When Data Science met Medicine!
“The groundwork of all happiness is good health”
- Leigh Hunt
Only in great health can one achieve anything they want in life. Rightly said that ‘Health is Wealth, for without which is no happiness nor true success. In one of his papers, evolutionary biologist, Caleb Finch describes the early life expectancy in the 1800’s throughout Europe and USA was about 30 and 40 years of age. What is the scenario today? It is expected that the average life expectancy of females in the USA alone is 81.2 years and that of males is 76.4 years.
What caused this huge change over the years? It can clearly be correlated to the advancement in the field of medicine and technology that have enabled us to live longer and healthier. Data Science, a field of scientific methods and processes to extract data is playing a crucial role in the healthcare industry. With a proper medical record of a person, a doctor can easily determine the correct disease that person is going through. This article explores the ways data analytics can be applied to the healthcare and how it can reshape and improve it.
Why data science is needed?
Before diving into the application of data analytics in healthcare, we have to answer a big question. Why healthcare needs data science? What is its significance? To answer this, let’s divide the “Big” healthcare data into two parts; clinical data and patient behavior and sentiment data.
Let us see some stats here;
“According to the National Hospital Discharge Survey, there were 35.1 million discharges with an average length of 4.8 days a stay in 2010.”
“In 2011, according to the National Hospital Ambulatory Medical Care Survey, there were 125.7 million outpatient visits and 136.3 million emergency department visits.”
This sort of annual data can be really helpful to build a concrete strategy to avoid diseases or any specific health threats. Annual hospital data of a country can give us some really big clues as to what is happening all around and can provide us with important points to which we can develop a new strategy around it. That’s one of the main reasons we need the data science.
Clinical data comes from doctor notes, lab results and medical images. This sort of data is gathered pretty much every day. The hospital uses certain algorithms to analyze patient records to identify certain individuals at risk for medical conditions. Clinical data is huge and constantly grows. Without proper organization and structure of the data, the doctor can diagnose the wrong disease to the wrong patient. Clinical data is sensitive. So data science is needed to properly extract the correct information from the data.
Patient’s behavior and Sentiment data is the type of data that helps in maximizing prevention over cure. The advanced technological devices such as wearable watches that measures heart rate and breathing patterns can pretty much give much more than a mere number. They can help detect slightest irregular patterns and can predict certain disorders based on them. Let me tell you one interesting fact, a human body generates around 2 terabytes of data every single day. That’s a lot of information and these wearable techs can collect all this data and can produce some amazing results. There are around 600,000 people suffering from sudden heart stoppages in the US every year. By the proper use of data science, we can reduce this number.
Applications of Data Science
So far, we have examined why data science is needed in the healthcare field. Now we are ready to dive into the applications of data science that are reshaping the healthcare. The USA has a 30 percent obesity rate in more than 25 states which are giving rise to chronic diseases such as diabetes and hypertension.
Omada Health developed a product to target this specific problem. They call this product a “first digital therapeutic” It is a data science-driven product aimed at reducing the risk of preventable health issues. Omada uses smart devices such as scales and pedometer to process patients’ behavioral data and develops a highly customized program based on the results. It is continuing to improve itself every day.
Another application is Enlitic, an organization that uses data science to increase the accuracy and efficiency of the diagnostics.
“According to the National Academics of sciences and engineering, about 12 million adult patients are misdiagnosed each year in the US.”
Enlitic uses deep learning algorithms to produce accurate data from CT scans, X-rays etc so that it could reduce the number. Enlitic claims to produce up to 70 percent accurate results. Overall, the use of data science is essential and crucial here. There are many other applications of data science that are working towards the betterment of the healthcare.
What’s the Skill Set Needed to Progress?
So far, we have discussed why data science is needed in the medical field and what are its applications but now let us see some of the important generalized skills required to make out career as a healthcare data scientist;
1. Dimensionality reduction,
2. Supervised machine learning
3. Time series analysis
4. Natural language processing.
These are just some skills relative to machine learning that are required. These four skills are related to manipulating data to extract out the best possible results. But on a foundation level, some skills such as ensuring security, understanding the process of accessing data and understanding health graph are required. Security of the data is much more important. When thinking about performing queries or building production systems, security needs to be present at stage zero. In healthcare, mediocre data science skills and great security is better than great data science skills but weak security.
(This article was authored by Research Nest’s Technical Writer Zeeshan Mushtaq)
