[Posted on 24 October, 2017 by Phil Wilson]
“All screening programmes do harm; some do good as well, and, of these, some do more good than harm at reasonable cost. The first task of any public health service is to identify beneficial programmes by appraising the evidence.” – Muir Gray, 2008 (1)
In the light of this powerful statement by one of the world’s leading public health specialists, isn’t it odd that so many health service resources are devoted to child health checks when the evidence for their effectiveness is so weak? In the case of neurodevelopmental screening, the harm may be limited, perhaps restricted to parental anxiety in certain ‘false positive’ cases, but the costs of screening programmes are significant and we all have a duty to ensure that public funds are spent effectively.
To most people in Scandinavia, criticising the preventive child health programme is a bit like swearing in church – quite unthinkable! But in Scotland in 2006, the government completely abolished all universal health checks after the age of six weeks because of lack of evidence of effectiveness. The Scottish Government recently reinstated a preventive child health programme, but there are some very important lessons to be learned, and the most important of these is that we really need to work hard to gather robust evidence for the effectiveness of neurodevelopmental assessments in the whole population of children.
So why did Scotland get rid of child health checks? There were several reasons, including:
• For many of the conditions being assessed, there was no published evidence that early intervention is any better than just waiting for something bad to happen and then treating it.
• A view that well-informed parents could potentially identify many of the problems without a doctor or nurse being involved. In other words a programme of health education might be sufficient and parents would seek help if their children needed it.
• It was thought by policymakers that risk factors for many conditions would be known by the time that children were six weeks old – so child health nurses would be able to identify and follow up the children who were at risk of neurodevelopmental problems anyway.
• Finally, uptake of child health checks was very low among families living in areas of economic deprivation – so if the checks were effective they would increase social inequality!
Many clinicians in Scotland tried unsuccessfully at first to resist these changes but a crucial step in getting the government to reverse its policies was gathering research evidence to refute the arguments for abolition of the child health programme. One important example of this process came from studies by Lucy Thompson and other members of our group in Glasgow. We were able to produce irrefutable evidence first that severe language delay was going unidentified and untreated – in other words parents were not bringing children with big problems for help, and second, that health visitors were not able to identify the majority of high-risk children in early life. (2) The impact of this was felt in kindergartens where increasing numbers of children with very little language ability at age four were beginning to be seen, and kindergarten staff were struggling.
Now a sceptic might argue that language assessment does not meet international screening criteria: in fact a Cochrane Review recommended against universal language screening in 2003. (3) The main weakness in the argument for screening in this case is that no-one has yet published a decent trial comparing screening together with an early intervention programme with a “wait-and-see” approach. Some people might consider such a trial unethical but in many countries of the world there is no language screening at all, so it is difficult to see why such a trial couldn’t be conducted in one of those countries. In fact there is some emerging randomised trial evidence that early intervention in autism does make a long term difference (4) giving support to the view that autism screening may be a good use of resources, but such trials are difficult and costly, and should certainly not be an absolute requirement for justifying developmental surveillance. (5) That is not to say we shouldn’t try our best to produce the best evidence we can.
Equally importantly, some arguments about the effectiveness – or lack of effectiveness – of early intervention actually miss the point in relation to some problems. For example, it may not be possible to fix the underlying neurodevelopmental problems of a child with an IQ of 65, but knowledge that the child will have a terrible time when they start a normal school without special support could lead to appropriate educational planning which would make all the difference to that child. So the argument sometimes is not about the effectiveness of screening programme in classical terms, but more about the overall wellbeing, or even the human rights, of the child. Other arguments for the benefit of child health checks include non-disease-specific factors such as the creation of an environment or therapeutic relationship in which parents feel able to raise their developmental concerns, or the opportunities afforded for health promotion activities.
The fact that every country has its own unique approach to child health surveillance is testament to the weakness of the evidence base. Scotland proved to have been very wrong in getting rid of health checks in the way that it did – lack of evidence of effectiveness is not the same as evidence of lack of effectiveness – but we should aim to build an evidence base just in case someone, somewhere else, has the same bad idea. We should not stop doing development checks unless there is evidence of lack of effectiveness but it is about time we started to have some shared understanding about what we are trying to achieve. Then we should start assessing whether we are actually achieving it!