(U.S. Air Force photo by Staff Sgt. Jake Barreiro)

Article by Dennis P. Wall

In the last two decades, autism prevalence has increased by more than 600%. Although Autism manifests early in development, the average age of diagnosis in the US continues to hover stubbornly above four.

With the incidence of autism now at one in 68 and climbing – due to:

  1. Changes in diagnosis
  2. Greater awareness
  3. Parental age
  4. Anyone’s guess

– the need for solutions that reach the risk population faster and more effectively has never been more real.

For most families the diagnostic odyssey of autism has more stops and starts than the 16 Odysseus endured, with two main sirens: the initial detection of risk and the clinical confirmation of risk. The path to each is long.

The first is not standardized and often rests on the family’s shoulders. The second has more structure, but the process is cumbersome and scales poorly. Worse, these two steps are generally decoupled in today’s system of healthcare.

As the incidence rate of clinical confirmation of risk converges to roughly 2%, within five years the prevalence of autism could increase in the US by as much as five million individuals.

This translates to about one million children a year requiring conservatively about four hours of clinical attention to reach a diagnosis, roughly four million clinical hours in total, which breaks down to about 1,400 years worth of eight-hour work days.

Just by simple back-of-the-napkin calculations, it is clear the system is setup to fail families.

There are approximately five developmental pediatricians for every 10,000 kids at risk for a diagnosis of autism. Just by simple back-of-the-napkin calculations, it is clear the system is setup to fail families. The risk population sharply outnumbers the clinicians available to administer the standard-of-care.

With this kind of math, we might even be “happily” surprised to learn the average wait time to receive the official clinical evaluation is about 13 months, but clearly without new approaches this timeframe will increase, the waiting lists will continue to grow and the time to reach a diagnosis that in many cases activates therapy will continue to climb.

With respect to getting to and beyond the first stop in the diagnostic odyssey – detection of risk – leaders in the field have rightfully stressed that all children should be screened for autism at routine 18 and 24 month well checkups.

Many valuable screening tools have been developed for early screens of risk [1]. These include the Social Responsiveness Scales (SRS), the Social Communication Questionnaire (SCQ), and the new version of the Modified Checklist for Autism in Toddlers (M-CHAT) [2].

All have merits but, despite endorsement of M-CHAT by the American Academy of Pediatrics, the adoption rate in general pediatric care remains limited. One reason for this has been low accuracy, and particularly low specificity.

If we can agree the field is now equipped with a clinically acceptable screen(s) for autism, the next immediate area of focus must be on adoption.

The Autism Diagnostic Observation Schedule (ADOS) has become the de facto standard for aiding diagnosis for an extremely good reason. It involves a direct observation of the child, which is vital to understanding risk.

But since ADOS must be administered by clinicians at clinics, its use simply cannot scale to the size of the risk population.

If we can agree the field is now equipped with a clinically acceptable screen(s) for autism, the next immediate area of focus must be on adoption.

A primary reason why screeners face challenges with consistent and standardized adoption (during windows of development that matter most for the child and the family) is because of issues with access and ease of use.

Mobile tools can help. Approximately 1.75 Billion people have a smart phone with video and sophisticated “app” capabilities. Millions, including an increasing diversity of the globe’s socioeconomic ecosystem, are embracing the mobile health (mHealth) revolution.

This revolution is marshalling in FitBit, Google Glass, Sproutling, Nest [3], technologies, devices and wearables that are meteorically restructuring how humans manage health.

The mHealth market is projected to reach $49.1 billion in 2020. These tools present an unparalleled potential to improve patient engagement and care at substantially reduced costs.

We have an opportunity to create a way for families not only to have a one-time screen but also to generate and store information about amplitude of risk over multiple time points.

To reach the booming population of children at risk for developmental delay and autism, we must find mHealth solutions that promote greater adoption and consistent screening (minimally) at 18 and 24 months.

Creating a simple app for direct-to-parent screening would not only encourage adoption but would enable structured, consistent and straightforward screening.

Tethering this app to persistent and secure cloud storage would provide a fully mobile way to save the results – minimally a risk profile, breakdown on behaviors from a questionnaire, and short home videos of their child – and give parents personally controlled medical records that they store privately but can also share automatically and as they choose.

For example, the parent could safely share the record with their primary care doctor or child specialist at a developmental medicine center who then can immediately examine it in the context of other records. This gives both parties powerful knowledge faster and more conveniently than is possible today.

Importantly, an accurate, fast and simple mobile solution that families can use from home need not be used only once. mHealth tools are designed to be used repeatedly.

