An interview with Dr. Joshua Kibera, founder and CEO of The Pathology Network.
The majority of hospitals and doctors in Africa do not have access to the diagnostic tests they need when faced with patients who have suspected cancer, autoimmune diseases, extrapulmonary tuberculosis, suspected leukaemia, lymphomas, etc. Doctors in those cases send samples to external labs and right now the process is chaotic. The Pathology Network connects hospitals to a network of advanced labs and pathologists for diagnostic testing to ensure that the reality where results get delayed or lost and samples get lost in transit no longer exists.
There is so much chaos in the market and The Pathology Network solves that through our online platform where hospitals simply order their tests from us. We take care of all the backend processes. We pick up the sample, choose the best lab, supervise the lab, and then assign the case to a pathologist who could be remote. The pathologist logs in online and reports the case. Everybody has their results and all of that is done within a set period of time. We have successfully reduced the turnaround time for one type of test called histology from a three- to four-week average to about seven days. And we now have a two-day option as well for urgent cases. We are currently serving around 40 medical facilities, over 100 doctors per month, with 10 active pathologists and have recently started working with our first international pathologist.
We have always wanted to develop AI solutions to address our internal issues, especially as we scale. When we are distributing work among two or three labs and 10 pathologists, we can do it manually. And we have a team of about 10 customer service and logistics staff who choose the best lab and the best pathologist and help to coordinate all the travel and logistics and sample movement. But as we scale to hundreds of pathologists, it’s just not possible to keep track of all the permutations and possible combinations of assignments. We need AI to help us choose the best lab and best pathologist for that specific sample.
Beyond that, there is use of language models in assisting our pathologists when they are writing up their comments and reports. We would like to use AI to help pathologists find the words that they want to say and write it in the most effective way. AI can help to contextualise effectively and take some of that thinking burden away from them, which then just speeds up their ability to sign out cases faster and allows them to spend time reading scans and operating at the top of their license.
Finally, we are increasingly using digital pathology and getting digital pathology images and those can help us in the future to develop AI that helps us in scanning and screening slides before the pathologists actually have a look at them. The pathologist can basically have a list of slides that they need to look at, and AI has already provided an initial diagnosis, so they can prioritise more serious cases. This makes a difference in the workflow and patients’ treatment and management.
We are also creating data platforms and looking at ways in which the data that we have can be anonymised and working with various government agencies and legal providers to find a way that we can make this data available to other players who are in the ecosystem who have AI models to train but don’t have access to the data that we have. I think it’s the right thing to do for the country and for the continent. We need to have all this African disease specific data available to inform AI in order to support medical researchers, doctors, medical students, etc.
At the beginning of 2023, The Pathology Network raised $650,000 in funding and were able to hire some amazing people to prepare the company to scale beyond Kenya. We had a checklist of things to be done with around 60 different action items and we are now around 70% done. This means that sometime early this year, we will be ready to serve a second country.
We have a Nairobi office, which is a resource center that other members of the diagnostic community within the industry also use to promote the adoption of digital processes in the diagnostic field. We have more clients in Nairobi, we have a new lab in Kisumu and we are geared up to begin validating pathologists for direct primary diagnosis in histology and cytology using digital pathology. We have also produced our first documentary, which talks about how The Pathology Network process has not only improved the diagnostic process for patients and doctors but has also improved surgical training by giving pathologists feedback on their work.
A lot of our challenges at The Pathology Network are mostly related to talent. Good talent is expensive and some of the skills we need are so unique that we end up having to train people in-house, which takes time. When we hire new people, it takes them three to six months to really understand what we’re doing and start having a visible impact on the business.
Additionally, the depressed economic environment within the country hasn’t helped. A lot of people are struggling to pay, contracts are taking longer to sign. But we are continuing on and know that we can overcome these challenges.
I have learned so much about the ins an outs of running a business, from hiring to marketing to fundraising and more. One thing I can say si that I think this is a good time to be an innovator. The barriers to entry are low. You don’t need to start making a big difference immediately. And there’s so much information out there, there’s so much opportunity and you really don’t need funding to get started. You need funding to grow, but I don’t think you need much funding to get started, especially in the service industry.
The other thing I would say is to be ready and willing to ask for help. I think that’s one thing that I got scared of in the early days. I could have asked more people for help. Whenever I have asked, people always have helped me. When I eventually asked, those people turned out to be awesome for us and it’s just by talking to people that you get the help that you need. And then always, always remember to pray. If you’re doing the right thing, God just ends up blessing that.
Well, first of all, the funding has really helped. It extended our runway, gave us the time to build out our tech, on top of which we’re going to build AI. It’s opened my eyes to the world of AI. I’ve learned a lot just by listening in and interacting with various players in the field and we’ve learned the opportunities that lie ahead, not just in developing our own AI but also being a source of useful information for people who need data training sets to develop their own AI. This is an area where we are in a position to really help.
I’ve learned that pathology AI is here today and it is not something that will happen sometime in 10 years. A recent development in Harvard University completely flipped my understanding of the role AI will play in the future of pathology. A team working at the university developed an AI algorithm which can read slides with an 86% accuracy. This is better than a lot of pathologists I know! It was very scary. I believe that the AI4H program primed me and prepared me for this moment, otherwise I would have had a heart attack. We have already reached out to the professor who developed the algorithm seeking for collaboration. We have also made a complete revision of our strategy to include AI into our next fundraise in a significant way rather than waiting to invest in it in 4 years.
The Pathology Network team is really grateful for this opportunity. NVIDIA has been amazing in offering the inception program and the credits and our mentors have been a wealth of information. So it’s been very good to learn from the team.
The AI4H program has helped me develop strategic thinking around the use of AI in our business. Without the AI4H program I don’t think I would have had the skills to deal with the fast-changing pace of the AI landscape, or to take advantage of new opportunities to advance our company.