Four lessons I’ve learned as a woman in technology

Even when I was little, I wanted to know how things worked. I loved to experiment with “mixtures” in the kitchen while my parents cooked, keeping track of my results, searching for some perfect consistency of goo.

As I grew from a “mixture” scientist into a data scientist, I remained fascinated with finding the explanation for how things work — the process behind the pattern.

I spent the first 12 years of my career in academia, working on my PhD in statistical ecology, researching, and teaching as a postdoctoral fellow and lecturer. My research in applied statistics and mathematics focused on understanding patterns in time series data. I created underlying models of the mechanisms behind changes in the data, then compared predictions from the models to the real world. I used these kinds of methods to predict species extinctions and pest outbreaks in agriculture, explain chaos in time series data and patterns in human evolution, and estimate the effects on climate change on insects.

Although I enjoyed my research, I wanted to be closer to the impacts my work could have on people’s everyday lives. I also wanted to be part of a larger organization where my contributions could have greater reach. The technical background and expertise I gained in mathematics and statistics overlapped with those of a data scientist, so I gravitated towards technology. I started to apply my expertise and knowledge of ways to understand the meaning in data to real-life applications in healthcare.

In healthcare technology, I use data to understand and improve the way doctors and nurses use software applications, helping them get information more efficiently and improving the way they understand a patient’s journey. I work on innovation projects where we start with a customer problem and experiment with all of the possible ways we could solve that problem. The work involves a lot of collaboration and out-of-the-box thinking. In this role, I work at the intersection of data science, engineering, and product to design and operationalize data science solutions. I also work on improving our data science best practices and processes so our data scientists have a better understanding of what problem we are solving and how that solution will provide value for our customers.

In this role, I rely heavily on my non-technical skills: communicating with a cross-functional team, writing proposals or documentation, and giving presentations to explain what we are doing and why it is important. I developed a lot of these skills in academia writing grants and teaching. Explaining complex data science to stakeholders is easier after teaching Intro to Calculus at 9am to university freshmen.

1. Trust your voice.

A common theme I hear when speaking with women is the inner critic who tells them not to rock the boat or speak too loudly. Inner critics are especially strong for those of us who grew up in the minority in a situation. When I was the only girl in my university math class, I never raised my hand. Making a career transition to tech motivated my inner critic to appear again. But I have learned to trust my instinct and voice to set the direction and lead projects.

I realized a few things along the way that helped me silence my inner critic and trust my voice. One realization was that I could be more confident if I prepared thoroughly and backed up my opinions with evidence. Another was realizing that no one knows everything — we are all still learning. It’s a real strength to be able to ask the right questions that make people think more deeply about a problem and move everyone closer to a solution. I also found allies who explicitly asked me for my opinion, and that made me more confident that my point of view was valuable.

Now, if I could give advice to my younger self, I would tell 15-year-old Katie to stop caring what other people thought — to stop waiting for approval or permission to do what she wanted to do. To trust her gut and set her own course through life.

2. Lift each other up.

Trusting your own voice is easier when you have a chorus of supporters behind you. It’s critical to have networks of allies who support you as you navigate through your career. Women — and people who identify as a minority gender — have to work harder to succeed, both at their individual jobs and at the general "glue work" necessary in an organization. This is compounded for people of color and people affected by other disparities. Finding an organization where that work is recognized and rewarded is important, and building communities of supporters who can lift you up and encourage you is invaluable.

I’ve found different types of communities that help in different ways. Large employee resource groups keep me connected and remind me I’m not alone, while a few trusted individuals are always there to listen. Mentors in different parts of the business give me new perspectives, and small feminist backchannels reinforce my beliefs when I start to doubt myself. I’ve even started to build my own personal “board of directors” made up of friends, family, co-workers and past colleagues.

To build your network, seek out people who understand the challenges you face and can offer help and advice when you need it — allies you can count on in professional situations. Similarly, reach out to people you feel you can help or mentor. You need to invest time and care in these relationships in order to build a community of people dedicated to helping each other succeed.

3. Innovation cannot exist without inclusion.

Innovation is at the core of our strategy at Elsevier Health. Our goal is to create products that help clinicians learn and care for their patients, ultimately improving patient outcomes. Our teams think outside of the box, finding new ways to improve our products — enabling clinicians to find relevant information more quickly at the point of care, for example. Our search teams use cutting-edge technologies pulled from the latest research on search, knowledge graphs, and natural language processing to interpret user queries and deliver the right information for a clinician treating a patient. This kind of innovation lives in a space of unknowns where new ideas can arise.

If innovation requires new ideas, then fostering innovation requires creating a space where people feel comfortable using their voices and sharing those ideas. The most innovative teams I have worked on have been diverse groups of smart, passionate people with a healthy amount of disagreement and huge capacities for listening and understanding. These spaces always have strong psychological safety, respect and trust.

Communities that innovate hold diversity, equity and inclusion as core tenets. This promotes psychological safety and collaboration while encouraging diversity of thought and perspectives. If innovation needs a new perspective on an old problem, then we need to make sure people with those new perspectives are in the room — and that we listen to them and embrace the new ideas that arise.

People don't come to work as a slice of who they are — they come as their full selves whose experiences and perspectives are informed by gender, race and ethnicity, among a million other things. That diversity of experience and perspective is a strength, both for how our diverse employee community improves our organization and for our ability to innovate.

4. First, do no harm.

Data science in healthcare requires an ethics-first mindset to manage every step of a project. Data sets often contain sensitive patient information that is protected by laws such as HIPAA(新しいタブ／ウィンドウで開く) and requires higher security and auditing capabilities. To decrease risks of working with sensitive data, the data is often deidentified by removing details that could be used to identify the patient, such as name or birthdate. Removing those details can obscure bias in a data set. Data sets can be biased towards people of a certain gender, race, ethnicity or any other group or demographic. Groups can be underrepresented, which poses problems in extrapolating the results or predictions to people not represented in the original data set.

The patient population of any healthcare system is constantly changing as people move between neighborhoods and cities or as hospital systems acquire new clinics. Medical code standards evolve, and best practices change. These drifts in the patient data cause any model to be less accurate over time. In approaching these problems, we need to adhere to the ethical principle of not causing harm and keeping the patient’s safety paramount. We need careful monitoring of models to ensure that they are really doing what we say they do. We also need thoughtful interpretation of model predictions and the factors that contribute to those predictions. Applying data science in health requires careful thought and consideration and a true ethical commitment to the patients we are ultimately trying to help.

My future journey

As I look ahead to the next steps in my career, I see endless potential within healthcare data science as a field and at Elsevier Health. We are just beginning to be able to make sense of the mountain of data we already have, to be able to understand from data how patient outcomes can be improved. With new data pouring in every day from clinical trials, wearables and digital medical devices, our ability to turn patient data into meaningful insights for clinicians relies on developing innovative data science methods and applying data science with an ethics-first mindset.

I’m excited to see where the next 20 years takes us as a field and as a company, and I’m thrilled to be in one of the drivers’ seats of this exciting field.