It is everywhere! From information technology to computer engineering, from banking to marketing, the word 'big data' seems to appear in many different sectors. It is one of the most popular words of the decade. It is so commonly used that it loses meaning most of the time, and it can be challenging to figure out what is meant by 'big data in a particular context. In this article, we looked into the term's usage in the medical field and several medical areas that can be useful.
We use big data to define any collection of 'information that is too big to process with conventional methods. An often-cited example of this is the data generated by smartphone users every day. Think for a moment how many times you interact with your smartphone every day. If you are an average user, that number should be around one hundred. Every such interaction generates a data point. Considering there are billions of smartphone users worldwide, data generated only by our smartphones every day quickly overwhelms the storage and processing capabilities of even the brand-new super-computers. Luckily computer science came up with solutions such as cloud computing in a relatively short time.
Medicine is an area that is full of big data. Every patient generates hundreds of data points each time they visit a doctor, get hospitalized, or even have a simple blood test to detect cholesterol levels. Most of the time, practitioners use this data immediately after it is collected. The doctor or other healthcare provider sees the data, determines its significance to the specific patient, and acts accordingly. Sometimes she decides to change the medication, increase the dosage or offer an alternative treatment modality. This sequence is the practice of medicine in the traditional sense. For the big data to kick in, we should process millions of pieces of information simultaneously. Of course, the human brain cannot handle this type of information, so we have to rely on computer programs and intelligent algorithms to deal with it.
Now think that your blood cholesterol tests are processed with test results of millions of other people who are undergoing some disease process or treatment. With some background information provided, data scientists – people trained to handle big data – could go through this dataset and come up with new disease associations that would be impossible to do with traditional means. For example, they could discover a specific geographical variable related to where a patient lives, which could dramatically alter the patient's response to a particular medication. Studies like that are being published even now and changing treatment ways.
By the achievement of decoding the human genome thanks to the Human Genome Project, we know where each gene resides in a chromosome and what it does. The next big challenge is to find how each gene interacts with a patient and his environment. Scientists can profile a human’s genome from a small blood or tissue sample with the current technology. Data scientists can look at this dataset and try to find meaningful information within tons of seemingly unrelated and unimportant data points. They try to make a functional model that could predict that a person might have a disease at an early age, might be more prone to side effects of a particular medication, and even has a higher risk for alcohol dependency!
Every day millions of X-rays, Computerized Tomography (CT) scans, and Magnetic Resonance Imaging (MRI) are done in hospitals. These imaging studies are assessed in detail by expert radiologists trying to understand what is wrong with the patient by looking at those images. Although they are pretty good at what they do and come up with impressive conclusions by looking at black and white photographs that make no sense to the layperson, they have human brains, after all. They cannot process millions of images at once. It was recently shown that using machine learning algorithms; computers can be taught to help doctors in assessing medical images.
All these advancements bring to mind the question of whether computers will replace doctors soon. That seems highly unlikely because the medicine is more than coming up with a correct diagnosis and prescribing the proper medication. Although computers are better than humans in processing data points, they are not capable of understanding the patient as a whole with their emotions, fears, and concerns to respond in a 'humane manner. Nonetheless, it won’t be surprising to see computers, intelligent algorithms, and even robots helping doctors in their day-to-day commutes within the next couple of years.
Dr. Genco Fas and Dr. Egehan Salepci