Does your hospital system need a data scientist?
The day finally arrived: the leadership team at your hospital system has acknowledged that there’s a treasure trove of data – both unstructured and structured – that has either been little used or untouched despite the valuable insights, savings, and benefits it could provide. But your departments have different expectations on which problems are the highest priority, who should have access to it, how to interact with it, and even where to start. You know other organizations have been hiring data scientists, and perhaps you’re thinking that seems like the best place to start…
But is it?
The market for data science talent is tight, which means those with the necessary skills and experience can command large salaries across multiple industries. Bringing one to your team, let alone building entire data science teams, can be costly. Before doing so, it’s vital to determine if such an addition is even necessary, at what point in the process it’s necessary, and if there is another role that could meet your organization’s needs and budget better?
So before you write up that job post for a new data scientist, let’s talk about who they are, what they do—and how to determine if their skill set is actually the best fit for your needs.
Definition of a data scientist
What is a data scientist? Is it a quant in finance, a researcher in healthcare, or statistician in business?
The answer is: all of the above. One size does not fit all in data science. According to Harlan Harris, Sean Murphy, and Marck Vaisman, authors of Analyzing the Analyzers, there are at least four types of data scientists:
- Data business people
- Data creatives
- Data developers
- Data researchers
Before you begin the hiring process, you must understand the skills you need – be it the skills of a data scientist, an analyst or a data engineer – to meet your organization’s strategic needs.
Typical skill set of a data scientist
Before looking at what type of data scientist background would be appropriate for tackling your organization’s needs, let’s first look at a general data scientist’s profile. They tend to have similar skill sets and tasks, such as…
- Pulling insight from data to answer questions and solve problems
- Creating models from the data to tell a story
- Possessing strong understanding of the scientific process and ability to handle unknowns
- Typically, these professionals will have PhDs, and will prefer scripting to coding.
Typical skill set of a data engineer
Many people think they need a data scientist, but what they really need is a data engineer. Although this role is complementary to that of a data scientist, it is distinctly important and significantly different. For example, the skill sets and tasks of a data engineer typically include things like…
- Focusing on systems that store and retrieve data
- Building robust databases
- Possessing strong engineering/programming skills
- A background in software engineering
Although data engineers also understand big data tools and technologies, their focus is different than that of a data scientist – they are often more focused on building and maintaining the infrastructure that data scientists need.
Why is talking about the difference between a data scientist and data engineer important? Because quite often an organization will want both skill sets in one person. Although not impossible to find, one person is unlikely to be strong in both data science and data engineering (much like full-stack engineering); even if he or she was equally able in both areas, it’s unlikely this person would want to do both. This is a very serious distinction to make.
What if I don’t need a data scientist or a data engineer?
Do both of the descriptions above seem like overkill for what your leadership team had in mind? Many organizations want to have better and more real-time access to their data, but they already know what kind of answers they’re looking for to make decisions – they don’t need or aren’t ready to analyze for new insights.
If this is the case, you may need a data analyst – and the good news is that you may have someone in house who can be trained to fill this role. It requires far less experience in scripting, coding, statistics, and database building; and relies more heavily on an understanding of your EMR, hospital systems, and workflow.
(image courtesy of Dataconomy)
This brings up one final consideration that will greatly affect your candidate pool, regardless of the role: Do you really need a candidate who comes from a healthcare background, is familiar with a specific EMR, or specialized health-related data sets? Or could you pair that position with someone on your staff who is already familiar with your specific systems? Or is it something they could learn on the job? Just remember that if you need a data scientist who has experience with Epic and analyzing image sets from radiology departments, your candidate pool will be more limited.
Interviewing your big data expert
Now that you know if you need a data analyst, a data engineer, or a data scientist (and hopefully, which of the four kinds of data scientists), it’s time to narrow down the skills and tasks you expect from the ideal candidate and create a job description. Too many organizations try to wing it here and use the interview process to determine their needs. This won’t be effective for you, it will be frustrating for your candidates who will likely find it difficult to understand the role they’re interviewing for, and your reputation could suffer as a result.
If you have a data expert on staff already, leverage his or her knowledge during the interview process. If you don’t have someone like that on your team, you may want to consider seeking outside help from a consultant or a recruiting firm that specializes in data science.
Finalizing the hire
You found the right person. Congratulations! You’ll want to increase the odds they accept your job offer by building a competitive compensation package. While you and your HR department may have a feel for what’s fair for your current workforce, data scientists and engineers are in high demand – inside and outside of healthcare, so consider industry salary standards, not just healthcare salary standards. If you wish to bring one on, I highly suggest you consult an up-to-date resource like Greythorn’s 2016 salary survey in order to see what kind of salary range would be appropriate.
Armed with an understanding of the differences between data scientists, engineers and analysts, you’ll now be able to identify the role and talent you need to leverage your collected data for cost savings and improved patient outcomes.