Go-Aut, Early diagnosis of neurodevelopmental disorders

Being able to diagnose childhood neurodevelopmental disorders (NDD) at an early stage is crucial for kids to achieve a better quality of life. Parents and paediatricians around the world are missing the early signals due to their own biases or lack of knowledge.

We offer a semestral AI-based online assessment that runs over the first 5 years of the child for early identification of risk signals of neurodevelopmental disorders. If the child is at risk, a referral to a certified professional is issued.

The potential of the solution is huge in impact and distribution. Impact is big because NDD cases that today go unnoticed will be diagnosed for rehab to start sooner and achieve better results. The potential of distribution is that, being a 100% online assessment, it can scale to every country and area of the world, even in rural areas far away from the big cities.

Neurodevelopmental disorders (NDD) start to show in children since they are as young as 2 months old. It is currently very hard for parents and paediatricians to identify the early signals of NDDs. Parents frequently struggle to accept that their kids are developing slower than the average. Paediatricians usually find at most only a couple of these cases per year, so it is hard for them to identify the signals as well due to the lack of exposure. All of these situations contribute to NDD being diagnosed at a later stage -maybe when the kid is 4 or 5years old- and it is too late for correct rehabilitation.

The most spread NDD are:

- Autistic spectrum disorders (ASD)

- Attention deficit/hyperactivity disorders (ADHD)

- Developmental intellectual disability

According to the site “Our world data”, in the world in 2016:

- 62 million have some Autistic spectrum disorder (0.8-1.1% of the population)

- 63 million had ADHD (0.5-2% of the population)

- 115 million had Developmental intellectual disability (0.4-3% of the population)

In Latin America, the total prevalence is about 2.15% of the population, which is about 13 million children.

We are serving children of ages 0-5 and their parents in Latin America (and later, the rest of the world). We want to improve their quality of life by allowing them to diagnose NDDs, get help earlier and get fully recovered, so they can avoid a lifetime of problems. 

In Latin America, there are currently 65M children aged 0-5, and around 11 million are born every year. All of them should get screened for NDDs as early as possible and Go-Aut is here to help.

Go-Aut will address children and parents needs by:

- Running regular assessments on children, with a minimum frequency of twice a year

- Early signalling and diagnosing children that are at risk.

- Helping children at risk to find help from a certified professional for treatment.

- Helping parents understand how to act early and help their children develop.

We offer a subscription service for scheduled risk assessments. The subscription can be bought directly by parents or children’s family (as a gift) or bought in bulk by paediatricians to outsource the diagnosis of NDDs or offer value-added services.

The scheduled risk assessments run minimum twice a year while the child is 3mo-5yrs old. A typical assessment includes some initial screening questions for parents and some tasks that the child needs to perform while being recorded by their parents. All the content is then uploaded online and a team of experts, aided by a machine learning algorithm, perform the risk assessment. Depending on the result of the assessment, extra follow up screenings might be scheduled.

The risk assessment has three possible outcomes for parents: 

1. The child is not at risk, for which we schedule a new assessment in around 6 months. 

2. The child needs monitoring, for which we schedule a new assessment in 1 month and we give homework for parents to do with their child.

3. The child is at risk and needs help. In this case, we offer the parents a set of certified professionals that can follow up the problem and provide treatment if needed.

Today, many parents don’t have access to top-tier health professionals with expertise in NDD. With Go-Aut, any parent with an internet connection -from Nigeria to USA, from rural areas to densely populated cities- will be able to get a NDD risk assessment for their children.

This solution is innovative because it changes the process in which the diagnosis of NDDs is done today, making it more scalable and accessible to everybody. The source of innovation, according to Drucker’s framework is “process needs”.

Today, the diagnosis of NDD is somehow random: parents see some signals, paediatricians might see some other while on appointments. The truth is that many NDD cases go unattended until a later stage when it’s too late for a full recovery of the child.

This new approach changes the current process in 3 different ways:

1. Better process: the process is now more scalable, due to its online nature and its machine learning models to aid decision making. This allows Go-Aut to evaluate children from anywhere in the world from a single hub, faster and at a much lower cost than any private practice.

2. Better Discipline: it’s usual for parents to forget when doctor appointments need to be made and many times they skip or miss the appointment. Go-Aut sets regular reminders for parents not to miss any assessment and follows up when needed.

3. More inclusive: thanks to its scalability and not requiring brick-and-mortar facilities, Go-Aut is much more efficient than private practices or hospitals. On the logistics side, parents will avoid travelling long distances to have an expert examine their children. On the cost side, we expect many more parents will be able to pay for the subscription than those that today can afford to pay for a NDD assessment for their children.

Technology is key for our success, as it is what will allow us to scale and have a bigger impact on society.

First of all, we will need a self-service application to allow parents to run the assessments and upload all the materials. We expect to develop a web application using Progressive Web Apps, and iOS and Android apps, most likely using Flutter.

Additionally, our NDD risk assessment will use technology to allow scalability, precision and cost reduction. We plan to use machine vision to interpret the videos that parents upload and speech recognition to run a voice test over toddlers older than 3 years. We expect these technologies to increase by 5-7x the amount of videos that can be analysed per hour.

