The airline and healthcare industries have some important features in common. Both call for large capital investments in infrastructure, rely on teams of skilled and semi-skilled workers to achieve their goals, face unpredictable levels of demand and need to maintain stringent safety standards.
However, there’s one thing airlines do that hospitals don’t – and that’s repricing services on the fly. As everyone who’s booked a plane ticket knows, the prices you scouted tonight might change substantially by tomorrow. That’s because carriers engage in a strategy called known as airline revenue management which constantly optimizes their pricing models.
To optimize their revenue, airlines continuously adjust fares using an algorithm that takes factors such as past bookings, remaining capacity, average demand for certain routes and the probability of selling additional seats in the future into account.
As the healthcare industry develops more sophisticated data analytic tools and leverages AI, it’s becoming increasingly more feasible for hospitals to do something similar. Any hospital with an EHR in place can theoretically begin the process, as patient health data alone can offer some useful operational insights.
Hospitals are already laying the groundwork for this. A couple of years ago, for example, I shared the story of a French hospital system which used machine learning to predict admission rates as much as 15 days in advance. The four hospitals involved, which analyzed 10 years of hospital admission records and other internal and external data, intend to find ways to forecast admissions by the day and even hour over time.
The next logical step, in my view, is to begin pricing hospital services in a way that relates to at least some degree to demand, staffing, clinical trends and optimized care models for patients with a given clinical path. In other words, once they can anticipate how many patients are likely to show up, they can begin pricing based on the best resources use models they can develop.
Of course, there are very few cases when the time a traveler flies has life and death significance, while that’s often the case where healthcare delivery is concerned. On the other hand, hospitals serve up a significant amount of elective services for which this approach might be a good fit. Why not give patients who already plan to pay for your services a chance to make (clinically acceptable) tweaks in their plans in exchange for lower fees?
After all, it’s not as if the way prices are being calculated is sacrosanct. In fact, from what I can see, existing hospital pricing models represent at best a very crude proxy for the level of need an individual patient might have. Starting with these somewhat arbitrary numbers, and lopping some percent off for contracted health plans, just bakes some rather fuzzy numbers into the system.
I realize that many hospitals probably don’t have the data in hand to create models that address all of the moving parts involved. Developing good models for the intensity of care needed, staff scheduling, resource utilization and countless other factors will be brutal at first. But when it comes down to it, health systems that take on value-based care will need to better optimize their ability to address shifts in demand, and I think developing on-demand pricing models could help them do so.