Why invest in the pre-school years?

Phil Wilson: GP and senior lecturer in infant mental health at the University of Glasgow

Phil Wilson argues that pre-school influences shape children's social, physical and emotional development and that evidence shows that wise investment to improve outcomes for pre-school children can pay a rich social dividend.*

Children's experiences in early years lay down the foundation for the rest of their lives. Pre-school influences shape children's social, physical and emotional development long after they start school.

Costing pre-school benefits

Spending on pre-school years gives the highest rate of return on investment in human capital (Heckman and Masterov 2005). But, despite this, most is spent on universities and least on pre-school service provision (Alakeson 2005). Our pattern of educational expenditure thus seems to be designed to produce the minimum return. General public expenditure on the under-threes is minimal when compared with expenditure on any other age group, with specific expenditure being largely restricted to health visiting and some preferential benefits payable to unemployed mothers. This pattern of low expenditure on this period of life is perhaps most exaggerated in mental health provision. In most parts of the UK, mental health professionals rarely see children under three, and services tend to be provided by general practitioners, health visitors and paediatricians. Although some health visitors are now being trained in infant mental health, few doctors caring for young children have had such training.

Getting a good start

Many long-term studies, particularly birth cohorts, (Thompson et al. 2010; Jaffee et al. 2002; Jones et al. 2002; Murray et al. 2010) have identified preschool factors associated with poor mental health and violence later in life. These may be:

• Genetic (e.g. vulnerability to autism, Attention Deficit/Hyperactivity Disorder (ADHD) or anti-social personality disorder) (Caspi 2002)
• Antenatal (e.g. maternal stress hormones, smoking and alcohol consumption)
• Located in the family/upbringing (e.g. postnatal depression, harsh or inconsistent parenting, parental discord)
• Located in the wider environment (e.g. relative poverty, neighbourhood problems)

These factors may interact in different ways. Sometimes the risks may simply add up, sometimes they may amplify each other, and sometimes they may apparently reduce each other's individual effects. For example, some of the adverse effects of antenatal smoking are accounted for by the fact that mothers who smoke when pregnant are more likely to have postnatal depression (Maughan 2004).

There are factors which increase the resilience of children's brains to adversity. Positive parent-infant interaction protects against childhood psychological problems (Bradshaw and Tipping 2010). Higher intelligence, particularly in verbal abilities, is also protective (Emerson, Einfeld and Stancliffe 2010).

Child neglect

UK policy has tended to prioritise child abuse over child neglect, partly because of its greater visibility and media impact, and partly because of the relative ease of definition and simplicity of planning care pathways (Wilson and Mullin 2010). Neglect is, however, a much greater social problem and probably causes more long-term psychological difficulties. For example, Kotch et al. (2008), based on a robust study from the US, concluded that ...child neglect in the first two years of life may be a more important precursor of childhood aggression than later neglect or physical abuse at any age. While the impact of neglect is likely to become apparent when children enter school, there is a compelling case for earlier identification. Making this case has proved difficult (partly because of the difficulty of research in this area) (Wilson et al. 2009) but there is a considerable indirect evidence supporting the view that we should be assertive in seeking cases of neglect in order to support families and reverse its long-term effects. The fact that only a tiny fraction of children living with problem drug or alcohol use in the family are subject to even basic child protection procedures, shows that we are failing to protect the most vulnerable children.

The value of early interventions

Early support to vulnerable families by nurses is highly effective, and cost effective. For example, David Olds' landmark trials of the Nurse-Family Partnership† offered to vulnerable mothers in the US have demonstrated that about 30 hours of input between mid-pregnancy and the age of two, at an approximate cost of £3,500 (Olds et al. 1993), can halve criminal behaviour, substance use, smoking, running away and high-risk sexual behaviour by age 15 (Olds et al. 1998). Each of these behaviours has been shown in various studies to be associated with future morbidity, both physical and mental. Other work has confirmed the strong association of such behaviours with mental health problems. Nurses are much more effective than paraprofessionals (Olds et al. 2002), and continuity of care has been found to be crucial. Olds' work is unusual in that study participants were followed up meticulously for many years, and there are no other examples of such rigorous assessments of nursing interventions. There are, however, excellent evaluations of early nursery-based interventions, most notably the Carolina Abecedarian project (Campbell et al. 2002), which produced dramatic long-term benefits in academic achievement and problem substance use. The Scottish Collaboration for Public Health Research and Policy has recently published an excellent summary of the impact of early interventions (Geddes, Haw and Frank 2010).

It has not proved possible to achieve results as good as Olds' when interventions have been offered to teenagers to reduce unhealthy behaviours such as smoking, drug use and high-risk sexual behaviour. Other compelling evidence suggests that violence and antisocial behaviour is best tackled in the pre-school years (Wilson et al. 2009). It seems that we 'learn' how to be violent very early (Tremblay 2008) and it is also easier to undo behaviour patterns at a young age.

There are many examples of problems where intervening early is better than intervening late, but language delay is a particularly clear example. There are 'critical periods' (Bailey et al. 2001) in language development, and if we miss the window of opportunity, a child's language will be permanently impaired. Also, early language delay is a very powerful marker of psychological vulnerability. In a large Swedish study (Miniscalco et al. 2006), children aged 30 months, who could not make two-word utterances and who had fewer than 50 words, had a 70% probability of having a psychiatric diagnosis, most commonly autism spectrum problems or ADHD, at age seven.

Language delay is a potential sign of neglect - most young looked-after children have such problems, and language problems are extremely common in children excluded from school (Ripley and Yuill 2005). So language delay is a very important early warning sign, and does not just go away. We must identify and assess these children quickly and carefully, and offer intervention (usually more than speech and language therapy) if we are to avoid major future problems.

