Recent global events have made all of us more aware of our healthcare options and needs than ever before. We’ve been researching symptoms, discussing vaccines, embracing wellness apps and virtual visits, and developing a strong desire to understand our own treatment options.
Patients everywhere are searching for accessible healthcare solutions that are truly personalized - after all, each and every one of us wants to be treated as more than just some test results or a medical chart. But is personalized healthcare actually achievable, or is it merely a pipe dream?
Thanks to patient similarity networks, personalized healthcare is more accessible than ever for the average patient. These powerful networks give us the ability to safely connect with others on a similar healthcare journey, share experiences, provide emotional and social support, and learn from each other along the way.
So how do they work, and how can they help you achieve better medical care?
Simply put, patient similarity networks are clusters of patients, carefully grouped together based on similarities they share.
In a patient similarity network, you may be grouped with patients that share not only your medical condition, but also reflect your age group, genetic background, gender, other health conditions you may have or medications you are taking, and more.
Consider a mom in her mid-40s who gets plenty of exercise chasing her kids around and does her best to put healthy dinners on the table. Although she may share a diagnosis of Crohn’s Disease with her neighbor, they are in their 70s, overweight, and are also suffering from heart disease. Patient similarity networks take all these factors into account to ensure that both individuals receive medical care that’s tailored to their specific (and very different) needs.
All theory aside, improved patient outcomes thanks to this type of personalized approach have been seen in a wide array of clinical studies. In 2015, Li et al. used electronic medical records (EMRs) and patient genetic data to identify three distinct patient clusters of Type 2 Diabetes subtypes. Interestingly, the clinical features we normally associate with Diabetes (including obesity and high plasma glucose levels) were limited to one subtype. Other subtypes were linked to cancer and neurological diseases.
In 2020, Curtis et al. used similar methods to identify five distinct phenotypes of rheumatoid arthritis, all with different symptoms, disease progression, and treatment requirements. These studies (and many more of their nature) stress that we must rethink the dogma of a one-size-fits-all, generalized approach to healthcare and instead develop personalized treatment plans for patients.
The concept of personalized medicine has been around for a while. Doctors have long-known the importance of matching donors for blood transfusions or organ transplants, but large-scale patient similarity networks have only recently entered the frame, for several reasons.
Firstly, our understanding of the human body has become much more thorough in recent years. Human genome sequencing and other new technologies offer unique details and help to identify subtle but important similarities (and differences) between patients.
Secondly, patient similarity networks require complex computational tools with advanced analytical capabilities. Many patient similarity networks are constructed using patient medical files and EMRs, which contain a vast amount of data. The complexity of the algorithms and machine learning tools used to create patient clusters have only been made available in recent years, facilitating an exciting new shift towards personalized healthcare.
As well as advancements in medical research technologies and data analytical capabilities, legislation around EMRs is also facilitating the advancement of patient similarity networks. Just this past April, new measures were introduced under the 21st Century Cures Act to empower patients with greater access to their health records. This new legislation aims to provide us with more control and transparency over our medical records, allowing us to make more informed medical decisions.
At Alike, we are excited to be at the forefront of the journey towards personalized healthcare. With de-identified data from EMRs, Alike uses a proprietary algorithm to create an Alike SimScore™, which computes the similarity between every two individuals in our system. You can explore your medical records to gain a better understanding of your own health, and connect with others in the system (your “Alikes”). Your Alikes may have the same health conditions, be taking the same medications, or undergoing the same medical procedures that you yourself are familiar with.
The immediate benefits of Alike’s patient similarity networks are easy to identify. Connecting with real people sharing your healthcare journey helps you become more informed of your treatment options. This approach can help you achieve better health outcomes by avoiding a treatment plan solely designed to treat one condition, without accounting for your overall health picture.
The long-term benefits of patient similarity networks are equally important. Doctors can use data from patients similar to us to predict future disease onset or progression, which helps us to take the necessary steps to stay as healthy as possible and prepare for the journey ahead.
By embracing these exciting new technologies, we hold our healthcare destiny in our own hands in a way that hasn’t been possible until now. Patient similarity networks give us a real chance to change our health outcomes by providing a medically accurate glimpse into our own futures.
.png)
.png)
