When a rheumatology patient misses an appointment, it can disrupt the clinic schedule and create a missed opportunity for the patient’s access to care. Previous research on patient reminder calls has demonstrated their value in reducing patient no-show rates.1 Recent research leveraging electronic health record (EHR) data applied to predictive models has also shown success in reducing the patient no-show rates. Such predictive modeling of missed appointments in clinical practices has improved scheduling efficiency by enabling the practice to fill open schedule slots.
You Might Also Like
Also By This Author
As predictive modeling approaches to data analytics becomes more prevalent in healthcare, rheumatologists can benefit from understanding how their clinic data can be used to improve scheduling, according to Scott R. Levin, MS, PhD, associate professor of emergency medicine at Johns Hopkins University School of Medicine, Baltimore.
Put Data to Work
Working with a team of researchers from the Johns Hopkins University’s Malone Center for Engineering in Healthcare, Dr. Levin and colleagues developed a new EHR-integrated algorithm that can reduce no-show rates and increase appointment availability for physician offices.
In 2018, Dr. Levin shared the results of using the algorithm in several physician practices within the Johns Hopkins Community Physicians system.2 These practices included a pediatric practice that achieved 70 additional appointments per week and a 16% reduction in no-show appointments using the predictive data gathered from the algorithm. The algorithm is now available for purchase through StoCastic, a Baltimore-based healthcare improvement company co-founded by Dr. Levin.
Much of the innovation around predictive data modeling has targeted acute care settings and larger EHR vendor systems and has been less accessible to smaller physician practices. Dr. Levin notes, this accessibility is likely to change as EHR interoperability becomes more ubiquitous and private practices see the value of using predictive analytics to improve their operations.
Some research suggests data from an individual clinic can provide effective predictive information. A 2018 study by researchers from Duke University on missed appointment predictive data modeling compared the accuracy of predictive data from the health system, specialty and clinic levels. Researchers showed that the more specific patient population data from the clinic-level modeling increased prediction accuracy.3
Using Predictive Data Successfully
Dr. Levin suggests four important points to help rheumatologists apply predictive modeling to improve their missed appointment rates.
- Define missed appointments in an operationally meaningful way: Dr. Levin gives the example of a patient who cancels 24 hours before an appointment. “Although this is not traditionally considered a no-show [appointment] within the data, if this slot cannot be backfilled because of lack of time, the effect on the practice is identical to a no-show,” he says.
- Include all pieces of the data puzzle in missed appointment modeling: Often, the likelihood of a missed appointment is driven by social, clinical and use factors, such as prior missed visits or scheduling lag that are represented in EHR data, he notes. “Although not always required, it’s generally important to include all three pieces of the puzzle to accurately predict no-show [appointments].” Examining these predictor data no-show relationships alone can be useful in supporting re-design of scheduling practices, he adds.
- Establish the central objectives of the predictive analytics: Dr. Levin gives the example of performing more or newly targeted outreach to reduce the number of missed appointments of current patients, as well as designing smart scheduling strategies that account for no-show appointments as common objectives.
- Have a plan to operationalize the predictive insights of predictive data modeling: Having a clear plan to meet the objective(s) of applying predictive analytics is equally important to the success of the predictive technology itself. Dr. Levin says, “Just predicting and displaying no-show [appointment] risk without a strategy can be futile.”
Dr. Levin and colleagues are currently working on the operational piece of predictive analytics to help practices, including for smaller practices, put their no-show appointment data to work. He welcomes practice leaders who want to reach out for more information about the algorithm.
Editor’s note: Under a license agreement between StoCastic and the Johns Hopkins University, Dr. Levin and the university are entitled to royalty distributions related to the technology described in this article. Dr. Levin is a founder of StoCastic, and both he and the university own equity in the company. This arrangement has been reviewed and approved by the Johns Hopkins University in accordance with its conflict of interest policies.
Carina Stanton is a freelance science journalist based in Denver.
- Beach S, Goglin S, Margaretten M, et al. No more no shows: Improving the appointment reminder system at an urban county hospital outpatient rheumatology clinic [abstract #2499]. Arthritis Rheumatol. 2015 Oct; 67(suppl 10).
- Graham C. New tool helps doctors determine which patients are most likely to forget or skip their appointments. The Hub. 2018 Oct4.
- Ding X, Gellad ZF, Mather C, et al. Designing risk prediction models for ambulatory no-shows across different specialties and clinics. J Am Med Inform Assoc. 2018 Aug 1;25(8):924–930.