We have an opportunity to create a way for families not only to have a one-time screen but also to generate and store information about amplitude of risk over multiple time points. This creates the potential for monitoring during therapy to track progress.

Another enormous opportunity with mobile systems for risk detection and monitoring of developmental delay and autism comes in the form of data.

The earlier parents receive actionable information and move through the diagnosis and to the therapeutic starting line, the sooner they can begin to engage in activities that equip them and their child with invaluable skills to thrive.

If the families used a mobile app(s) to capture every risk assessment, every clinical diagnosis, and a standard set of data points along a time continuum as their child progresses through behavioral therapy, we would – relatively quickly and cheaply – have the “big data” necessary to create powerful predictors and much more.

Giving families a way to communicate a standard outcome to their physician could have considerable impact on arriving at the first stop on the diagnostic odyssey – detection of risk – earlier and more systematically than what is possible today.

Yet, to reach the second stop – clinical confirmation of risk – and achieve a meaningful reduction in the average age of diagnosis, a better bridge between the families and the clinicians must be made.

artificial intelligence medicine mobile autism

[Figure.  Kids like mine, patients like mine. Data delivered immediately to parents and families with a handheld…

Forming the bridge requires at least three pieces:

  • First the clinicians must be able to get the results from a screen. Empowering families with personally controlled risk data that they can deliver to their pediatrician could be a powerful way forward.69% of US adults are willing to communicate with providers by email, and as much as 40% are willing to do so by mHealth applications. 84% of doctors use tablets. Innovation is occurring both in private and public sectors (e.g. the Blue Button Initiative [4]) that will serve to increase these percentages and increase the potential for families to equip doctors with data on which they can act.
  • Second, the clinicians must understand the pluses and minuses of the screen. Even if the screen provides just a few outcomes, such as high, medium and low risk, this may be enough to give pediatricians greater confidence overall and a set of clearly structured next steps.
  • Finally, the bridge requires that clinicians can use the results from a screen. A focus of my own work has been to move towards quantitative scales of risk, probabilities that have more information than a simple yes or no and instead provide a digital phenotype.

These probabilities can already be plotted into a distribution [see Figure] that provides statistical context and a more data-driven understanding of the risk, e.g. “patients like mine”.

Visual depiction of a child within a distribution can give clinicians a more precise way to react, for example to triage children off the waiting lists and onto one of several potential paths. This kind of picture may resonate with clinicians in a way that assists rather than interrupts their daily flow.

The human touch and the clinicians needn’t and shouldn’t be removed from the process, but the earlier parents receive actionable information and move through the diagnosis and to the therapeutic starting line, the sooner they can begin to engage in activities that equip them and their child with invaluable skills to thrive.

The sooner we can install a solution that ends the diagnostic odyssey the sooner we can shift more of our collective focus and our funding to what happens after a diagnosis and developing tele-medical, mobilized, e-health solutions for therapy that can reach remote populations both inside and outside of the US.

 

FOOTNOTES

[1] [http://sfari.org/news-and-opinion/news/2011/how-many-tests-does-it-take-to-diagnose-autism]

[2] http://sfari.org/news-and-opinion/viewpoint/2014/no-ideal-tissue-for-gene-expression-studies-of-autism

[3] http://techcrunch.com/2013/09/16/wearable-baby-monitor-developer-sproutling-raises-2-6m-from-first-round-and-others-to-raise-parenting-iq/

[4] http://www.healthit.gov/patients-families/blue-button/about-blue-button

 

dennis wall stanford artificial intelligence medicine

Dennis Wall is an Associate Professor of Pediatrics at the Stanford University School of Medicine, where his lab is developing novel approaches in systems biology to decipher the molecular pathology of autism spectrum disorder and related neurological disorders.

Dr. Wall received his doctorate in Integrative Biology from the University of California, Berkeley, where he pioneered the use of fast evolving gene sequences to trace population-scale diversification across islands. Then, with a postdoctoral fellowship award from the National Science Foundation, he went on to Stanford University to address broader questions in systems biology and computational genomics, work that resulted in comprehensive functional models for both protein mutation and protein interaction.

Dr. Wall has acted as science advisor to several biotechnology and pharmaceutical companies, has developed cutting-edge approaches to cloud computing, and has received numerous awards, including an NSF postdoctoral fellowship, the Fred R. Cagle Award for Outstanding Achievement in Biology, the Vice Chancellor’s Award for Research, three awards for excellence in teaching, and the Harvard Medical School Leadership award.