On top of those technologies, we will build an AI expert system that will aid the decision making of our screening professionals that will use machine learning and classification algorithms. This system will learn from previous tests to make better predictions in the new assessments. The input of the model are the parent’s responses to the questionnaire and our professional’s assessment of the children videos. The outcome is child’s expected risk of suffering NDDs. This value will aid our professional’s decision on the final assessment of the child. This technology is expected to increase the precision of our assessment by 50% and improve 10x the speed in which it is run.

As Doctor McNeil mentions in his article (1), disorders affecting children and adolescents tend to present unique challenges to families, physicians, and others who seek answers and mitigation of symptoms. On the same line, the National Scientific Council on the Developing Child mentions (2) that “While a range of strategies is available to treat or prevent difficulties, diagnosis in early childhood can be even more difficult than it is in adults“. Both views validate our problem statement and the need for a solution.

There are many validated procedures to diagnose NDDs in children, our differentiation is the systematic process in which we follow up children, generate data and track their evolution.

Then, once the child has been diagnosed and is under treatment, the National Scientific Council on the Developing Child mentions (2) that the possibility of significant improvement in quality of life, as well as in both cognitive and social functioning, as a result of prompt intervention provides a strong argument for the early detection and treatment of these developmental disorders.

References:

• Disorders in Children and Adolescents: Challenges in Diagnosis and Treatment With an Emphasis on Neurodevelopmental Disorders, Shawn E. McNeil, M.D., The American Journal of Psychiatry Resident's Journal, Volume 12, Issue 9,  Sept. 2017

• National Scientific Council on the Developing Child. (2008/2012). Establishing a Level Foundation for Life: Mental Health Begins in Early Childhood: Working Paper No. 6. Updated Edition. Retrieved from www.developingchild.harvard.edu.

We are currently operating our beta program and we are testing it with 100 patients. In one year our goal is to have 7,000 subscribers in Latin America. In five years, we want to have 2,000,000 subscribers. After Latin America, the next countries to launch on are the US, Canada, India, Australia and New Zealand, some of the countries with the highest prevalence of NDDs in the world, according to ourworldindata.org. 

Our goal for the next year is to settle in Latin America as go-to-place for signalling NDDs. For that, we need to find product-market fit and find the right sales channels. It will be a stage of strong product development and sales exploration. We expect to get 7000 annual subscribers and revenue for around 1M$.

After the first year is over and as we achieve product-market fit, we want to double down on sales and launch in new countries. In years 2 to 5, we expect to experience hypergrowth as we scale the product to new markets. By the end of the fifth year we expect to have 2 million annual subscribers and revenues for 150M$.

The biggest barrier for scaling is legal. Since healthcare is a highly regulated industry and we are directly impacting people’s lives, in each country where we want to launch we need to make sure we are not generating liabilities for the company nor for ourselves. This is a barrier that can be avoided with enough money to pay for lawyers.

Another barrier that we need to overcome is the awareness of parents of the risk of NDDs. Not all parents know the importance of diagnosing NDDs as early as possible and this is a critical factor at the moment of purchasing our subscription.

The main way to overcome the legal barrier is to position our product as an NDD risk assessment as opposed to an NDD diagnosis. We will not diagnose NDDs but we will process all signals and point out children that might be at risk, to then refer them to certified professionals. The second way to overcome the barrier is by hiring a good buffet that can help us hedge against liabilities. For this, we will certainly need funds.

To overcome the awareness barrier we plan to have proper communication, that can be in the form of marketing or PR campaigns. To be successful in it we will also need funds.

Although overcoming both barriers will require funds, this funds are required at the moment of launching a new country. As such, we can control how and when we spend the money by deciding when to launch in new countries. We plan to raise capital to increase the speed at which we expand into new economies.

-

At this moment the Go-Aut team has 3 people:

1. Martin Panelati: full-time, leading the company and product development

2. Dr. Malena Corral: full-time, leading the clinical operations and the assessment development

3. Facundo Panelati: part-time, leading marketing

Dr. Corral has extensive work experience as a psychiatrist for children and adolescents. She has the right mix between old-school medicine and wild and innovative spirit. 

Martin Panelati is an MIT Sloan MBA Candidate 2021. He has been working for the last 5 years as a Product Manager in an online classifieds company for emerging markets, managing products with more than 80M monthly active users. In the past he had his own startup, pitched investors and raised money in Argentina.

Facundo Panelati leads sales intelligence at Mercado Libre (NASDAQ: MELI) and has 4 years of work experience in e-commerce brands, leading online marketing.

Together, we are looking forward to having a long term impact on society and to improve the lives of millions of children around the world.

We are currently partnered with FETEM (Foundation for the study and treatment of mental disorders, as per the spanish name). FETEM provides support and validation in clinical matters. For example, they contribute, review and provide peer validation for the NDDs risk assessment that Dr. Corral is leading.

Our main beneficiaries are the children that get a better health and quality of life after early diagnosis and treatment of their neurodevelopmental disorder (NDD), and their parents as they also get a better quality of life. For both of them, it will certainly be life changing to diagnose an NDD early enough. 

