A collaborative endeavour between a technology business, a university and the NHS
is seeking to help speech and language therapists identify the 2 children in every classroom with the
often hidden disability of Developmental Language Disorder.
Developmental language disorder (DLD) is a condition where children have problems understanding and/or using spoken language. There is no obvious reason for these difficulties, for example, there is no hearing problem or physical disability that explains them.
Children with DLD often go unrecognised and without appropriate support can go on to have significant difficulties with education and social relationships with consequences for their wellbeing. However, access to specialist support and interventions at the right time can make a big difference, and reduce the chances of these negative consequences.
Funded by the National Institute for Health Research’s Invention for Innovation programme, specialist technology company Therapy Box is leading a collaboration to take their machine learning based app to the clinic so that speech and language therapists in England and Wales can more quickly assess children’s language, supporting them to identify children with DLD and then plan the support that child needs. The game-based app will present an animated story to be retold by children aged four to eight, and the child’s spoken story will be analysed to understand how their performance stacks up to age-matched peers from around the country.
The innovative collaboration between Therapy Box, the Bristol Speech and Language Therapy Research Unit, Newcastle University and three NHS sites in Hackney, Newcastle and Bristol will take place across two years. At the start of the project, the team will be seeking 600 children to record their stories, to help train the app to recognise the children having trouble with their language. In the second year of the project, the app will be put through its paces by speech and language therapists and families coming to clinics for assessment.
“Our aim is to help speech and language therapists evaluate more children in a shorter space of time,” project principal investigator Rebecca Bright shares, “The ambition for this project is to help children with DLD get the support they need as soon as possible”. Dr Yvonne Wren from the Bristol Speech and Language Therapy Research Unit is leading the evaluation of the clinical decision making tool, “The team involved in this project will ensure that we develop a tool which is fit for purpose and has been robustly evaluated, fulfilling the need for evidence-based tools in the assessment of children’s language skills.”
Dr Cristina Mckean whose team at Newcastle University will be working on the project, adds, “Identifying children with Language Disorder and designing tailored interventions to meet individual children’s needs is a highly specialised and time-consuming process. Together with the Speech and Language Therapy Service at Newcastle Hospitals NHS Trust, I am excited to be part of the team harnessing the potential of technologies to make access to these specialist assessments more time-efficient and informative. In this way, our hope is that more children with Language Disorders can access the support they need to overcome their difficulties and thrive” “As the Head of Service for a large, busy Speech and Language Therapy Service, I welcome a digital innovation that will support clinicians to quickly identify children with developmental language disorders,” adds Geraldine Bates, Head of Service at Bristol Community Health, “This will save precious clinical time, that can be redirected into improving access to therapy”.
The NIHR was established in 2006 to improve the health and wealth of the nation through research and is funded by the Department of Health and Social Care. In addition to its national role, the NIHR supports applied health research for the direct and primary benefit of people in low- and middle-income countries, using UK aid from the UK government. Award code: NIHR200889
We are so delighted to have been rewarded by the Digital Public Health
Innovation Prize last week in Marseille where we went,
last November, to introduce the Language Explorer project
The DPH Innovation Prize marks a celebration of success, innovation and business in digital public health whether that be through the establishment of game-changing partnerships across sectors/disciplines, creation of cutting-edge technologies or novel use of big data or data systems.
Read more about the awards and the Digital Public Health event here https://www.acm-digitalhealth.org/daily-agenda/dph-2019-innovation-prize/
Rebecca and Lucia will be at the 2019 ASHA Convention with information on hand about The Language Explorer project. We’d be keen to talk to researchers and clinicians from around the world about how The Language Explorer could be standardised for different regions. To set up a meeting email Rebecca at firstname.lastname@example.org
Last week, Therapy Box hosted in London the kick-off meeting of ATLAS project. The purpose of this day
to gather the different partners and discuss the plan for this 2-year project funded by the NIHR
for Innovation programme; starting with demonstrating the progress made in the development of Language
Explorer, an engaging app aiming to collect 600 language samples from children across the UK and Wales.
