Want to Turn Your Health Analytics Background into a Data Science Career? It’s Not Easy!

*I may be compensated when you click on my links, but you will not be charged.

Over the last few months, I interviewed over 20 people involved in careers where they trained for public health or health informatics, and have been working with healthcare data. All of them wanted to move in a more “data science” direction with their career, but they met with barriers. This article presents the barriers I found from talking to them, and the solutions available.

Barrier #1: Healthcare is Siloed, so Health Data Analysts Don’t Gain Skills for Working on Data Science Teams

I found this barrier to be particularly interesting. Look at this graphic below.

On the left, you see how teams are set up in healthcare and public health. In green, you see the positions that people trained in healthcare and public health usually take on those teams. But then, on the left side of the diagram, you will see that there is this red circle with “IT people” in it. These are data science teams, and they are usually running data or other technology systems from which the data come that the team will analyze.

But what goes on inside that circle is a “black box”, so to speak. One of the people I talked to had a PhD and would work on projects at the health insurance where she worked, but would be left out of the activities in the red circle.

As you can see on the right side of the graphic, what’s in that red circle is likely a data science team. The person I was talking to could have served in one of those green positions – but she didn’t know how to serve on such a team. If she were to be able to gain the skills to do this, she could move over to a team like this. Then, she could probably make more money and have more interesting jobs. She said she already had ”maxed out” her level in her organization. When I asked what would happen if she tried to integrate herself into the “red circle” team, she said the organization was siloed and there were too many barriers to doing this in her current position.

Why it’s So Hard to Cross Over

Usually, the people hired into those positions on the right side of the graphic have a background in business, and are taught to serve on teams like this. This is a “matrix management” approach, which we generally do not do in healthcare and public health.

So if you are in the position of the person I was talking to at the insurance, you are pretty much stuck. There is no obvious “bridge” or “transition” program anyone I talked to could find that could help a person like this gain the teamworking skills so they could start working in such a flat hierarchy like they have on data science teams.

Barrier #2: Culture Clash: Healthcare is like Healthcare, and Data Science is like Business

I made a video and blog post about the “differences” between public health and data science as fields, because I was very fascinated about what I found.

As with the hierarchical nature of healthcare, there is a lot of governance. So, there is bureaucracy, lots of protocols, and lots of slowness. But for all these flaws, this approach causes things to be relatively organized and predictable. It also ensures that protocol is followed, and that experts weigh in on decisions.

Data science is totally different, because it’s not very hierarchical. People are expected to jump in and solve problems, not wait on the sidelines for their turn to do something in a larger process, which is what we do in healthcare data analytics. In data science, everyone is expected to bring whatever knowledge or skills they have to the table to solve problems. This can lead to a very chaotic workplace with frenetic energy, which can be overwhelming to someone used to sitting in an office and going to scheduled meetings!

It's a culture clash – the way going to the Middle East from the United States or vice versa may feel like a culture clash. But people are people, and if you are actually move from one culture to another, soon you will get used to eating new food, wearing new clothes, and saying new phrases. So if you immerse yourself in a data science culture, you will soon do as data scientists do.

Overcoming the Culture Clash

But how do you do that outside your workplace? And how do you get over this culture clash before you apply for a data science job, so you can convince a new employer that you are already a data scientist, and have already adopted the culture?

This ended up being a huge barrier for the people I interviewed. I talked to someone who had worked in health informatics, and they explained that there really weren’t any data science teams at their workplace to try to infiltrate (for lack of a better word). The person I talked to at the insurance – where there were data science teams – said there were structural barriers (e.g., reporting lines) to joining one of these teams, or even just working with them temporarily.

Barrier #3: What Do you Do to Transition Without Starting Over at the Beginning? Lack of Viable Options!

If you have years of experience in healthcare data work, then you don’t want to start over when trying move to a more data science career. And you shouldn’t have to, because you already have a lot of skills needed in data science. I created this graphic to better explain what I mean.

The circle represents health data analytics as a field, and the zig-zaggy cloud represents data science (because it is sort of irregularly-shaped, in my view). The topics I colored dark red are subjects we are very good at in health data analytics, and the ones I colored blue are subjects in which people tend to excel in in data science. The purple topics are ones we do with some regularity in both fields.

In reality, we need all of these subjects to be covered in all these fields to do effective data science or health analytics. So people will want to build their careers by transitioning across this diagram. The person at the insurance was only on the left side of the diagram, and wanted to move to the right and bring their skills with them – which are sorely needed on the right side of the diagram.

