A few years after graduating college, I found myself working in the healthcare field. Although I had realized that working in medicine did not fit my career goals, I became fascinated with the applications of data science within healthcare. At the pain clinic where I worked, we often encountered patients in severe pain with a risk of developing opioid addiction; the intake process was designed to identify risk factors for addiction. If we learned the patient was at high risk for addiction, the physicians decided it was better to try steroid injections, a common yet effective tool often used as a preventative measure to avoid opioid pain medications. I witnessed firsthand the application of analytics utilized within medicine to determine the warning signs of disease to achieve a simple and inexpensive solution. Experiences like these kindled my interest in employing technological tools to alleviate problems that affect a large population of people, which eventually led me to the Institute for Advanced Analytics (IAA) to earn my master’s degree.
Data science and machine learning concepts being taught in this program are applicable to healthcare due to the volume of data that is collected. A 2020 report from Deloitte estimates that the increased efficiency gained through data science can save up to 400,000 lives in Europe alone.
Using data science in patient care can lead to faster and more accurate diagnoses along with simplified operations and decreased costs. Based on my time in healthcare, there are three applications of data science within medicine that I am most excited about.
- Medical imaging
The subspecialty of radiology utilizes various types of medical imaging, such as X-Rays, CT scans, and MRIs to diagnose ailments including bone irregularities, cancer, and nerve pain. Data science can be used to detect patterns within these images and help make a diagnosis. This is especially helpful if there are difficult-to-detect bone fractures, displacements, or growths that physicians may potentially overlook.
- Drug discovery
Creating new drugs is a long, challenging, and expensive process for pharmaceutical companies. In fact, according to the Congressional Budget Office, the average cost of bringing a drug to the market is $1.3 billion dollars. Furthermore, the probability that a drug will be approved for clinical development is 12%. To increase the likelihood a developed drug will be successful, researchers can derive insights between patient data and data available on chemical compounds to maximize success. By utilizing machine learning algorithms, researchers can develop models that compute the prediction given these variables to create effective drugs.
- Predictive analytics
Our devices collect a plethora of information on us. Whether it be what we search, say, or the places we visit, companies have been using insights from this data to reach actionable conclusions for their business models. Recently, smartwatches have become equipped with features to detect low heart rates, irregular heart rhythms, and low blood oxygen. Patterns from data collected from devices such as an Apple Watch are used to create predictive models by measuring the correlation of these variables and disease states.
Although this blog post only highlights a limited view of exciting data science applications within medicine, the opportunities are endless. The skills gained from the IAA will help me and my other fellow students build the next generation of leaders within healthcare. We are capable of combining communication, data, and healthcare sciences to unlock the value of data science in healthcare.
Columnist: Mahad Munawar