Unless your life is in imminent danger, a trip to any emergency room across the country is likely to include a lengthy wait before receiving care. But pervasive ER overcrowding is more than an annoyance: It can lead to worse health outcomes, an increase in medical errors, and nursing burnout.
Professors Carri Chan and Jing Dong envision a different paradigm, one where data analysis streamlines hospital operations and transforms patient experiences. Chan is the faculty director of the Healthcare and Pharmaceutical Management Program at Columbia Business School, and Dong is an associate professor of business in the Decision, Risk, and Operations Division at CBS. For years, they have championed data-driven strategies to optimize healthcare operations — and watching COVID-19 buckle the US healthcare system only emphasized the need for better tools.
“Healthcare is an area where there’s a lot of opportunity for improvement,” Dong says of their research. By harnessing a wealth of data and operations research tools, she and Chan have found solutions with the potential to significantly impact how we move through the healthcare system. In addition to exploring how hospitals can shorten wait times for ER patients, they’re using mathematical tools to understand how to integrate telehealth appointments with in-person appointment scheduling to enable an effective hybrid care model, propose an evidence-based approach to balance ICU nurses’ workloads over time and address other challenges that compromise quality healthcare — along with considering how their research might translate to other industries.
More Than a Hunch: Guidance Grounded in Math
Chan and Dong’s research focuses largely on operational aspects of care, such as patient scheduling, nurse staffing, and hospital capacity planning — areas often overlooked in healthcare research despite their wide-reaching impact.
Traditionally, hospitals have operated based on staff experience and intuition, lacking a quantitative approach. By applying a statistical tool called the econometrics method, Chan and Dong can establish causal relationships of a hospital’s operational measures and create advanced predictive models to anticipate future issues. With this information, they can offer actionable recommendations to hospital stakeholders. “Operations research not only gives you high-level insights about what you need to do, but it could also give concrete decision support,” Dong says.
Their research has to overcome a significant hurdle, though: Performing experiments in medical settings — like assigning specific wait times to a sample of patients, regardless of their condition — would be unethical. But thanks to digitized patient records and aggregate medical databases, healthcare professionals have unprecedented access to data. So in lieu of controlled experiments, Dong and Chan can now leverage large amounts of data from electronic health records to generate more accurate and meaningful insights.
Keeping Enough Healing Hands on Deck
In the wake of the pandemic, the impact of overworked providers and overwhelmed hospitals became painfully clear. In their research, Chan and Dong use real-time predictive models to show that hospitals can proactively plan staffing to prevent understaffing or overburdening staff in the emergency room. Using their proposed algorithms to adjust staffing in real time allows hospitals to provide faster and higher quality care. Plus, improving workloads can better safeguard the welfare of healthcare providers — an important focus of the duo’s research after seeing unprecedented levels of burnout during the pandemic.
However, even using hard data, the researchers found that healthcare has a human element that’s vital to consider. Hospital administrators usually establish staff schedules weeks in advance, allowing people to plan their lives. Adapting to real-time predictive models would mean responding staff would have to be on call, relinquishing some control over their schedules.
“If you have a busy shift starting in, let’s say, the next 12 hours, then you might have the opportunity to call in some extra nurses, but you have to pay them extra, at what they call the incentive pay,” Dong says.
Hospitals will have to evaluate the tradeoff: Is an upfront investment in more resources worth the payoff? If hospital operators adopt data-driven ideas that improve the nurse and patient experience, they might find it easier to hire and retain nurses and more people might choose that well-run hospital over others — ultimately generating more revenue.
Next Up: A Better Telemedicine Experience
After watching remote care surge during the pandemic, Chan and Dong started to look at how telemedicine was reshaping healthcare delivery, revealing new challenges. Again, one of the central issues they’ve encountered is wait times. They observed that late doctors may lead to patients abandoning their telemedicine appointments entirely. Compared with office visits, “patients are more likely to leave without being seen when waiting for the physician online,” Dong says.
A deeper investigation revealed two potential factors influencing this behavior: Patients attending in-person appointments invest time and effort in commuting to the clinic, which makes them more willing to tolerate delays. In-clinic patients also observe the busy environment and gain a sense of how long they may have to wait, bolstering their patience.
In this arena, the duo have been able to begin testing online interventions through controlled experiments. “We don’t want the patient to abandon appointments, because doing so could interrupt their treatment plan,” Dong says. Measures such as delay announcements and sunk costs — like watching educational videos on the importance of getting flu shots or filling out surveys — can help alleviate this. If the results of the experiments significantly impact patient behavior, hospital operators could easily implement these strategies.
Embracing a Data-Driven Future
As these examples prove, leveraging the volume, real-time velocity, and variety of big data can improve decision-making in healthcare, enabling providers to make decisions based on concrete evidence rather than a gut feeling. “With the wealth of data that’s becoming available, we can fine-tune and improve the quality of decisions and guidelines by incorporating that information,” Chan says.
Combining these data-driven insights with other medical advances, such as diagnostic tools enhanced by artificial intelligence (AI), could revolutionize healthcare management and patient outcomes. “It would be really interesting to understand how these AI tools will affect the workflow of physicians, and how you can use the information you gain from these AI-based diagnostic tools to better allocate the resources,” Dong says.
Adapting learnings from healthcare, Dong and Chan’s research can be translated to other industries with similar issues. As in nursing, there’s currently a significant staffing shortage in the manufacturing industry, for example. “Even with higher pay and better benefits, many people aren’t interested in manufacturing jobs because schedules are inflexible,” Dong says.
Adjusting workplace culture and operations to align with these preferences can boost talent attraction and employee satisfaction, performance, and retention. “If you are planning for the future, you should take this into account,” Dong says. “It’s no longer the case that by simply offering higher compensation you can incentivize people to work the way you want them to.”
For now, there’s plenty of research still to be done in healthcare. Dong and Chan look forward to the day when evidence-based decision-making is the standard in hospital care, positively impacting patients’ outcomes, and consequently, our communities. “If you can improve the efficiency of the healthcare delivery system,” Dong says, “then that benefits the whole society.”
Watch Carri Chan, the John A. Howard Professor of Business at CBS and the Faculty Director of the School’s Healthcare and Pharmaceutical Management Program, discuss her data-driven research into how healthcare operations can be improved, particularly in emergency departments: