Predictive scheduling

From Clinfowiki
Revision as of 02:43, 28 October 2014 by Jamesgaor (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search


TIme is a valuable currency in the modern age. One that is likely evident to providers, patients and payors. How can we leverage current and future technologies to reduce wasting time and and maximize time efficiency in daily clinical workflow and operations?

Healthcare providers have a long history of frustration with no shows and late patients. The problem is amplified in the context of very tight schedules where every minute is of tremendous value. As physician panels increase in size, their time will become more valuable, and this problem will continue to grow. Research into why patients may be late or not show up to appointments point to a patient’s perceived lack of respect from the health care system, their lack of education around the scheduling system, or a lack of a personal relationship with their provider.1

Physician offices have a notorious reputation for long waits for patients. Patients are often instructed to arrive 10-20 minutes prior to their appointment time, but find they are waiting for various amounts of time to be seen by the provider. Healthcare providers often find themselves waiting 5-20 minutes for patients to arrive to their appointment which can put their entire schedule for the day behind creating further patient dissatisfaction.

Late and no show visits cost a lot of money to an organization. Physicians who run behind schedule create frustrated patients with low satisfaction scores and potential loss of business through negative patient reviews.

Current Solutions:

Many provider offices use the Televox system for automated reminder phone calls. A study of about 10,000 patients showed a 5.8% reduction in the no show rate for those who received automated calls.2 This same study showed that when the staff called patients personally the reduction was 9.5%. Kaiser permanente implemented an SMS text message reminder which when reviewed for approximately 32,000 patients showed a decreased in the no-show rate by 0.73%, which was equivalent to approximately $275,000 in one office, based on a now show cost of $150 per visit.3 Although there is tremendous value in a personal phone call, the cost and work force required to complete this may not be cost effective.

Open access scheduling solutions have been gaining traction in the last few years. Patients are now able to view their provider’s schedule and choose the appointment time most convenient to them. Kaiser Permanente reports that in 2010, 1.8 million physician appointments (5.5%) were booked online.4 Allowing scheduling transparency for patients and enabling them to make their own schedule theoretically can enhance practice workflow efficiency and reduce no-show rates. Healthcare technology companies such as ZocDoc ([1]) help match patients to physicians based on insurance as well as schedule availability. Other companies such as Smart Scheduling ([2]) are leveraging data analytics to improve clinic workflow efficiency by predicting which patients that are scheduled are likely to be a “no show” and plan the rest of the office schedule around that. Data mining patients’ medical records and scheduling history is necessary to create algorithms that can predict high risk no-show patients and trigger staff to communicate with the patient regarding their upcoming appointment. Smart scheduling is reporting approximately 70% accuracy in predicting no-shows utilizing 722 variables for its algorithms5.

Future Solutions:

Data mining and analytics has the potential of optimizing clinical workflows beyond its current state. More sophisticated algorithms could be utilized to not only predict at risk no-show patients but also to predict how much time will be required for any given patient using multivariable analysis including patient demographics, reason for visit, medical history, and scheduling history. Currently, many assumptions are made regarding the amount of time allotted for any given patient, such as new patients are given an arbitrary 45 minute time slot, whereas a known patient is given 15 minutes. These assumptions are not necessarily data driven and using data analytics to help predict the time needed for any particular patient for an upcoming visit could optimize clinical workflow and optimize the time needed to address a patient’s medical needs and preventing the healthcare provider from being rushed from one encounter to the next. Thus, the scheduling paradigm will be patient centered and data driven, with the benefit of decreasing the burden of the healthcare provider and staff from managing unexpected wait times to be seen.

Submitted by James Gaor


1. Lacey, Paulman, et al. Why We Don’t Come: Patient perceptions on no shows. Ann Fam Med November 1, 2004 vol. 2 no. 6 541-545 2. Parikh, A. et al. The Effectiveness of Outpatient Appointment Reminder Systems in Reducing No-Show Rates The American Journal of Medicine Volume 123, Issue 6 , Pages 542-548, June 2010. 3. Giselle Tsirulnik, Kaiser Permanente cuts patient communication costs with SMS. [3] 4. Galewitz,Phil. More Patients Making Appointments Online as Doctors Embrace Web. [4] 5. Farrell, Michael B. Data-driven scheduling predicts no-shows. [5]