In This Section      

The Breathtaking Possibilities of Big Data

By Paul B. Rothman, M.D.

DNA sequencing

Others look to Johns Hopkins, a leader in academic medicine, for answers on how to address some of the greatest questions we face in health care. Chief among the concerns in 2017 is: How do we best store, process and protect patient data, and use it to guide our work?

I believe that data science and informatics stand to transform the practice of medicine, public health and biological research. In the future, scientists, engineers and information experts will help doctors customize treatment for each patient by collecting and analyzing very large databases of clinical information, factoring in a patient’s anatomy, lifestyle, medical history and genetic blueprint.

This has implications not only for clinical practice and research, but also for training. The next generation of physicians and scientists undoubtedly will need to develop competencies around understanding and applying the tools of big data and artificial intelligence.

Three things have catalyzed this movement toward “big data” approaches to medicine:

  • Advanced technologies that produce complex new data sets, such as DNA sequences, methylation analyses, protein structures and high-tech images

  • The shift to electronic medical records

  • The need to improve outcomes and efficiency in health care

So far, we in the medical field have not been able to harness the power of data to achieve that last goal as effectively as we must.

Look at the way data collection and data mining have transformed other fields, such as retail. Amazon has powerful algorithms to predict what a book a shopper is likely to enjoy based on his or her prior purchasing habits and the behaviors of other similar shoppers. For many diseases, we’re unable to customize treatment even remotely as well as Amazon can tailor book recommendations, and that’s a problem.

To offer another example, an equity analyst would not dream of advising a fund manager to invest in a stock without first running a series of highly complex models, analyzing past performance of the company along several dimensions and comparing it to other similar companies to make informed predictions. So why can’t we apply the same rigor when choosing which drug to prescribe for a given patient?

We need to bring big data to bear not only on treatment decisions, but also on screening decisions based on predictions about patient trajectories.

At the moment, there is generally very little way to predict which complications will arise in certain cases and how serious they will be. Instead, we rely on our own experiences and our memories of similar cases, and while technology is no substitute for a caring, seasoned physician, it certainly can help reinforce and, in some cases, redirect our hunches.

The Johns Hopkins inHealth team likes to use the metaphor of the air traffic controller. We can apply information science in the same way air traffic controllers use radar: to help our patients avoid traffic, turbulence and other dangers by guiding them along the very best path.

For example, our radiation oncology team members have collected 3-D images of every head and neck tumor they treat and linked those images to other trackable data around patients’ anatomy, comorbidities and outcomes. Then, they have applied what they learned from past patients in each new case.

There is no reason we can’t apply that same rigorous, comprehensive approach in every discipline, for every disease. As an institution, we just need to provide the right infrastructure, guidelines and skills to make that possible. No problem, right?!

This past fall, we held a leadership meeting to gain a clearer understanding of the opportunities and challenges surrounding big data and informatics as applied to medicine and biomedical research. We also worked to collectively develop a set of action items that we can execute on over the next two years. Time is of the essence, as change is happening quickly in medicine and informatics.

There are thousands of reasons we convince ourselves that things can’t be done — not enough time, not enough money — but if we don’t explore the possibilities, we will end up with mediocrity. Remember the old saying: A good plan implemented today is better than a perfect plan tomorrow. And when it comes to big data, it’s not hyperbole to say that the possibilities are breathtaking. They also require massive coordination, which makes it so important that we all get on the same page. We need to boldly confront the challenges ahead and continue the tradition that is our hallmark.