ATLAS project is a unique collaboration. The team is multidisciplinary and includes app developers, data scientists, user experience designers, academics and clinicians from Therapy Box, Bristol Speech and Language Therapy Research Unit, Newcastle University and three NHS sites in Hackney, Newcastle, and Bristol. Together we are aiming to create an app tool that would support Speech and Language Therapists in the assessment of children’s language skills and detect quicker potential Developmental Language Disorders.
The use of technology to support child learning is becoming increasingly prevalent. There are now apps
available and in development which can be used for a variety of purposes, including the assessment of
communication needs. So where does an actual Speech and LanguageTherapist (SLT) fit into this? If there
is an app to assess a child’s needs, do we even need a SLT at all? We might ponder this question in
relation to apps such as Language Explorer.
In the current climate, SLTs face a number of challenges- none more so than time. Budget cuts and increased caseloads have sadly resulted in pressurised staff having to provide high-quality assessments within a limited timeframe. This is where our new friend, technology, comes in. SLTs might use technology to provide the solution to such challenges; this includes the development of assessment apps.
So how would an app like Language Explorer ‘save time’ for SLTs? Firstly, it records andtranscribes what the child says electronically, saving valuable SLT transcription time. Secondly, it analyses this data and provides a useful, succinct report detailing key aspects of the child’s grammar and vocabulary. One might question that it is not only time that is saved here- but also SLT access, as surely they are not needed if the app has done their work for them?
SLTs are most definitely still needed. Yes- an assessment app may be quicker at recording and analysing language samples. This does not mean to say that a trained ear is not needed though; we all know that technology is not perfect. This is why in Language Explorer the SLT has the opportunity to appraise the language sample and make alterations based on subtle and nuanced observations that are only evident to the trained ear.
We must also remember that the SLT role in assessment is not simply deduced to the analysis of data; there is much more to it than that. It is what they do with this information, what it really means, that matters. Ultimately, it is about the human beings behind the data. What is the impact on the child’s participation in class? What difficulties are effecting their friendships? Whatdoes the child see as their main area of need? One of the most vital skills a SLT has is the ability to look at the bigger picture beyond the impairment; a computer simply cannot do this.
So, what can we do going forward? Let’s consider how technology can enhance the work SLTs do, rather than replace it. Imagine the impact of saving valuable time and how this time might be utilised; shorter waiting lists, increased intervention time, easier access to services. We can combine the relatively new benefits of technology with the unique and irreplaceable skills of the SLT to improve the quality of life for our children. After all, isn’t that really what it’s all about?
Specialist SLT and NIHR intern
Despite Developmental Language Disorders (DLD) being prevalent amongst Children in the United Kingdom,
with research finding that on average 2-3 children in each classroom being afflicted by the condition
(Lyons and Roulstone 2018), many misconceptions around DLD still exist. A combination of the fact that
DLD have no known cause and are often difficult to identify,has resulted in a poor understanding of the
condition (Bishop et al, 2016). In this article, we seek to tackle five common misunderstandings
Myth 1):Being Bilingualism increases the likelihood of having a DLD
A prominent concern amongst bi-lingual parents is that talking to their child in two languages can stunt their language development. This belief stems from the notion that instead of developing a child’s knowledge of two languages, it can stunt their development in both.However, no empirical study has ever found any truth to this claim (Petitto & Holowka, 2002; Paradis, Genesee, & Crago, 2010). Children require high-quality language input from their parents, regardless of the language this comes from.
Myth 2): DLD is a fancy term for disruptive
Those who are not informed on the causes and consequences of DLD are often tempted to simply class Children living with the condition as naughty or lazy (Guardian 2016). They view the condition as an excuse made for trouble some Children.
However, this is not the case. Children with DLD struggle to understand instructions given to them, especially when these instructions are spoken quickly or in an environment where there are plenty of distractions. This is coupled with the fact that it takes them longer to formulate a response to the instruction than the average child. As a result, the Child’s inability to comply with the task set out for them is often misconstrued as disobedience, when in actuality it is dueto their condition.
