Early screening for failure to thrive in infancy
The detection of failure to thrive in infancy is an important goal in routine surveillance of children. Failure to thrive is normally identified by slow weight gain, but is associated with feeding problems, and may lead to developmental delay and enduring intellectual defects. The prevention of these consequences is likely to depend on earlier detection of the condition than is currently achieved using traditional methods. An experimental computer based early screening method for the detection of failure to thrive was implemented in a two-year birth cohort (1,966 infants) in 18 general practices in the Easington area. The methods utilised an ACCESS database incorporating the British 1995 growth reference, which was used to convert the infant’s weight to a z score (conditional on age and sex). A 'thrive index' (a z score for weight gain conditional on age, sex and birth weight) was then calculated for the period from birth to the six to eight to week check and the infants in the slowest growing 5% automatically identified. The projected number of births in the 18 practices over the period 1 April 2001 to 31 March 2003 was 1800, and the actual number identified from health visitors' birth registers was 1966. For the 1966 infants, records of both a birth weight and a six to eight week weight were identified for 1880 infants. One hundred and twenty one infants met the criteria for FTT over this period (thrive index <-1.17) and of these, 102 term singletons were eligible to be recruited to the study. Those who participated had their development and weight gain followed to one year of age. Infants were tested at four months and again at nine months using the Bayley Scales of Infant Development (2nd ed). Mental development index (MDI) scores and psychomotor development index (PDI) scores of case infants and controls were compared and a mean difference was found between cases and controls in MDI scores at four months of 3.52 which was statistically significant. The mean difference in PDI scores at four months was 3.59, which was also statistically significant. At nine months the mean difference in MDI scores was 2.26 and the mean difference in PDI scores at nine months was 2.25, which was not statistically significant in either case. Information about demographic characteristics, health and feeding behaviour was obtained by using a structured questionnaire with the mothers. There were no statistically significant differences between families of case and control infants in indicators of affluence such as home or car ownership, nor were there any statistically significant differences between their mothers in their levels of educational achievement. There were no statistically significant differences between cases and controls in whether infants had ever been breast fed. Case group infants, however, were significantly more likely to be slow feeders than controls, and were more likely to take only small quantities and to be weak suckers. They were also more likely to be described by their mothers as having feeding problems. The screening method described provides a practical procedure for weight screening at the six to eight week check that allows identification of children who fail to thrive in the early weeks of life.