Artificial intelligence (AI) could reduce the cost of healthcare by $150 billion by 2026 in the US alone, according to the Stanford Medicine 2018 Health Trends Report. To achieve this, the industry relies on access to the AI skills of data scientists. However, recruiting one full-time is not always easy and can be costly. Here Ramya Sriram, digital content manager of online platform for freelance scientists, Kolabtree, explains why no healthcare business is too small for data science.
Whether it’s medical records, electronic billing or information from wearable devices, the healthcare industry produces vast amounts of data every day. Big Data is changing the way that we conduct our businesses and data scientists are at the center of this transformation. Even if you didn’t previously invest much in technology, you will find these experts are now a valuable asset due to their ability to interpret and manage the data that you encounter.
As Big Data and AI becomes more prominent, the healthcare industry demands data that is sound, reliable and accurate. A big part of the data scientist’s remit is developing, testing and validating mathematical models and this skill can be vital for businesses across many different sectors, especially in healthcare where the chance of failure must be minimised for patient safety and confidentiality.
A data scientist can carry out data-type checks, code and cross-reference validations, range and constraint validations and structured validations on combined data types. They can verify the consistency and value of a range of data types, which will vary depending on the type of healthcare data — be it a physician’s notes or more structured datasets.
Cost and flexibility
If you decide to hire a data scientist, for example for a validation project, you will find this is an expensive task. According to Prospects, in the UK, salaries for junior data scientists start at around £25,000 to £30,000, while data scientists with experience can earn between £40,000 and £60,000. According to Glassdoor, in the US, the average salary of a data scientist is $117,345. For some businesses, this is too expensive, particularly if you are just starting up.
There are also issues of flexibility in recruiting long-term employees. If you only require a data scientist to create a database of clients for your new GP practice for instance, they will become superfluous once they have done this. Similarly, small businesses with a fluctuating need for a data scientist, who don’t need one all year round, may be discouraged from hiring one when they do.
As well as being expensive, data science is a business that’s in demand. According to a January 2019 report from Indeed, demand for data scientists has increased by 29% annually and since 2013 there’s been a 344% increase. While job postings continue to rise, searches by job seekers who are skilled in data science climbed at just 14%. You see the problem? There’s a huge gap between supply and demand.
If you have a task that’s urgent, such as designing a clinical trial, it’s important that you can access data scientists when you need them. However, if you can’t attract these experts then your project risks delay, which is not always an option. The answer may be to hire a freelance data scientist, so you can access the required skills for a set fee and short time frame. On Kolabtree, for example, data scientists charge between $35 to $200 per hour and the typical rate of a fixed fee project is $2,000.
If AI is going to reduce the cost of healthcare over the next seven years, we need to make the best use of the data we are collecting.