Macro to Micro: The road to personalised medicine
Alan Payne, CIO at Sensyne Health discusses how greater access to data is key for achieving personalised medicine.
We live in an age where personalised experiences are now both expected and commonplace. From Spotify suggesting playlists based on our music choices, Netflix giving us a recommend list of movies to watch and online retail stores pushing adverts linked to our last purchase. We take for granted our preferences to not just be taken into consideration but remembered and used to tailor the services we receive. My Amazon purchases increase is a reflection of this!
It’s the same story when it comes to the way we care for ourselves. With wearable devices such as fitness trackers continuing to grow in popularity, interrogating our heart rate and activity levels is simple, and we can use the information to inform how we eat, sleep and exercise. Beyond that, an Apple Watch can even provide its wearer with an ECG heart report at the push of a button.
The pharmaceutical industry is also taking an increasingly personalised approach to how therapies and treatments are designed, and to predict and manage what health conditions may arise amongst certain patient groups. While the industry is starting to achieve this, there is still a way to go for healthcare to be tailored to the needs of every individual. And this isn’t surprising, as the amount of data and insights needed to create truly personalised medicine and care is too vast and complex to be collected and analysed by traditional means of data processing. This, combined with the complexities of human biology, mean we still have a very poor understanding of how the human body works. This is where sophisticated technology such as machine learning, is essential.
Luckily, we are in a position where this technology is available to us. We just need to apply it in the right manner to take full advantage of the insights it can provide to potentially save lives and revolutionise healthcare as we know it.
The current state of personalisation
Achieving truly personalised medicine at scale is less than a decade away and AI technologies will play a fundamental role in this. There has been a boom in data generation in recent years and the rate of data collection is only on the rise. In fact, IDC research predicts that the global datasphere will grow from 33 zettabytes of data in 2018, to 175 zettabytes by 2025. To put that into perspective, to download 175 zettabytes of data on the average internet connection speed, it would take 1.8 billion years!
This huge dataset includes healthcare data, in the form of genetic information and electronic health records (e.g. medical history, demographics, allergies etc), which has given clinicians the opportunity to look more closely at individual patients and their conditions, in combinations they could never have done before. They are now able to leverage machine learning to spot trends, patterns and anomalies in the data that can help them make informed decisions based on multiple factors like medical history, genetic makeup and data recorded by wearables.
The application of data analytics is also important for personalising clinical trials and experiences for those enrolled on them. Many trials are still undertaken by giving the same drug or treatment to lots of different people and using a statistical approach to how the majority react. This is not a ‘personalised’ approach, as human beings are far more complex than just statistics, and every one of us has a unique genetic make-up and specific biomarkers. As a result, drug efficacy can differ from person to person – and this should be reflected in the way trials are carried out.
Getting to know us better
Each one of us has a unique variation of the human genome, so being able to understand which gene mutations or differences may cause specific illnesses, will be instrumental for clinicians to predict a health condition, prevent diseases earlier, and develop more comprehensive disease management plans to mitigate risks when they do arise.
For example, in cancer treatment, the same treatment was once routinely given to every patient with the same type and stage of cancer. However, we now understand that different people may experience different genetic changes in their cancer cells and/or their genetics will effect how their body responds to cancer; both these factors will affect how their cancer progresses. With better understanding of disease progression through analysis of patient data, precision medicine and targeted therapies can be developed and used to help predict which treatments a patient’s tumour is likely to respond to and spare them from receiving painful treatments which are unlikely to help.
In order to personalise medicine in this way, we must first build a full view of every patient, a digital clone if you will, by collating their daily data, health records and lifestyle behaviours from disparate sources. This data is crucial to understand and analyse the biology and needs of each patient, which can be used to inform both how drugs are developed, and the type of care that a patient receives. It is these huge datasets that hold vital clues to how chronic diseases manifest so pharmaceuticals and clinicians can identify patterns between lifestyles and illnesses developing.
However, the ability to do this, hinges on being able to collect, map and garner insights from vast amounts of data across different sources – a process that cannot be done manually. To put it into perspective, we currently require the equivalent of the sun’s output power for a whole week just to model a single human’s genome. Clearly, this is not a sustainable model, and will not allow us to personalise healthcare at scale.
AI and the promise of personalised medicine
This is where automation tools like AI can be of huge benefit to solve the key challenges healthcare providers face when it comes to big data – velocity, volume, variety and veracity. In fact, nearly 80% of respondents to a recent Oracle Health Sciences survey said they expect AI and machine learning to improve treatment recommendations for individuals.
With AI and machine learning capabilities, pharmaceutical companies can collect, store and analyse large data sets at a far quicker rate than by manual processes. This enables them to undertake research based on information about genetic variation from a huge wealth of patients, and develop targeted therapies faster. In addition, it facilitates learning more about how small, specific groups of patients with certain shared characteristics react to treatments, and how to precisely map the right quantities and doses to give to individuals.
As a result, patient care can be optimised. In an ideal world, we want to prevent disease. By having more information at our fingertips about why, how and in which person diseases develop, we can introduce preventative measures and treatments before a patient even starts to show systems.
The future of personalised healthcare
Personalised medicine has the potential to improve, and even save the lives of many people. Technologies such as AI will be crucial in this change and the driving force behind making future breakthroughs. By harnessing the power of AI and machine learning in healthcare we can also then begin to reap the benefits of other innovative technologies starting to emerge within the industry such as personalised, 3D-printed drugs offering the right dose for each patient.
As wearable technologies and IoT devices continue to rise in use, with an expected 1.3 billion IoT subscriptions expected by 2023, and 26.6 billion IoT devices in use in 2019, the amount of personal data we collect on ourselves will only grow – opening more opportunities for bespoke healthcare experiences for patients.
There are still many challenges that lie ahead for personalised medicine to be perfect, but as automation technology continues to advance and becomes more widely adopted across the pharma and life sciences industry, I am increasingly confident that a future of workable effective personalised healthcare is closer than we think.