Multi-Disease Screening Tool

A consumable free, point of care, disease screening tool. Which leverages AI to interpret common digitised clinical sensors. Initially focused on providing underserviced patients access to an improved tuberculosis screening tool.

Braden van Breda

AI Diagnostics co-founder and study PI.

Our solution is positioned to overcome the problem of limited access to sufficiently sensitive and specific disease screening. See Definitions of sensitivity and specificity in the supplied insert. Initially, we are focused on alleviating these problems for the TB epidemic. 

An estimated 10 million people fall ill from TB annually, with roughly 1.4 million dying from the disease. The current diagnostic efforts miss an estimated 3 million patients annually.

South Africa (Third-World countries alike) have limited access to lung X-ray screening facilities and personal and are therefore unable to fulfil this WHO’s TB screening recommendations. Unfortunately, this is especially true for the rural low-income regions where TB is most prevalent. Instead, a symptom-based questionnaire (77% sensitive and 65% specific) is used to screen for TB.

In South Africa 360000 people fall ill from TB annually, of which an estimated 24% are left undiagnosed, predominantly due to poor screening sensitivity. This significantly contributes to the 60000 annual TB deaths.

South Africa spends an estimated £36.8 million on TB diagnosis. The insufficient screening specificity results in superfluous secondary diagnostic testing on TB-Negative patients, incurring unnecessary costs and capacity congesting. 

Figure 1 in the supplied insert represents this general screening problem.

There are three categories in which our solution is designed to serve.

• Beificery: TB sufferers and low-income and underserviced communities.
• Customer: National Departments of Health.
• User: Nurses and public health workers.

We will provide access to sufficiently sensitive screening, in the regions where it is most needed. This will result in earlier TB diagnosis for infected patients and thus earlier treatment. This will improve treatment outcomes while protecting the community by reducing disease transmission. We’ve followed both TB advocacy and the international organisation’s prescribed needs, however, more actual patient engagement could be incorporated.

An increase in screening specificity will serve the National Departments of Health by reducing the percentage of TB-Negative patients being sent for secondary TB diagnosis testing. Preserving the state TB budget will in turn also benefit TB sufferers as funds can be redirected to improving TB treatment etc. We have maintained an ongoing conversation with the state to ensure that we aline with their needs. 

The screening tool has been designed to empower the user by extending their diagnostic input while ensuring the tool remains fast and easy to use. Nurse input played an important role in the development of our product User Needs.

Our study results and methods will be published. Moreover, all our (patient-disassociated) surveillance data will be shared with the relevant health authorities alerting them of breakouts be it in TB, COVID-19 or the next pandemic. Any new model architecture will be made public. 

Our target population already uses the public clinics for their primary health care interaction. South African policy dictates that all public health clinic walk-ins are screened for TB. The current screening tool used is a symptom-based questionnaire which at the get-go misses 23% of TB-Positive patients. Our sensitivity feasibility study suggests we should be able to halve this number of patients missed.

TB has a significant link to poor social-economic environments. Moreover, the financial burden of the disease (out of work, travel for treatment etc.) is significant on the families of those infected. This interdependence clearly creates a nasty spiral. To combat this spiral we aim to capture these patients as early as possible improving treatment timelines and outcomes and thus improving patient morbidity, mortality while reducing the financial burden as well. Moreover, this early diagnosis and treatment works to reduce the chance of further spread of the disease to other vulnerable individuals in the communities.

First Year: 

Primarily focused on product and production development, verification and validation. The impact will be limited to the research papers detailing our findings. 

Next three years:

Horizontal scaling

After product launch throughout South Africa, we plan to spread geographically to the other top 7 contributors to the TB epidemic. In South Africa, we plan to have a separate company that is responsible for sales and installation. To support transformation and facilitate uptake and scaling in South African, our initial sales and installation satellite will be majority black-owned. This satellite company will license our technology and service. To facilitate uptake in new countries we will copy this arrangement by creating sales and installation satellites owned by individuals from the country in which we are looking to deploy.

With language updates, the core technology should scale easily. Small pilot studies may be required in new countries to ensure there are not significant race/area-related biases that affect the predictive accuracy.

Verticle scaling

New diagnostic potential using the same hardware will continually be investigated. Moreover, new modular hardware addition will also be investigated to improve and increase the range of diagnostic capabilities.

The first marker of success is to end the data collection study with a Sensitivity and Specificity in excess of 90% and 70%, respectively,  as per WHO's endorsement requirements. Sensitivity and Specificity will be determined using the K- fold cross-validation methods on the training database.

Maximising the capture and efficiency ratio will be the main metric indicators of success.

Capture Ratio: Number of bacteriological confirmation test conducted on TB-Positive patients vs the WHO's national TB incidence rate. 

Efficiency Ratio: Number of bacteriological confirmation test conducted on TB-Positive patients vs Number of bacteriological confirmation test conducted on TB-Negative patients.

The TB screening tool will ultimately be considered a success once it becomes obsolete.

Next year:

Financial barriers - we need to entice investment prior to any revenue. We will approach grants and investors that share the cause and see the potential social and financial returns.

Wasting and muscle atrophy are typical side effects of TB, leaving many patients very thin. Clear auscultation recording on the contoured ribcage of these patients challenging. This is overcome by our novel two-part impressionable sound conducing stethoscope diaphragm, presented in figure 6 of the supplied insert. 

Public clinics are typically noisy environment, therefore to give our model the best chance to identify TB specific signatures we have active noise cancelling. Using two microphones, with one directed at the stethoscope diaphragm and the other towards the external environment, we digitally scale down and subtract the external noise from the auscultation recording. 

South Africa does not regulate Class B devices. Thus to obtain home county approval we are following the FDA's 510K approval mechanism which South Africa recognises.

Next three years:

It may be difficult to convince foreign nations to invest and implement our solution. Therefore we plan to partner with sales and installation satellite companies owned by individuals who are from the country we plan on deploying in. 

Working for AI Diagnostics.

We are applying to the Trinity Challenge for financial support. If we could secure enough to allow us to complete the data collection study - £103,310 (without founder salaries). We believe we would be able to leverage the results thereof to attract more grant funding or equity investment.

Partnership with the likes of some of the Trinity Challenge members would certainly add to our company's actual and perceived credibility. Helping with future fundraising and approval efforts.

Secondly, our core team is primarily technically based. Winning this challenge could allow us to partner with organisations that specialise in clinical, legal etc. fields. This diversity in personnel and skillsets should help us identify our blind spots earlier.

Instead of picking specific partners, I'll state the three organisational categories in which we could most certainly look to partner:

Academic institution - support our clinical study setups and publications

Philanthropic or investment institution - to support us with funds

Database service companies - to support our database security needs