Patterns of vulnerability

Our research group at the University of Glasgow recently conducted a pilot project in which families were offered two new universal contacts with their health visitor when children reached 13 months and 30 months.

The project used structured assessment tools because this is the best way to ensure social equity; otherwise there is a high risk that interventions would be offered to those who least need them as detailed in the 'inverse care law' (Hart 1971). The work was designed to assess need (including unmet need) for parenting support in the community and offer appropriate levels of service to families. We assessed parental wellbeing and the parent-child relationship (Wilson et al. 2010) at 13 months; language delay through a two-question screen (Miniscalco et al. 2006), behaviour problems and parental stress at 30 months; and family background and demographic factors at both ages.

We identified a great deal of previously unsuspected need. For example, 8% of parents who had been assigned to the lowest risk category by health visitors had strong evidence of depression. At 30 months, 10% of children were found to have some degree of suspected language delay: 47% of these children had been assessed as being at low risk at the start of the visit.

Further work has been conducted with the Scottish Government and Glasgow City Council to develop the assessment of children's emotional and behavioural wellbeing at school entry using the 'strengths and difficulties questionnaire' (Goodman 1997; Goodman 2001; Youth in Mind n.d.). This data has allowed us to describe the emotional and behavioural wellbeing of children entering school in Glasgow. We now have maps of the distribution of emotional problems, hyperactivity/ inattention problems, conduct problems, and peer relationship problems across Glasgow The prevalence of conduct and hyperactivity problems is roughly 50% higher in the most deprived parts of the city compared to the most affluent, but some of the most deprived areas appear to have excellent childhood mental health. The data will allow us to identify local and individual factors predictive of problems likely to interfere with children's school attainment and will provide a baseline for future comparisons.

How should we respond in the UK?

It is important to view the great achievements of David Olds' Nurse-Family Partnership in context. In the US, there is no universal health visiting service and, consequently, no mechanism for identifying actual need in individual families in the community. Offering the Nurse-Family Partnership intervention to all families is clearly impractical, expensive and unjustified. It has to be a targeted, rather than universal, provision. Directing attention to families on the basis of predicted vulnerability (using, for example, lone parent status, teenage pregnancy, and economic adversity) without further assessment is inefficient at best: it gives resources to families who do not need them, and misses many children with substantial need who do not fall into the 'right' demographic group. Our recent evidence from Glasgow confirms this view. We have the potential for an efficient and flexible use of resources through an 'active filtering' approach in which professionals and families together determine level of need with reference to standardised assessment tools (Wilson et al. 2008a). Resources can thus be directed to those most in need. In other words, we need an intelligent system for 'case-finding', an assessment of the level of child/family need and appropriate resource allocation, often called progressive universalism.

The professionals routinely in contact with all children under the age of three years are: midwives (usually until the child is 10-28 days old); general practitioners (GPs); and health visitors. Each has the advantage of universal access and, consequently, contact with them is not associated with stigma. These professionals lack routine training in infant mental health, and profess a desire to learn more (Scottish Needs Assessment Programme 2003). In recent years, several policy developments have tended to reduce GP and health visitor involvement with the preventative care of children to the extent that many children do not see either, except on an opportunistic basis (for example, during illness) after the age of four months. One argument for universally offered, regular, child health surveillance contact with both sets of professionals is that there is strong evidence that vulnerability is not a static characteristic, but can become apparent at any time in a child's early years (Wright et al. 2009). There have been some welcome recent developments both in England, where a commitment to increase health visitor numbers has been announced (Department of Health 2010), and in Scotland, with a new mandate for universal contact with children, focused on language and behaviour.

Finally, once vulnerability is established, there must be clear care pathways available to families, with almost immediate accessibility. There is no excuse for a wait-and-see policy in early childhood social and emotional development.

Conclusions

There is a large body of evidence that wise investment to improve outcomes for pre- school children can pay a rich social dividend. The evidence available (Geddes, Haw and Frank 2010), supports the intuitive view that we get the biggest payback from investing more in supporting those children with the biggest needs. However, as Marmot makes clear, it is also important that these targeted services should be underpinned by universal services, the need for which is less intuitive. The scale and intensity of services needs to be proportionate to the level of need.

Assessing need, however, requires targeted investment in the universal services for children under three - health visiting and general practice.

There needs to be a commitment to training and professional supervision, to universal health surveillance programmes involving direct contact with children, and to functioning information systems. In the interests of social equity, there is a strong case for universal assessments using validated tools to assess need at several stages in the pre-school years. This would bring benefits beyond facilitating an equitable approach to support. It could allow efficient information sharing between primary care professionals, currently beset with problems (Wilson and Mullin 2010; Wilson et al. 2008b), and help policymakers and managers evaluate how well early years' services are performing.

Professional training for health visitors should have a strong focus on infant and child mental health and early brain development (Wilson et al. 2008b; Wilson et al. 2008c). Universal adoption of the Solihull Approach to infant mental health (Blackwell 2004) would promote effective communication among professionals about the mental wellbeing of pre-school children.

Child neglect needs to be identified early and addressed wherever possible, and certainly before the child begins to display serious problems which are difficult to contain. Behavioural symptoms in pre-school children should be taken as seriously as they are in adolescents and adults. Persistent aggression and indiscriminate friendliness, for example, should provoke professional concern and detailed investigation.

Once identified, there is a need to provide care pathways for problems such as child neglect and language delay. We are currently providing inadequate services to children with these difficulties.

* A version of this article first appeared in Thinking Ahead: why we need to improve children's mental health and wellbeing, a Faculty of Public Health publication 2011. Thank you to the Faculty of Public Health for permission to reproduce this article.

† In the USA, this programme is called the Nurse-Family Partnership. In the UK, it is generally referred to as the Family Nurse Partnership.