the Both Hands Measure (BoHM), an interval scale of 0-100 HAI-units and the main outcome of the HAI.

The internal construct validity of the HAI was examined by the use of Rasch measurement model analysis and we found that the HAI has a unidimensional construct and good fit to the Rasch model requirements for both items and persons. A high separation index shows that the HAI can separate between infants of different levels of ability. In Study II normative reference values were created through compiling 489 HAI assessments of typically developing infants 3-10 months. Results show that typically developing infants improve their bimanual hand use, as measured by the HAI, at a rapid pace and that a clear majority of typically developing infants does not demonstrate an asymmetrical hand use. We can now compare the results of an infant with suspected signs of unilateral CP to the reference values to identify infants with atypical hand use. In Study III the reliability of the measures of the HAI was investigated both for test-retest and inter rater reliability. Both test-retest an interrater reliability of the outcomes of the HAI was excellent as shown by the range of the reliability coefficient (Intraclass Correlation Coefficient) 0.97-0.99. For all individual items reliability was good or excellent with an ICC range of 0.81-0.99. The smallest detectable difference of the HAI is 2.36 meaning that a difference in results in 3 HAI-units can be considered a true change.

In summary the studies in this thesis shows that the HAI provides valid and reliable measures of the object related hand use of infants with signs of unilateral Cerebral Palsy. The HAI can be used to identify a possible asymmetry and differentiate infants with an atypical hand use as well as to quantify asymmetric hand use over time in infants with clinical signs of unilateral cerebral palsy.