Saving children's lives with data and predictive analytics

Patient no-shows are a big deal for health care providers. For hospitals already grappling with tightening margins, heightened demand for patient care and persistent staffing pressures, every missed appointment has a ripple effect felt throughout the organization. The patient must be rescheduled, as the open appointment goes unused while other patients wait to be seen. Caregivers' time is wasted, and revenue is lost because US insurers do not pay providers for unattended appointments. Just as important, the patient who didn’t make the appointment misses an opportunity to get the care they need.

At Children’s Specialized Hospital (CSH), part of the RWJ Barnabas Health system in New Jersey, hospital leaders had been grappling with patient no-shows for years. “We tried everything from reminders to overbooking,” says Colin O’Reilly, Vice President and Chief Medical Officer at CSH. However, these tools still didn’t have the desired impact – no-shows continued to take a toll on patient care, staff morale, and revenue. For some specialty services, the hospital would book appointments six to nine months in advance – a last-minute cancellation was especially regrettable in these scenarios, when many other patients were enduring long waits to receive the care they needed.

SAS partner innovation: predicting patient no-shows

Like every health care provider, CSH has access to significant amounts of data connected to individual patient no-shows. How far must the patient travel to access care? Do they have a history of no-shows? Are there social determinants of health that may influence the likelihood of a no-show? Which types of communications (reminder calls, texts, etc.) do they prefer? Are current weather and traffic patterns presenting obstacles to patients trying to reach their appointments?

Although all this data existed, it was scattered across various systems. The challenge was figuring out how to connect the data points and turn them into something truly useful.

That’s when CSH discovered the Predictive Health Solutions Patient No-Show Predictor, developed by SAS partner Pinnacle Solutions. Built on a SAS Viya foundation and hosted on AWS, the predictive analytics solution can draw data from a wide range of the hospital’s proprietary and commercial solutions, as well as third-party data sources, to construct patient-level no-show risk profiles and suggest specific actions.

At CSH, that required integrating data from the hospital’s Epic systems with scheduling and appointment history data from other systems to identify which upcoming visits presented the greatest risks of no-shows. The No-Show Predictor offers intuitive user interfaces that enable schedulers and other staff to easily identify risks and drill down into the underlying data as needed – all while ensuring HIPAA compliance.

Features that point the way to action

“The Patient No-Show Predictor has been very easy for our staff to use,” says O’Reilly. “It features interactive dashboards that enable our staff to identify the most likely no-shows and then customize intervention strategies – whether it’s a weekly text, a single call or a virtual appointment when that makes sense.”

Some of the capabilities CSH staffers rely on every day include:

No-show scorecards – The No-Show Predictor features a daily scorecard where each appointment is scored based on the no-show probability, allowing schedulers to quickly assess which patients are most and least likely to miss their appointments. This resource is particularly important for same-day scheduling, because it points staff members to the appointment slots that are most likely to be available for overbooking.

Individualized reminders – With the No-Show Predictor, reminders can be tailored to individual patients according to their needs, preferences and past activities. Where one patient may benefit from weekly reminders through both emails and texts, a single phone call might be best for another patient. The No-Show Predictor can help schedulers determine the correlation between reminder protocols and compliance – and craft the optimal mix to ensure patients receive the care they need.

Patient-specific scheduling insights – The No-Show Predictor can also provide patient-specific insight at the time of scheduling to help lower no-show risk. Schedulers can see which patients are at risk, gauge the level of risk they present, identify key factors affecting risk and determine which intervention options might be most successful. For example, for patients who live far away from the appointment site and routinely miss early morning appointments, schedulers may be alerted to suggest a later appointment time. 

Provider-level coordination – While there are only so many ways to influence patient behaviors, hospitals like CSH have a significant opportunity to work with care providers to find ways to improve appointment attendance. For instance, does one provider consistently have a higher patient no-show rate than their peers? Using the No-Show Predictor, CSH can identify provider-level trends as a starting point for useful, data-informed conversations about no-show rates. 

Real results

Children’s Specialized Hospital has 14 locations across New Jersey. After the first three months of deploying the No-Show Predictor, leaders measured an 8.5% reduction in no-shows across its clinical network. “In one clinic, we’ve seen a 63% decrease in no-shows,” highlights Chuck Chianese, Vice President and Chief Operating Officer at CSH.

The hospital expects to achieve additional reductions as it refines its approach using the No-Show Predictor. “It’s enabled us to really take a deep dive into what’s driving our no-show cancellations and map it to real solutions, to effectuate meaningful reductions and also a significant increase in revenue associated with those reductions,” says Chianese.