5 Excuses that Needlessly Kill Data Science in Healthcare

5 Excuses that Needlessly Kill Data Science in Healthcare

by Ryan Heyborne, MD, MBA, Proskriptive Chief Medical Officer

Data science is revolutionizing how we live. Online shopping, financial systems, travel, education, and myriad other institutions are being transformed. Never have we had more opportunities for insight into the world around us or what it may become. Nowhere is effective data science more vital than in the field of healthcare. Conversely, nowhere are the obstacles more challenging.


The stakes are high in healthcare.  It has been estimated by the Institute of Medicine that around a third of healthcare spending is wasted, meaning it does not improve health. This translates to close to a trillion dollars a year. The same entity has estimated that nearly a hundred thousand deaths a year occur due to medical error in hospitals. While these figures are certainly up for debate, it is clear that healthcare is more expensive than it needs to be, and outcomes could be improved. Combine this with a rising rate of burnout among physicians (around 50% – higher in some specialties, lower in others), and the need for systematic change becomes even more apparent.

There is no magic pill and no easy fix.  Simply telling providers and healthcare systems to “do a better job” isn’t going to accomplish anything. All of us in healthcare (with very few exceptions) truly want to do the right thing. We want to do it for the lowest reasonable cost. A major obstacle for providers and systems, however, is access to accurate and actionable data. Effective health data science can unlock opportunities that lead to higher-quality, lower-cost care.


In talking with others and examining my own perceptions, I have identified five obstacles that can prevent us from even trying to access the power of healthcare data (with the contrasting perspective in italics):

1.  Healthcare is too specialized – Medicine is technical. It has its own language.  While medicine is specialized, so are other disciplines. Data science principles can be tailored to any field. Health data science experts can help navigate the path.

2. The data is too complex. Healthcare data is fragmented. Formats vary. Systems don’t play nice with one another.  The data is complex, but that doesn’t mean it can’t be analyzed. Unmonitored collection and storage methods may not work, but data management techniques can be implemented to make healthcare data consumable.

3  Privacy issues are insurmountable. HIPPA won’t let me. A data leak would be devastating.  Security protocols are ever strengthening. While there are always risks when dealing with sensitive information, those risks can be mitigated. The potential benefits outweigh the potential harms.

4. Providers won’t play. Doctors don’t have time to talk about data. They won’t believe it if they do.  Doctors are busy but they are busy trying to take care of their patients. They don’t have time for futile exercises, but they yearn for actionable data. Physicians are scientists and will change practice when the data supports it. The data must be validated, clearly presented, and tied to real outcomes. When it is, doctors will listen and act.  

5.  Data science is a buzzword, we have real problems to solve first – Patients are waiting. Audits are looming. Budgets are straining.  Data science is a key component to solving the problems facing healthcare. The issues are too complex to attempt any other course.


This, of course, isn’t a comprehensive list and I am sure my views will continue to evolve.  I do know that these are real challenges.  It is hard to gain momentum when going up against such daunting obstacles. Even when we commit to utilizing data science in healthcare, the path is a hard one to travel. It is crucial, however, that we recognize these obstacles for what they are, and find creative ways to break down barriers that prevent us from delivering effective and efficient care. Some principles that have helped me better leverage healthcare data include:

  • Commitment – Yes, health data science is difficult. We can’t let the challenge deter us from jumping into the data, asking the hard questions, and iterating until we see improvement (even if it isn’t readily apparent). The stakes are too high to shy away.
  • Humility – Healthcare is unique. It is specialized. That doesn’t mean we can’t learn from other industries. It doesn’t mean we can’t ask for help and let validated tools guide our improvement efforts.
  • Flexibility – There is no such thing as perfect data, and we need to understand the limitations inherent in any model.  Tough questions need to be asked, and we have to be willing to change what we do based on the answers we find. Improvement depends on it.

Again, I am not trying to pretend that I have this all figured out.  I can say, however, that I want to make the system better and effective data science is key in getting there. Let’s keep the conversation going. Let’s work to find solutions, and not let the challenges prevent us from moving forward.

If you see other challenges or opportunities for data science in healthcare, I welcome your replies to this post.


Got Something To Say:

Your email address will not be published. Required fields are marked *