Paediatricians are also benefited in two ways: first, they make sure they don’t miss any diagnosis of an NDD in their patients. Second, they get more patients diagnosed with NDDs who will visit their practice more often.

Finally, companies are also benefited. Companies are interested in the long term wellbeing of their employees and their family, as happy employees are more productive. An employee with children with an NDD is an employee with less productive. It needs to leave early, it’s mind is on its child, the stress level is higher. If that disorder lasts forever, it means a perpetuity of loss for them. Early diagnosis then means for children, the opportunity of a full recovery, for parents lower stress levels and for companies more productive employees.

Our product is an ongoing NDD risk assessment that runs from 3 months to 5 years old, under a subscription model. We estimate the price point at between 100 and 150$ (US dollars) per year.

Parents can subscribe their children, family can give it as a gift for newborns, companies can offer it as a benefit for employees and paediatricians can buy bulk subscriptions for their practices at discounted prices.

In the beginning, we plan to apply for grants and raise some seed capital. We need money to develop our product and to market to the initial set of customers. The initial commercial strategy will be to partner with a key account (like a maternity hospital or paediatrics hospital) to gain volume and early revenue. 

Then, our goal is to achieve financial sustainability by selling our service at a price where customer lifetime value is higher than our marketing costs. If this proves successful, we expect to achieve break even during year 1. Then, we will only require raising capital to speed up our expansion process.

We are applying to Solve because we value the network and mentorship that we can get from it. Although Martin has experience as an entrepreneur and Malena has experience as a Psychiatrist evaluating and researching NDDs, we have never ventured in the healthcare industry. Solve will be able to get us mentorship in the healthcare industry that can provide valuable advice. Additionally, our network is big in Argentina but not enough in the US, where Solve can help the most. Finally, the ultimate growth hack for us would be to close a deal with a big paediatrics clinic that can use Solve to increase their NDDs diagnosis accuracy and treat more patients, or with an enterprise that is willing to add Go-Aut as an employee benefit. We believe Solve can help with these introductions as well.

-

In order to build our product, we would like to partner with:

• A legal partner that can help us understand the requirements to run a healthcare startup in the US.

• Some renowned paediatrician or child psychiatrist that can become our advisor or even a member of our board, to provide clinical guidance and professional reputation. This can be either in Latin America or the US.

• One or more Autism, Asperger, ADHD or Developmental intellectual disability non-profits, to help them with the screening of children at risk and get validation for our assessment.

In order to increase our sales and build a sustainable business, we would like to partner with:

• One or more enterprises that may be willing to include in their corporate benefits the subscription of their employees’ children.

• One or more paediatrics clinics that is willing to use our product to increase the accuracy of their NDDs diagnoses.

These two can be either in the US or any country in Latin America.

AI sits at the core of our business strategy. We aim to process 4M assessments/year in the next 5 years, and we believe it can only be achieved with AI aiding professionals in the processing of risk signals. 

Our neurodevelopmental disorders (NDD) risk assessment has two parts: parent questionnaire and child task assignments (video recorded). Our vision is that AI will help us in 3 main areas:

• Building an expert system to aid the decision making of our assessment professionals

• Evaluate children behavior and movement in assignment videos using computer vision.

• Evaluate children’s voice using speech recognition during a voice recording assignment.

We plan to start with a regression model of parent questionnaires as a cornerstone of the expert system. Then we will enhance the regression based on the operator’s observations in the videos. A classification model will follow, to point our areas where the operator needs to take a closer look. 

For computer vision, we want to compare how movements in children differ based on their NDDs risk. Detecting anomalies in children movements can help us improve the information we get from the assignment videos and get faster and more accurate assessments. Similarly, the same can be done with the children voice.

To conclude, AI is the technology that will help us scale to impact millions of lives and differentiate from traditional NDDs diagnoses, which are slow and human-error prone. We believe we can make our world better by improving the lives of children that suffer from NDDs.

-

AI sits in the core of our business strategy. We aim to process 4M risk assessments per year (coming from 2M subscriptions) in the next 5 years, and we believe it can only be achieved with AI aiding professionals in the processing of risk signals. 

Our neurodevelopmental disorders (NDD) risk assessment has two parts: parent questionnaire and child task assignments (video recorded). Our vision is that AI will help us in 3 main areas:

• Building an expert system to aid the decision making of our assessment professionals

• Evaluate children behavior and movement in assignment videos using computer vision.

• Evaluate children’s voice using speech recognition during a voice recording assignment.

We plan to start with a regression model of parent questionnaires as a cornerstone of the expert system. Then we will enhance the regression based on the operator’s observations in the videos.

For computer vision, we want to compare how movements in children differ based on their NDDs risk. Detecting anomalies in children movements can help us improve the information we get from the assignment videos and thus, get to faster and more accurate risk assessments. Similarly, the same can be done with the children voice.

To conclude, AI is the technology that will help us scale to impact millions of lives and differentiate from traditional NDDs diagnoses, which are slow and human-error prone. We believe Go-Aut leverages advanced artificial intelligence to create a differential and disruptive impact for our society, by improving the lives of children that suffer from NDDs.