Abstract
Purpose
To investigate determinants of no-show rates in an academic pediatric otolaryngology practice including appointment time, age, sex, new patient status, payer mix, and median household income by zip code.
Materials and methods
Retrospective chart review of clinic no-show rates and patient demographics in a free standing children’s hospital and affiliated outpatient clinics across eight providers in a one-year period.
Results
Analysis shows that the overall no-show rate across all providers was 15% with the highest rate of 19% in the zip code with the lowest median income. Highest no-shows are in June, but overall, seasons did not play a significant role in no-show rates. Male gender, morning appointments, and having public insurance appear to significantly predict no-shows. Lost revenue on no-shows range from $191K to $384K per year. The average percentage of the amount billed paid by insurance range from the lowest by out-of-state Medicaid at 16% to the highest by managed care at 54%.
Conclusions
No-show rates account for a significant portion of lost revenue in the outpatient setting for an academic practice, and can be predicted by lower median income, male gender, morning appointments, and public insurance. Such patients may need different appointment reminders. Future clinic templates should be optimized for no-shows to increase productivity and access to care.
1
Introduction
High rates of nonattendance at clinics are a significant problem across all specialties. In the US, clinics have reported nonattendance rates as high as 55% . Nonattendance accounts for a substantial source of wasted healthcare costs and inefficiency, in addition to decreased revenue and lost time for physicians and staff . In 2009, the average cost per office visit ranged between $132 and $340, depending on the specialty . In the face of rising healthcare costs and budget cuts, clinics experience an increased pressure to see more patients in less time. As a result, nonattendance needs to be addressed — not just to increase practice productivity, but also to ensure quality and access to care for patients.
Evaluating predictors of nonattendance can lead to the development of targeted interventions that will reduce costs, improve clinical outcomes, and increase provider productivity. Miller et al. showed that male race, younger age, and African American race significantly predicted repeated nonattendance in a general otolaryngology practice in Michigan . Furthermore, they found that Medicaid insurance and closer distance to appointments were correlated with these no-shows. Demographic and socioeconomic predictors of nonattendance have not been well documented among pediatric otolaryngology clinics in the US . In this study, we investigate potential predictors of nonattendance such as age, gender, new patient status, AM appointment time, zip code, payer mix, and median household income. The purpose of the study is to investigate determinants of no-show rates in a diverse academic pediatric otolaryngology practice. Based on provider observations and prior research findings, our hypothesis is that lower income patients with public type insurance and follow-up appointments particularly in the morning time slot will have higher no-show rates than those in higher income brackets with private insurance.
2
Materials and methods
2.1
Study design
The study was conducted at The Children’s National Health System and its four affiliated outpatient clinics (ENT Specialists of Shady Grove, Upper Marlboro Outpatient Center, Children’s National Specialists of Virginia, and Laurel Outpatient Center) located in Washington, DC, and surrounding areas of Virginia and Maryland, also known as the Washington metropolitan area. This area has a population of over six million people and is 47% White, 25% African American, 15% Spanish, 10% Asian, and 3% are of other races .
We performed a retrospective analysis using the Cerner EMR system, tabulating all no-shows across an academic free standing pediatric hospital and its four affiliated outpatient clinics across eight providers during a one-year period from 2015 to 2016. Among the no-show patients, information on appointment time, appointment type (follow-up or initial new patient visit), insurance type (Medicaid or private), and demographics such as age and gender were collected. Clinic zip codes along with 2010 census data were used to estimate median household income. By choosing the clinic location for their appointment, we assumed that the patient lived within that clinic’s zip code. Payer mix from the various clinics was calculated from 2 years of billing data based on insurance type and collections information.
2.2
Statistical analysis
Data from providers who did not have a full 12 months’ worth of demographic data was excluded from analysis for consistency. No-show rates were calculated based on the total number of no-shows/total number of patients for each provider and location per month. Further calculations were completed to acquire a whole year rate for each provider and location. Several comparisons were made based upon the characteristics being compared. These included the following:
- •
If the proportion of male, AM appointments, public payer, and new patient no-shows was an expected 50% using a two-sided binomial probability test with p-values < 0.05 denoted as significant.
2
Materials and methods
2.1
Study design
The study was conducted at The Children’s National Health System and its four affiliated outpatient clinics (ENT Specialists of Shady Grove, Upper Marlboro Outpatient Center, Children’s National Specialists of Virginia, and Laurel Outpatient Center) located in Washington, DC, and surrounding areas of Virginia and Maryland, also known as the Washington metropolitan area. This area has a population of over six million people and is 47% White, 25% African American, 15% Spanish, 10% Asian, and 3% are of other races .
We performed a retrospective analysis using the Cerner EMR system, tabulating all no-shows across an academic free standing pediatric hospital and its four affiliated outpatient clinics across eight providers during a one-year period from 2015 to 2016. Among the no-show patients, information on appointment time, appointment type (follow-up or initial new patient visit), insurance type (Medicaid or private), and demographics such as age and gender were collected. Clinic zip codes along with 2010 census data were used to estimate median household income. By choosing the clinic location for their appointment, we assumed that the patient lived within that clinic’s zip code. Payer mix from the various clinics was calculated from 2 years of billing data based on insurance type and collections information.
2.2
Statistical analysis
Data from providers who did not have a full 12 months’ worth of demographic data was excluded from analysis for consistency. No-show rates were calculated based on the total number of no-shows/total number of patients for each provider and location per month. Further calculations were completed to acquire a whole year rate for each provider and location. Several comparisons were made based upon the characteristics being compared. These included the following:
- •
If the proportion of male, AM appointments, public payer, and new patient no-shows was an expected 50% using a two-sided binomial probability test with p-values < 0.05 denoted as significant.
3
Results
A total of 17,569 outpatient visits were scheduled across all five locations during this one-year period. A total of 2587 no-show appointments were tabulated for an overall no-show rate of 15%. More than half were male patients and 37% were new patient encounters (see Table 1 ).
