Why your EHR is just not enough

Health system leaders are increasingly looking to deploy sophisticated techniques to improve operational efficiency and mine the enormous amount of clinical, operational, and financial data that already exists within their electronic health records (EHRs). This begs the natural question—if the data mostly exists in the EHR, why can’t the EHR do these things for us?

Many EHR vendors will assert that their platform can perform these functions, or if not will soon be doing so in an upcoming release. The senior leadership in the health system would like to believe that the EHR they selected can deliver these capabilities. After all, they invested tens or hundreds of millions of dollars and a countless number of FTE-hours deploying the EHR.

The EHR is excellent at doing what it was designed and chosen to do—harnessing all of the clinical, financial, operational, and patient-specific data in one place to create a single source of truth for the entire health system. It provides a unified view of all patient encounters across time and across all points of care, showing clinicians and staff a current view of the patient’s unique characteristics; however, the EHR cannot perform tasks it wasn’t built to perform.

There are three fundamental reasons why your EHR isn’t enough:

1. EHRs have a flawed concept of an appointment.

Every EHR treats an appointment as if it were a “reservation of the specific resource for the expected duration of the appointment.” Hence, when a two-hour infusion appointment is being scheduled, the EHR enables the scheduler to “reserve” an infusion chair from 9:00am to 11:00am on a specific day. This is an excellent way to schedule tennis courts or conference rooms because the exact start and end time of the reservation is known at the time of making the appointment. Unfortunatley, this is inadequate for scheduling a clinical encounter which will likely not start or end on time and, if it does run long, cannot be terminated midstream. This is why the scheduling grids displayed on the EHR look wonderful when the day begins and fall apart as soon as appointments start to deviate from the planned schedule.

2. EHRs encourage a first-come-first-scheduled mindset.

Every EHR displays availability on each day into the future and encourages schedulers to “offer the patient their choice of available slots” believing that this is a patient-centric approach to scheduling. This is as helpful as attempting to solve a jigsaw puzzle by putting pieces on the table in the order they were handed to you. It also confuses the concept of scheduling with the concept of optimization. Scheduling is the act of putting an appointment on the calendar, whether on paper or online. Optimization is the method for determining the correct slot for that specific type of appointment—this must be done before the appointment can be placed on the calendar with whatever scheduling mechanism is being used.

3. EHRs do not have capacities for probability theory or discrete event simulation.

Complex scheduling environments rely extensively on advanced mathematical methods like probability theory and discrete event simulation. This is why airlines routinely overbook flights knowing that there will be cancelations and no-shows which will be offset by the extra standby and upgrade passengers. No EHR function accounts for these realities. You cannot book three appointments into two slots nor can you book a three-hour appointment into a two-hour slot. Instead, the EHR gives you a blunt “overbooking” instrument and leaves it to frontline staff to estimate the number of overbooked appointments and the placement of each one in a manner that would be minimally disruptive.

Understandably, health systems have challenged their EHR partners to help them solve some of these problems. In response, most EHRs have built robust dashboard and reporting capabilities and facilitated the smooth transfer of data into analytic tools like Tableau or Power BI. Dashboards are nice but they do not solve the problem. Some health systems go a step further and provide future-looking predictions and alerts. Those can be helpful but only if they predict something unique or alert you to something unexpected.

Enabling the frontline to make rapid decisions based on prescriptive recommendations requires building predictive analytics, probability theory, constraint-based optimization, discrete event simulation, machine learning and AI algorithms at scale in order to fully leverage the data that exists in the EHR.

These capabilities simply do not exist in your EHR, which is why your EHR isn’t enough.