…where fame and fortune await those who help keep pediatric practices alive and independent!

Pediatric No Show Analysis, Part I

September 04, 2018 / 2 Comments / in predictive analytics, appointments / by Chip Hart

Pediatrician Checking WatchCan you predict when a patient is going to no-show?  Perhaps not perfectly, but you can with a lot more accuracy than you might think.

I've split this blog update into two parts.  Part II will follow after folks have had a few days to digest this one.  There's a lot here to think about.

Earlier this year, Igor and I were talking about one of the great unexplored areas of the pediatric practice management map: appointment data. Your appointment book is the heartbeat of the office, yet remarkably little is known or understood regarding best practices, benchmarks, or performance. What is the optimal scheduling template? What is a normal no-show rate? What kind of policies and procedures does an effective office use?

We were asking ourselves these questions when I realized that I was saying, "I feel like..." and "I think that..." far too much. We needed to analyze data.

Fortunately, PCC has appointment data. Millions and millions of appointments' worth, in fact. And although we have the expertise in-house to make judgments about the data, we have some grand plans, so we employed the folks at Rexer Analytics. We met Karl Rexer at a Predictive Analytics Healthcare event last year and he introduced us to some different ways to think about all the data we have, so we were excited to pull in him and one of Rexer’s Senior Consultants, Heather Allen, to help us.

We started with a seemingly simple question: How well can we predict that a patient will miss an appointment? Our premise is that if we can identify the patterns that affect Missed Appointment rates, we can have a national discussion about best practices to help minimize or, at least, mitigate, missed appointments.

Rexer began by reviewing practice-level data for over a few hundred pediatric practices, and then later identifying a core sample of PCC clients representing different practice types (large, small, high Medicaid, low Medicaid, etc.) and analyzing about 1,000,000 appointments. Here are the highlights of what we found:

First, the benchmarks. The average practice has a missed appointment rate of 4.7%, but practices range from <1% to over 20%. However, if you exclude walk-ins and same-day appointments, the number is more like 10%. That's a large number and, for many practices, represents the difference between red and black on the books. This is why we are interested!

Together, PCC and Rexer identified a wide variety of factors that might impact missed appointments.  Then Rexer examined the data to let the data speak for itself.  Factors we looked at included:

  • Insurance Carrier
  • Historic no show rate for patient
  • Length of appointment
  • Age of patient
  • Time since scheduled
  • Month of appointment
  • Time of day of appointment
  • Day of appointment
  • Sick visit
  • Immunization only visit
  • Total number of historical appointments
  • Family historic no show rate
  • Practice missed appointment policy

What did we learn? 

  • No, your patients are not more likely to miss appointments on Mondays. Or on Fridays. It just feels that way. In fact, there's essentially no distinction among the days about the odds of a patient missing an appointment.  For that matter, neither the age of the patient, the month, nor the season have any particular effect on no-show odds.
  • Appts before 8am and after 8pm are almost never missed. Many of these appointments are same-day, of course, but there are enough that aren't that we should consider the impact of the other demands on patient time that interfere with making an appointment.
  • Finally, to preempt some expected questions: yes, we looked at the effect of having a "no-show policy" and we looked at the use of a reminder system. Our results were very limited for a couple reasons:
    • No-show policies are inconsistent and inconsistently applied. It is impossible to distinguish between someone who has a 1-strike-and-you're-out policy and someone who charges for missed appointments...sometimes. Further, we can't tell whether the very presence of a policy lowered no-show rates OR if the practices who had low no-show rates felt more confident in creating a policy. There does appear to be a correlation between having a policy and a lower rate, however.
    • We ran into similar challenges with the reminder system usage - to really understand the impact, we have to look at the before- and after- results for each practice using a similar reminder policy (practices vary widely with their usage). This may be our next effort.

You'll have to wait for the next episode to get the punchline, but let's review what we know so far.   If we want to predict the odds of a patient no-showing for a visit, we can ignore the day and month of the appointment.  You can ignore the age of the patient.  You can ignore the time of the appointment if it's before 8am and after 8pm.  In my world, that is very, very valuable information already.  I hope you think so, too.

Tags: predictive analytics, appointments

2 replies

Subscribe

PCC Pediatric EHR Solutions

Recent Posts

PODCASTS

TAGS

ARCHIVE