Myth 3): People with DLD aren’t intelligent
DLD is not an intellectual disability and therefore, doesn’t prevent a person with this condition from being smart (DLDandME 2019). Instead, it impacts a person’s ability to comprehend conversations and express themselves.
Myth 4): They will grow out of it?
Some parents are tempted to believe that their child’s language difficulties are something they will simply grow out of. Like Maggie Simpson, they expect that one day their child will simply start talking fluent prose and will catch up with their peers quickly. However, this typically isn’t the case. Research has found that on average children with DLD have a 2-3 year gap in Language abilities in the first 3 years of Primary school, and struggle to make this up this difference as they progress (Guardian 2017). Hence it is vital to identify and treat DLD as soon as possible.
Myth 5): Oh it must be the parents’ fault!
Another common misconception surrounding DLD is that they stem from bad parenting. People often are quick to blame the parents for a child’s language difficulties, believing they simply don’t talk to them enough (BBC 2012). However, this assumption is not correct. Although the causes of DLD are still some what unclear, we do know that DLD are primarily caused by genetics (Smith 2007). However, this isn’t to say that parents can’t be a part of the solution, and there are many ways that parents can help a child improve their language skills.
Analysts and people alike are increasingly aware that we are on the brink of the 4th
(Barr 2018). Healthcare is not an exception, with AI and Machine learning expected to play a prominent
role in the coming years. In this article, I will consider some of the main potential opportunities and
barriers to AI adoption in the NHS.
1) NHS and
openness to new
Government departments are typically risk-averse (Mcfaden 2016), there has been an acknowledgement of
need to embrace technologies and the benefits they can bring. Simon Stevens, the chief executive of
NHS, has called on tech firms to lead the charge for making UK Healthcare a World Leader in the use
and machine learning in healthcare provision (NHS News 2019). In addition, the NHS Long Term Planhas
advocated the role technology can play in helping them achieve their targets of reducing doctor
boosting efficiency and improving care (NHS 2019, p.6).Achieving these goals is becoming more of a
necessity due to growing budgetary constraints and population size.
Alongside the willingness to work with Tech companies to improve patient outcomes, the UK Government is investing heavily in improving the infrastructure necessary for such collaborations. In 2019 the Government pledged £250m to encourage the NHS to adopt AI, and in 2018 opened up 5 new Medical Centres across the country to research ways to use AI to diagnose diseases (Telegraph 2018; BBC 2019).
2) Availability of Data
Even the least tech-savvy individual understands that technology is reliant on data to function effectively. Health tech functions in very much the same way, with massive amounts of data being required for machine-learning algorithms to learn medical practice. Luckily for tech firms, the NHS holds massives amount of data on its patients,which in turn can be fed to algorithms to improve their functionality. Therefore, the opportunity exists for collaborations between the NHS and Tech firms to use this data to maximise the benefits of AI and machine learning technologies.
1) Data Security
While I’ve mentioned the potential benefits of using NHS patient data for AI, a number of risks surrounding the use of personal data also exist. Many will be aware of the suspicion surrounding Tech Companies use of personal data. Consequently, there are concerns about the NHS sharing data with private firms and the transparency of the process of data sharing (Steveton et al. 2019). For instance, the collaboration between the Royal Free Hospital and Google’s division Deep Mind was criticised for their handling of data of 1.4 million patients (PWC 2017). Therefore, it is vital that suitable Data Protection procedures are developed to deal with these collaborations.
2) Skill Gap
Another barrier to consider is that the NHS workforce simply isn’t ready to embrace these technologies. Although tech-savvy software developers may preach the benefits of health tech, in many cases the NHS neither has the infrastructure or the knowledge amongst the workforce to utilise some of the technologies. 69% of NHS staff have yet to undertake any form of training relating to how to use new technologies in their daily jobs(Hughes 2019). This epitomises the current skills gap in the NHS. Without training, AI adoption in the NHS is likely to be slowed down.
