You may not realize it, but nearly everyday you benefit from data science. It makes texting faster with autocomplete and ensures our packages are delivered on time with optimized shipping routes. Data science simplifies the unimaginably complex and predicts the future to help us make better decisions. Shouldn’t we want healthcare to leverage data science too?
In healthcare, the difference between the right and wrong decision can result in hundreds of thousands of dollars in spend, hospitalizations, or worse. It’s time we made healthcare smarter. Luckily, the healthcare industry is ripe for innovation through data science now, more than ever. Why?
There’s so much data to work with
First, because there’s so much data to work with. It’s estimated that 30% of all data in the world today exists within healthcare, which presents a massive opportunity to power insights and predictions. Every day, the healthcare system creates new mountains of data as every interaction is logged in patient medical records, imaging & test results, clinical trial results, and everything in between. As of 2017, nearly 90% of all practitioners in the US had adopted electronic health records, adding to the daily deluge of data.
Nearly every stakeholder in healthcare can benefit from smarter decisions unlocked by data science
Second, healthcare is a perfect environment for data science to break new ground because it is so inefficient and complex. Nearly every stakeholder in healthcare, including doctors, pharmaceutical & biotechnology companies, and patients, can benefit from smarter decisions and insights unlocked by data science. Doctors can leverage data science technology to improve accuracy and speed of diagnoses through smarter analyses of imaging and electrocardiogram results. Pharmaceutical & biotechnology companies can reduce time and costs needed to bring drugs to market by mining prior chemical, clinical trial, and participant data. And patients can leverage data science-enabled tools that nudge them to make healthy decisions around diet, exercise, preventive care, and others. Altogether, McKinsey estimates that applications of data science in healthcare could reduce overall spend by $300 to $450 billion annually.
But, there are significant challenges for data science in healthcare
Data, while voluminous, is siloed. For example, hospitals use different electronic health record systems that don’t talk to each other. This means that analyses on a single patient using multiple hospitals will yield different, inconsistent, or contradictory results because each system is using a different data set. Additionally, data is often unstructured (e.g. free text), meaning it requires further processing before it can be used, and inconsistent, as systems may have different definitions for the same concept. Finally, and arguably the biggest barrier, is a shortage of talent. Healthcare companies have struggled to hire the data science talent needed. One study estimated that there was a shortage of 250,000 data scientists in 2020.
Additionally, not all data science techniques or applications are created equal. For example, data science is frequently used to simply describe what happened in the past. This is helpful to understand what happened, but insufficient to help inform future actions. To properly aid in decision making, data science needs to predict what will happen in the future based on past data and the current situation, such as predicting how well a physician can treat a patient, given that physician’s historical performance and that patient’s current medical needs.
Despite these barriers, some applications of data science have emerged as particularly promising for healthcare, such as matching members to the right, high-quality provider for their specific needs. We’ll explore those applications more in our next blog post, where our data science team will talk about applications that they think will have the biggest impact in the future.
Mark your calendars! Register for our June 10th webinar: How Data Science Makes Healthcare & Health Benefits Smarter.