Data science has been a driving force in the growth of numerous industries in recent years. Banking and finance industries have made use of data science to improve fraud detection and better analyze risk. Other industries such as retail and entertainment have seen huge success in using it to create a highly personalized user experience. However, some of the most significant benefits are being seen in the healthcare industry, where data science is being leveraged to help create revolutionary new devices and improve patient outcomes while reducing the cost of delivering care.
So what are some specific ways that data science is being used in healthcare?
- Making sense of data from wearable health devices
- Adding intelligence to existing and new medical devices and services
- Using predictive analytics to better understand disease spread
Making sense of data from wearable health devices
The market for wearable health technologies has been growing steadily for the past few years and is expected to see a compounded annual growth rate of over 15% through to 2027. Broadly speaking, these wearables fall into two primary categories; those that help users track basic health and wellness metrics, and those considered clinical-grade that help to monitor or diagnose specific medical conditions. The amount of data that can be generated by the human body is enormous – researchers at IBM Watson predict this amount predict this amount could exceed one million gigabytes per person in a lifetime. With all of this human-generated data being produced, the next challenge is how to make sense of it and extract meaningful insights.
Many physicians report being inundated with patient data from consumer devices such as activity trackers and smartwatches, but find it to be of limited value from a clinical perspective. This is likely to change in the future as a result of initiatives from companies such as Apple with innovations such as the ResearchKit open-source software that allows for large-scale studies to be undertaken by researchers using volunteers from Apple’s large wearable user base. This is already yielding real-world benefits such as an app called EpiWatch, developed by Johns Hopkins University using the Apple Watch to gather data from epileptics to better understand potential triggers for seizures.
However, many doctors are seeing value in medically accurate wearable technologies when paired with machine learning and artificial intelligence technologies. This combination has seen great results in monitoring and managing a number of health conditions including diabetes and asthma.
Adding intelligence to existing and new medical devices and services
It isn’t just wearables that benefit from machine learning and artificial intelligence. Some existing medical devices are getting a smart makeover and other new connected devices offer improved diagnostic options. These, along with wearable devices form part of a connected ecosystem known as the Internet of Medical Things (IoMT) and includes:
Diagnostic medical imaging
X-Ray, MRI, CAT, and other medical imaging devices are able to capture incredibly high resolution images, however, they still traditionally need to be analyzed by trained healthcare professionals. Now machine learning and AI powered software has been shown through multiple studies to be as effective or better at detecting abnormalities and are helping to improve patient outcomes.
Since its invention in 1816 by Rene Theophile Hyacinthe Laënnec, the stethoscope has been a standard item in most medical provider’s toolkits. California-based medtech company Eko has brought it into a new era by adding hi-quality digital audio, ECG capabilities, and AI analysis, to allow serious heart conditions to be diagnosed quickly and accurately. With heart disease being the leading cause of death in the U.S., this technology offers the ability to save many lives.
Breast cancer screening
With 685,000 deaths annually, The World Health Organization reports breast cancer as the most common cancer worldwide. Traditional screening involves self-examination and mammograms, but a new device, NIRAMAI (Non-Invasive Risk Assessment with Machine Intelligence), offers a new approach. It is a non-contact device that combines thermal imaging and machine learning algorithms to detect breast cancer earlier than other methods. Its portability allows it to be used in remote or rural areas that may not have access to typical screening options.
Enterprise remote health monitoring platforms
A new FDA-approved health monitoring platform has been developed by Current Health, that utilizes a wearable device to monitor vital signs. The collected data is analyzed by AI algorithms and can alert doctors to any decline in a patient’s condition and even allow telemedicine conferences within the platform. The ability to monitor patients with chronic illness, or recovering from surgery, remotely reduces the burden on hospitals and helps to reduce overall healthcare delivery costs.
Using predictive analytics to better understand disease spread
When the enormity of the Covid-19 pandemic became apparent in early 2020, many countries resorted to the tried and tested tool of contact tracing to try to curb the number of new infections. Some countries like South Korea had great success, while other countries like the United States didn’t fare as well in their tracing efforts.
However, a team at the New Jersey Institute of Technology built a predictive analytics model to predict the spread of Covid-19 using data published by regional health agencies. This did not require them to gather personal information, as is the case with contact tracing, and the results largely matched the outcomes reported by the states. The team aims to further improve models to be able to predict outcomes based on whether specific preventative actions have been taken such as mask-wearing or regional lock-downs.
Advancing drug research and discovery through machine learning Traditionally, developing new drugs is a lengthy (on average 12 years) and costly process – a recent study of 47 companies put the median cost at USD$985M. Creating a vaccine, for example, typically takes 10-15 years to complete the process and bring it to market. The Covid-19 pandemic has proven to be a catalyst in speeding up this process, and machine learning played a significant role in the speed of vaccine development. Scientists at the Stanford Center for Human-Centered Artificial Intelligence were able to use neural network algorithms to help identify the most likely elements of the virus that would help to create an immune response. Companies such as Berg in Massachusetts are using AI to research drug development for various cancers, diabetes, and Parkinson’s disease.
As more companies and institutions devote resources to AI and machine learning in this field, it is possible that the time and cost of developing new drugs will go down significantly.
These applications for data science in healthcare are already delivering impressive results and the role it plays is set to be even greater in the years to come. The embracing of health-focused wearables by consumers eager to better understand their own body metrics is providing an enormous pool of data that will only help to generate the health breakthroughs of tomorrow.
When data science is used to make healthcare delivery more personalized, efficient, and cost-effective, to accurately predict the spread of a new pandemic, or set new records in safe vaccine and drug development, the opportunity to improve the quality of life for all is virtually limitless.