Article Text

Predictive value of preschool surveillance in detecting learning difficulties
  1. SHARON GOLDFELD,
  2. THERESA LAZZARO,
  3. FRANK OBERKLAID
  1. Centre for Community Child Health and Ambulatory Paediatrics
  2. Royal Children’s Hospital
  3. Flemington Road, Parkville
  4. Victoria 3052, Australia

    Statistics from Altmetric.com

    Editor,—In trying to predict learning difficulties using preschool data, Corrigan et al have overlooked some basic principles of health surveillance and screening for developmental problems.1 Detection of learning difficulties is a reasonable goal of health surveillance. However health surveillance does not include using data from one point in time to predict problems seven to eight years later.2 In fact, the efficacy of preschool developmental screening as a global phenomenon is still not clear and certainly has not been shown to be useful in detecting mild to moderate learning difficulties.3 There is also generally a poor correlation between perinatal events and subsequent learning outcomes except in extreme cases.

    Predicting learning difficulties using data from the preschool or even neonatal period also seems a paradox given that learning difficulties are by definition related to educational problems that occur in schools. The authors have supported this theory by specifically choosing children who had been at school a minimum of two years, thereby allowing teachers time to detect learning problems.

    The outcome measures of developmental delay or learning difficulties in the preschool period that are used in this study are vague and both require further definition. The diagnosis of developmental delay using referral to a psychologist or documentation in the child health record would fit best with surveillance methodology rather than an outcome measure. Preschool learning difficulties, once again seem to be an incongruent concept given the previous definition.

    The results do not seem to answer any hypotheses of clinical relevance and this is evident by the fact that being a single mother in the neonatal period appeared to be protective against learning disorders, when in fact we know the opposite is true.

    In conclusion, the authors have reiterated the supposition that has been part of the literature for a number of years; learning difficulties cannot be assessed until the child is in an educational facility. Mild learning difficulties may be a reflection of maturational variability and the importance of detecting this in the preschool period is still being debated. There is no doubt that it is important to detect learning problems early in a child’s schooling and services should be in place to support these children.4There seems little point in mounting extensive surveillance programs in an attempt to predict difficulties years later. Outcomes of health surveillance should be rather directed to interventions which can be implemented in the present.

    Drs Corrigan and Stewart comment:

    That Dr Goldfield et al have missed the basic premise of this paper is evident in their opening comment. Our aim was not to predict learning difficulties using preschool data but rather to challenge the scientific basis of the existing system of child health surveillance which claimed early detection of mild to moderate learning difficulties as a stated goal.1-1

    We agree entirely with their concerns regarding the efficacy of preschool developmental screening and the difficulties in correlating perinatal events and subsequent learning outcomes. It was these concerns, and the challenges of Professor Hall’s reportHealth for all Children,1-5 that prompted our original research. The preschool outcome measures quoted, which Dr Goldfieldet al correctly refer to as vague, are those actually recorded by the preschool surveillance team. This is the reality of preschool developmental screening that we set out to challenge.

    We demonstrated that not only did the system fail to identify children as at risk of later learning difficulties in the preschool period but it was impossible to use the data recorded to develop a useful predictive model. This challenges a basic premise of a system that is enormously expensive in terms of both health care resources and parental time and energy. In a world of financial constraints and evidence-based practice we would suggest that the clinical relevance of such a finding is self evident.

    Finally, we would agree that there seems little point in mounting an extensive surveillance programme to predict later learning difficulties if this cannot be shown to be both sensitive and specific and offer proved interventions to aid the children identified as at risk by it. Far from disagreeing with our conclusions we would consider these points integral to them.

    References

    1. 1-1.
    2. 1-2.
    3. 1-3.
    4. 1-4.
    5. 1-5.
    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.