Digital technologies are rapidly transforming the healthcare industry, offering novel approaches for improving healthcare results, potential for reducing long-term healthcare expenses, and fueling the growth of the digital health market.

In the video linked above, Carri W. Chan, the John A. Howard Professor of Business at CBS and the Faculty Director of the School’s Healthcare and Pharmaceutical Management Program, discusses her data-driven research into how healthcare operations can be improved, particularly in emergency departments.

In addition to her research in this field, Professor Chan teaches a number of related courses at CBS, including the US Healthcare Course, a consulting practicum called Healthcare Management, Design and Strategy, the Analytics Advantage, as well as the School’s core operations management course.

Watch the video above and read the transcript below:

CBS: Tell us about your research.

Carri Chan: I'm actually an engineer by training. I got my PhD in electrical engineering. I teach here at Columbia Business School about healthcare, and my research is in the space of healthcare operations. My primary collaborators are actually at health systems, many of them are doctors. And so my work takes a very interdisciplinary lens that fuses business, healthcare, medicine, engineering together to come up with new solutions to improve healthcare industry, to improve access to care, to improve quality of care, and to reduce costs. So one of the things that we have done is we have built a machine learning predictive algorithm to predict when patients are going to arrive to emergency departments. But critically what we do is we look at this prediction on two different timescales aligned with when staffing decisions are made. Even before the COVID-19 pandemic, almost all hospitals have some sort of surge plan. A surge plan is what do you do when patient demand exceeds supply of resources? These surge plans can have many different features, including bringing in overtime staff or agency staff, which are temporary staff, to help fill in gaps to meet the patient demand. So what we wanted to do was understand how predictive analytics could be used to guide these types of staffing decisions. With that in mind, the decisions are typically made months in advance of how many full-time employees you would have staffed. But as the shift gets closer, you may see that demand is actually higher or lower than what you expected it to be. Why would this be the case? Well, think about it. Every winter we know that the flu is going to hit. We don't know if the flu season is going to be bad six months in advance, but when you're in the middle of the flu season, you actually have that type of knowledge. And so real time information that can feed into your predictions can certainly then be used to improve your staffing rules. So what we do is we build a staffing algorithm that has these two decision epochs weeks in advance to determine your full-time employees. And then we feed in our real-time predictions to determine how much additional staff surge up. We estimate that our algorithm can save anywhere from 10 to 15 percent of staffing costs while maintaining a level of quality and access for patients just by using data and smarter staffing decisions.

CBS: How can data help improve healthcare?

Chan: In many parts of healthcare, but at other industries as well, a lot of decisions can be made by gut feel and often because people have experience that gut feeling can be a very good choice. But it's also important to understand that with access to data, with the wealth of data that's becoming available, we can fine tune and improve quality of decisions and guidelines by incorporating that information. On top of that, the importance of data itself is not just about having the amount of data, but the type of data. And so we found in our staffing model for emergency departments that actually the predictions of emergency department arrivals between predicting, let's say two months in advance versus right now at the beginning of the shift, you would imagine that there would be an improvement in the quality of the prediction. And there is, but it's actually quite small. But because that information is so critical to the staffing decision, it can translate into vast improvements in terms of your outcome and the quality of the decision you're making. And so it's not just about gathering data from everywhere. It's about finding what is the right data that'll help you guide the decision in the most effective way.

CBS: How does this research impact CBS students?

Chan: About 15 to 20 percent of CBS students are interested in healthcare, and a growing area of interest is in the digital health space. And so there are questions amongst our students to understand where are the opportunities? What are the barriers? And so, I live and breathe this in my own research, and so talking them through it to show them there are real opportunities where AI and machine learning can have a direct impact in improving healthcare is really exciting. I also have the opportunity to describe to them what are some of the barriers, how to get buy-in from clinicians, how to get access to the right data. And it's not just about getting more data, it's really about identifying what type of data is necessary to be able to utilize these tools in an effective manner.

CBS: How do you feel about the future of healthcare?

Chan: It's a really exciting time to be working in the healthcare space. The amount of innovation that has happened in the last couple of years, and the openness of the community to adopt the tools that are becoming available and are being used in many industries is growing rapidly. Admittedly, healthcare has been relatively slow to adopt the use of data analytics, predictive analytics, and these types of models, in part because data availability and the quality of data has been fairly limited. On top of that, the community and the industry is relatively conservative. These are literally life and death decisions. And so one wants to make sure that a machine learning algorithm, or some AI tool that is being used, has been really thoroughly vetted. That said, there's been a dramatic change over the last couple of years in terms of the availability of data. And with continuing pressures to reduce cost and increase access to care there's a real opportunity for these types of tools to just transform the way that healthcare is done in the US and across the world.