Can we identify infants at risk of becoming obese, and if so, should we?

To the Editor,

As our understanding of the risk factors for obesity have advanced, we may now be in a position to identify children at a very early age who are at risk of becoming overweight and obese in later childhood. We write to highlight this possibility and to stimulate debate on the practical and ethical issues that might arise.

To our knowledge, there is no published research that has aimed to use routinely collected perinatal data to predict risk of childhood obesity. However, several studies have used risk analysis to develop theoretical prediction models. In a prospective pre‐birth cohort study of 1110 mother–child pairs from the Boston‐based VIVA project, the risk of overweight at 3 years of age was assessed according to four risk factors: maternal smoking during pregnancy, gestational weight gain, breastfeeding duration and infant sleep duration (Gillman et al. 2008). In this prediction model, the estimated probability of overweight ranged from 0.06 among children exposed to favourable levels of all four risk factors to 0.29 with adverse levels of all four. A review of gestational and early‐life risk factors (Toschke et al. 2005) conducted using cross‐sectional data from 4289 German schoolchildren 5 to 6 years of age determined the most predictive risk factors. Only 11% of the cohort was overweight, suggesting that this population had a lower prevalence than comparable with UK and US child populations. While high early weight gains were found in about half of the population, the positive predictive value of this factor was only 25%. When this was combined with parental obesity, the likelihood ratio increased and a positive predictive value of 40% was indicated.

Current collaboration between the Centers for Disease Control and Prevention in Atlanta, USA and the University of Leeds has suggested the feasibility of developing a simple statistical obesity risk tool (ORT) to be used by primary care practitioners at well‐baby check‐ups to identify babies at risk of later obesity and target them for intervention before they begin to show unhealthy weight gain. The ORT was developed using data from the Avon Longitudinal Study of Parents and Children (ALSPAC, available at: http://www.bristol.ac.uk/alspac/) and the Millennium Cohort Study (MCS, available at http://www.cls.ioe.ac.uk/studies.asp?section=000100020001). The factors considered included parental body mass index (BMI), maternal age, ethnicity, education, smoking; infant breastfeeding, sleeping patterns, birthweight and infant weight gain. The ORT that was developed from this work could be incorporated into a handheld device such as a personal digital assistant or mobile (cell) phone and could provide data for surveillance and help inform service planning. It would offer the possibility of help and support to parents whose babies have most to benefit.

For use in a primary care setting, the sensitivity and specificity of the tool must reach a generally agreed standard to give the user the confidence to identify the majority of children who are at risk and minimize the numbers mistakenly identified as at risk. Secondly, it must be simple. It must be based on ascertaining a small number of parameters that are readily available either from existing records or provided at interview. It must not be intrusive or require complex measurement or assessment and must be made available in a convenient and portable form, ideally enabling data downloading.

There are clear practical issues around the use of an ORT in a primary care setting. Which risk factors can be consistently and reliably ascertained at a well‐baby check‐up? Pre‐pregnancy or early pregnancy BMI, gestational weight gain, birthweight, infant weight gain and breastfeeding should be available, while paternal BMI may be unreliable. Given the availability of a suitable tool, care would be needed by the primary care practitioner to ensure that the risk assessment was carried out with sensitivity to the perceptions and concerns of the parent. Research indicates that some parents are often dissatisfied with the care they receive in primary care (Edmunds 2006) so there is a clear need for further training before a tool could be introduced. The outcome from an obesity risk assessment of an infant is difficult to predict. Some parents may be so stigmatized and antagonized that they will refuse to cooperate with the primary care practitioner and may be driven to disregard advice provided at a population level. Another response may be to seek advice and reassurance from their primary care physician. Yet, another and more unpredictable outcome could be that the parents become so alarmed by a high‐risk assessment that they embark on a totally inappropriate regime that drives the infant down into the lowest BMI percentile. Clearly, a tool could only be used in conjunction with interventional strategies that would provide appropriate guidance tailored to a baby's risk. However, the effect of attempting to mitigate risk remains to be evaluated. Apart from the direct effect on parents and children discussed above, the central ethical issue for the use of a risk assessment tool is the availability of a remedy for those who had been screened and deemed to be at significant risk. Without one, a charge of unethical professional activity could arise. Once again, this brings us back to the need for a strong evidence base to support currently available interventions and access to goods and facilities, such as affordable and conveniently available fresh fruit and vegetables and accessible recreational facilities, to support routine preventive strategies. If a practical tool is to be developed, we need to further explore and analyse the accumulating mass of data from longitudinal cohort studies that have been established in a number of countries so that the statistical base for the tool is validated and made country specific. The sensitivity and specificity of the tool must be within acceptable limits, which are yet to be defined. If this can be achieved, the next step would be to study the effects of the tool in a primary care setting by consultation with primary care practitioners, service providers and parent focus groups. A strategy could then be developed to place what is essentially a screening exercise within the structure of existing primary care programmes. Finally, a comprehensive system of outcome assessment must be developed in the form of longitudinal studies so that the success or failure of such an innovative risk‐based public health measure can be determined. Above all is the need to achieve lasting benefit while minimizing harm and this is the nub of the question –we may have the capability of identifying infants at risk, but should we?