Meet the AI That’s Making Interpreting Your Blood Test Results Look Easy!
The preventative healthcare company Thriva has been working on Doc-chain, an internal personalised health platform that will revolutionise how doctor’s advice is given and blood results are interpreted. Doc-chain will take everything it knows about you and produce a health report specific to your goals and test results. Excitingly, this means that we can deliver the most relevant information to you at the different stages of your health journey in a highly personalised way.
What is Thriva?
Thriva is a home finger-prick blood testing service that helps you keep on top of your health. You can easily monitor your health with a straightforward test kit that’s posted to you.
After every test, you get a detailed breakdown of your results, advice and recommendations from a GP on how to improve your health. Our aim is to to give you the most personal experience, and as much relevant and accurate information as possible. Which is why we’ve recently developed “Doc-chain”.
How Doc-chain started
There’s so much information available from scientific research journals and large national health studies carried out by the likes of WHO and NICE. However, the downside is that none of this information is available in a useable format and in one place. So, how can we use the massive repository of health data that’s available to help people understand and improve their health? Doc-chain is our answer to this.
Doc-chain originally started as a data entry tool to help us improve the advice we give our customers. But, it quickly evolved into something much more comprehensive. It’s now a powerhouse of structured information, and through it we can see the power of the computer and human to deliver valuable insight.
The Doc-chain glossary
- An analyte is something in the human body that we can test and measure. For example, a lab test can check the blood for levels of HDL cholesterol, or saliva for the hormone cortisol, which we refer to as analytes.
- A phenotype is anything that you tell us about yourself when you fill in a short questionnaire in your health profile when you sign up. For example, a phenotype can be a family history of cardiovascular disease or following a vegetarian diet.
- A condition is something abnormal that affects your health. For example, atherosclerosis, which is the narrowing of the arteries.
- A disease is a pathological process in response to internal or external factors. For example, cardiovascular disease which can be caused by abnormalities in the body such as atherosclerosis.
- A symptom is a feature of a disease, which happens as a result of having a disease. For example, fatigue or shortness of breath, which can be symptoms of cardiovascular disease.
What exactly is Doc-chain?
Doc-chain is our contextual encyclopedia. Similar to the thinking behind Blockchain, Doc-chain provides a “digital ledger” of evidence-based links. Unlike transactional links in cryptocurrency, a link in Doc-chain ties:
- A disease to a symptom or condition — for example, cardiovascular disease is linked to chest pain.
- A blood result to a health area — for example, a high white blood cell count is linked to decreased longevity (long-term health).
- And an intervention to a specific blood analyte — for example, moderate physical activity can help reduce levels of harmful LDL-cholesterol.
To do this, we’ve pulled together a huge body of information. To understand the above relationships and their paths, we use a relational database to input the data and a graph database to visualise all of the connections. Understanding what the relationships are is the most important aspect of Doc-chain, in order to tie everything we know about our users together — the end goal is to give you the best and most relevant information.
“Doc-chain allows us to deliver this information in a personalised way to our customers at the different stages of their health journey.”
The data from Doc-chain is now automatically imported into a graph database — which allows us to visualise the relationships between blood results, diseases, and so on. How do we capture all of these relationships? We do it using links.
Links capture all the possible relationships between analytes, diseases, phenotypes, and interventions. Every link is backed up by scientific evidence — which we input when making a link. Each link must be verified and approved before it’s added to the record. The strength of the evidence is also graded.
Doc-chain sources and targets
Every link in Doc-chain has a:
- source — an analyte, intervention or condition.
- and a target — similarly, a target is another analyte, phenotype, health area or condition.
By inputting many links, we can build up a picture of cause and effect, and map these out to look at any analyte we test for. Doc-chain has meant we can systemise and store these maps in a way that is usable to our customers — so we can deliver this information at the right time and in a personalised way.
Health insights and interventions
At Thriva, you can tell us your main goals, which are divided into eight health areas:
- Longevity (long-term health)
- Mental wellbeing
- Nutrition and weight loss
- Women’s health
- Men’s health
When we share information and advice with you about your results, we want to make sure it ties into the health areas you care about. In Doc-chain there are two ways information about your results is surfaced to you.
Health insights show you what your results mean to you. Your test results can help identify why you might be feeling the way you do. For example, if you’ve told us you’re interested in sleep and your results show a low vitamin D level, we can flag that low vitamin D levels are associated with sleep problems.
Interventions show you what you can do to improve your results. They’re the recommendations we give you if your results aren’t in the normal range. For example, if your lipid results are high, there’s dietary advice, exercise advice, smoking advice, alcohol intake advice, and so on that you can implement.
One of the most powerful aspects of Doc-chain is that one intervention can affect multiple analytes. So, for example, by exercising more regularly, you can increase your HDL cholesterol and also lower your triglycerides. This has a knock-on effect for your risk of heart disease and will also increase your longevity.
“Better access to reliable, organised health data means better health advice.”
It’s reasonable to say that no human is able to consider all of this information all at once and in different contexts. The built-in search engine interpretation power of Doc-chain can uncover patterns emerging that humans aren’t aware of and surface the most relevant data. If we do it right, we’ll be able to provide much more specific information than a doctor could by looking at a report on its own, in isolation. Doc-chain has the potential to be a supportive tool for doctors, to enable them to comment on reports with the best information possible to take into account for the patient.
Why we’re excited about it
Doc-chain allows you to zoom in on what you want to look at. The advice that we give is tailored to you using ALL of the information that we know about you. The wider implications for clinical practice are very exciting, as it will save doctors time and help them to understand their patients with more evidence and clarity.
We’re empowering doctors and users with the best information, to allow them to take control of their health journey. Our advice is concise, fully automated, and scalable. Importantly, we can show the user the effects of an intervention on their results, and show them the reasons why we’re telling them this — in a completely evidence-based way. Doc-chain means Thriva can change with our users, as they change along their journey and we can be there at each stage, offering the most valuable advice for where they currently are.