I have also helped customers move from the right side – data science – into the left side – health data – but this direction is more challenging. That’s because you really need to understand the business of healthcare, and you can’t do that very easily if you haven’t been trained in it, and been doing a job in it for a while.

Boot Camps are Not Enough

All the people I talked to had either a master’s degree or a PhD, and although some had thought about going back and getting another “data science” degree, most decided that this was not a good idea. I agree with that. I don’t think a data science degree could fill in the gaps for these individuals.

Several of them had felt that if they learned the software data scientists use – such as R, Python, and SQL – they could then transition. They took boot camps, online courses (such as on Coursera and LinkedIn Learning), and used other training resources. These resources were all very good at teaching them the languages, but they all realized that this was kind of a dead-end path. You don’t really “know” a language until you do a project in it, and these individuals found it hard to quickly apply the knowledge they learned, so it would dissipate. If you aren’t doing a project in the language at school or work, how will you retain the knowledge?

Needing Guidance for Portfolio Projects

So some of them actually did try to do projects – called “portfolio projects” – to gain data science skills and retain what they had just learned. Since most of the people I talked to had had epidemiology courses sometime in their lives, I was surprised that they had a hard time doing these projects. Several expressed to me that they needed the experience of “doing a data project completely from start to finish”.

I realized that in order to do such a project if you have never done one, you need more than epidemiology and biostatistics. You need a certain type of business thinking that allows you to ask iterative questions, and then answer them with summaries and visualizations from your data. It is not obvious how to do that, and so gaining this skill is hard on your own. I have heard it called “storytelling” with data, but I’m not sure that exactly captures this skill. Even data scientists usually need a mentor for this – so not having a mentor for portfolio projects was a big challenge for these individuals.

Lack of Team Training

In high-quality data science master’s degree programs like the one at Boston University, they structure the degree so that you gain basic skills in the beginning, then serve on data science teams later in the program through co-ops and other group settings. That way, you actually practice serving on data science teams, and gain those teamwork skills.

If you have built your career in the health sector, then you don’t have these teamwork skills. And as good as it is, I would not advise a health data person with years of experience to start over from scratch with a new master’s degree from Boston University. If all you need is the team training, then there should be a program that gives you team training. However, none of the people I interviewed had found a program that did that.

Solution: “Public Health Rebrand to Data Science” Program

To fill the knowledge gaps these people were experiencing, I designed the “Public Health Rebrand to Data Science” program. The purpose of this one (1) year program is to provide people with career experience in healthcare and public health the specific data science experiences they are missing in order to rebrand as a data scientist, and “cross over” from the left side of the diagram to the right side.

So this is not a boot camp, nor is it another two-year master’s degree! It’s something completely unique, and lasts exactly one year.

This one-year program has the following features:

• Monthly one-on-one sessions with a data science mentor who works with you to develop a customized goal plan. The goal plan provides a roadmap for gaining the specific data science experience and skills you are lacking. It also gives me feedback about how to design customized data science experiences to fill your specific knowledge gaps.
• Online learning center materials exclusive to this mentoring program. These include custom courses available on a course management system only available to program participants, as well as other specific learning resources that will be designed specifically to fill the knowledge gaps of program participants.
• Simulations of working on a data science team to give you experience outside of your workplace. You will be placed in a small group called a “cohort”, and have teamwork learning and skillbuilding experiences throughout the program within your cohort. That way, you can get comfortable with the group, and establish a professional network outside of your job!
• Opportunities to do portfolio projects. These are optional and can be made part of your goal plan.
• Multiple opportunities each month for individual and group data science learning experiences that can fit a working person’s time and cost budget! The scheduled activities account for approximately five (5) hours per month. It is up to the participant to extend their participation in the program as far as they can through independent learning center activities, portfolio projects, and other non-scheduled learning activities. By providing intensive group learning rather than individual instruction, I was able to reduce the price to be much less expensive than a boot camp, software certification, or college diploma.

Does this program excite you, and make you want to join? Want to learn more about it? Then click here to sign up for a 30-minute market research Zoom meeting with me, where I will explain the program, and get your feedback.

Monika M. Wahi, MPH, CPH is a LinkedIn Learning author of data science courses, a book on how to design and build SAS data warehouses, and the co-author of many peer-reviewed publications. Follow her blog and YouTube channel for learning resources!