It is clear that AI and Machine learning are going to play a large role in healthcare for years to come. AI and Machine Learning technologies can improve efficiency, patient outcomes andquality of care when adopted right. However, in these early stages, there are still a number of barriers to overcome!
Having a Developmental Language Disorder (DLD) can inhibit a number of elements of a child’s non-physical
development. Language is essential to every aspect of human life, from social interactions to education,
which makes the early treatment of language disorders vital in orderfor a child to maximise their
While many are aware of the impact DLDs can have on a child’s learning outcomes, especially in relation to their literacy and comprehension skills, what is less known are the negative socialand emotional consequences. For instance, a report published by the Early Intervention Foundation found that alongside being detrimental to a child’s education and employment prospects, that DLD also increase the likelihood of antisocial behaviour and mental health issues in adulthood (Law, Charlton and Asmussen 2017). For instance, 74% of those in young offending institutions in the UK have communication difficulties, and 60% have issues surrounding language, communication and literacy (Law, Charlton and Asmussen 2017, p.26).Alongside this, DLD are also especially prevalent amongst Low social-economic groups.18–31% of children aged 19–21 months living in low social-economic households have been found to have language delay that warrants referral for specialist assessment (Law, Charltonand Asmussen 2017, p.7). DLD can act as a form of inequality trap, due to their impact on a child’s educational achievement and development.
Given the adverse consequences of an undiagnosed, and therefore, untreated language disorder, it is vital to identify DLD as soon as possible. As a child gets older, the impacts of DLD become more entrenched, and the effectiveness of Speech and Language Therapy decreases. A child can quickly fall behind if speech and language learning is delayed (ASHA 2012). Early identification increases the chances for improving communication skills by helping develop language when a child’s brain is most malleable and absorbent to learning (ASHA 2012). The earlier a child is exposed to Speech and Language Therapy, the more likely the risks associated with DLD are to be mitigated.
Consequently, the team here at Therapy Box embarked on the Atlas project, to improve the existing means of identifying DLD.
From phonics to prepositions, the insurgence of technology-based learning has gradually changed the shape
of Speech and Language Therapy over recent years. This cultural shift maybe regarded with excitement or
apprehension within the profession- for what is the actual evidence behind our clinical decision making
when recommending our favourite apps?
In a world where information is so readily accessible, it can be surprisingly difficult to pinpoint the level of evidence behind some communication apps. With a largely unregulated market and many apps claiming to be effective, it is more important than ever to delve deeper into the ‘why’s’ and the ‘how’s’.
However, a lack of transparency into how some apps are developed and evaluated is a significant barrier to this. As with any form of clinical input, it is important that we evaluate the existing evidence in order to make an informed clinical choice. We attempt to do this for other interventions we provide- so why not for the apps that we use?
One thing we can do is apply appraisal tools that are already available to us. For apps which have published research, we can use the free resources on the CASP website to guide our evaluation of this evidence. The check lists provide a useful tool for us to evaluate evidence at various levels, including Systematic Reviews, Cohort Studies, Randomised Control Trials and Case-Control Studies.
For apps with no published work, this is trickier. We might ask ourselves why there isn’t an evidence base- is it because the creators haven’t allocated the time and resources to do this? This is not to say that apps without published research are definitely ineffective- within our profession, we can often apply interventions where we simply know it works based on our experience. However, what we must do is be aware that our personal experience, as valuable as it is, is not a substitute for a rigorous and methodologically sound study. The Communication Trust ‘What Works’ database is an invaluable tool in this respect. It allows Clinicians to inform themselves about the variety of interventions available, whilst maintaining an awareness of the differing levels of evidence behind them. Importantly, the database now also includes app/computer-based interventions.
This is an exciting time for Speech and Language Therapy; we have at our disposal an increasing number of apps and technology-based programmes which aim to make our lives easier- and more importantly the lives of our children. However, we must continue to question, consider and challenge the interventions we provide, regardless of whether they are computer-based or not. Although a challenge in itself, we can aspire to do this by using the resources we already have available to us.
Specialist SLT and NIHR intern