Percentages were calculated as weighted number of ED visits with primary diagnosis of HZ over the total number of ED visits, multiplied by 100.
Proportion of ED visits for HZ and charge for HZ-related ED services from 2006 to 2013 overall (A), in patients aged less than 20 years (B), 20 to 59 years (C), and 60 years or older (D). All charges were adjusted to 2016 dollars using the Consumer Price Index adjustments for medical care.
All charges were adjusted to 2016 dollars using the Consumer Price Index adjustments for medical care.
eTable. Nationwide Herpes Zoster Emergency Department Visit Characteristics for All Ages From 2006-2013
Customize your JAMA Network experience by selecting one or more topics from the list below.
Dommasch ED, Joyce CJ, Mostaghimi A. Trends in Nationwide Herpes Zoster Emergency Department Utilization From 2006 to 2013. JAMA Dermatol. 2017;153(9):874–881. doi:10.1001/jamadermatol.2017.1546
What is the effect of vaccination on emergency department utilization for herpes zoster?
In this population-based, descriptive epidemiologic study of more than 1.3 million herpes zoster–related emergency department visits from 2006 through 2013, the overall rate of visits rose by 8.3%, driven by an increased number of visits by patients aged 20 to 59 years; populations recommended for vaccination (aged <20 and ≥60 years) demonstrated a decrease in utilization. Herpes zoster annual incidence rates fell after the recommended use of the herpes zoster vaccine in 2008 in patients 60 years or older.
Vaccination may be associated with a reduction of emergency department utilization for herpes zoster.
The effect of vaccination on emergency department (ED) utilization for herpes zoster (HZ) has not been examined to date.
To determine trends in US ED utilization and costs associated with HZ.
Design, Setting, and Participants
The Nationwide Emergency Department Sample data set was examined for temporal trends in the number of visits and costs for treatment of HZ in EDs in the United States from January 1, 2006, through December 31, 2013. Cases of HZ were identified using validated International Classification of Diseases, Ninth Revision–Clinical Modification diagnosis codes. Patients were stratified by age: less than 20 years (varicella vaccine recommended), 20 to 59 years (no vaccine recommended), and 60 years or older (HZ vaccine recommended). Population-based rates were estimated using sampling weights.
Main Outcomes and Measures
Population-based incidence rates of HZ-related ED visits, charge for ED services, and total charges.
A total of 1 350 957 ED visits for HZ were identified between 2006 and 2013, representing 0.13% of all US ED visits. Of these patients, 563 200 (51.7%) were male; mean (SE) age was 54.0 (0.1) years. Between 2006 and 2013, the percentage of HZ-related ED visits increased from 0.13% to 0.14% (8.3%). This growth was driven by patients aged 20 to 59 years (increase of 22.8% [from 0.12% to 0.14% of ED visits]) while the proportion of ED HZ visits decreased for patients aged less than 20 years and 60 years or older, from 0.03% to 0.02% (−39.6%) and from 0.28% to 0.25% (−10.9%), respectively. For all age groups, there was an increase from 2006 to 2013 in overall adjusted total (from $92.83 to $202.47 million) and mean charges (from $763 to $1262) for HZ-related ED visits.
Conclusions and Relevance
The number of ED visits and total cost associated with HZ increased between 2006 and 2013. Greater use was driven by an increased number of visits by patients aged 20 to 59 years, but populations recommended for vaccination (<20 and ≥60 years) demonstrated decreased ED utilization. Per-visit and total costs increased across all age groups. Vaccination may be associated with a reduction of ED utilization. Further research is required to confirm these results and examine the drivers of increased ED costs.
Herpes zoster (HZ), or shingles, is caused by reactivation of the HZ virus, the same virus that causes varicella.1 One-third of individuals in the United States will develop HZ in their lifetime, with the risk increasing sharply after age 50 years.2 Any individual who has had varicella or received the varicella vaccine can develop HZ, although the risk is lower in those who are vaccinated.3
The rate of HZ has been increasing gradually over many years in the United States and other countries. The reason for this trend is currently unknown.4-6 According to a hypothesis proposed by Hope-Simpson, exposure to primary varicella or subclinical latent varicella zoster virus can boost varicella zoster virus–specific immunity and reduce the risk of HZ.7 A live-attenuated vaccine for varicella was introduced into the US vaccination schedule in 1995 for children aged 12 months or older, with subsequent significant declines in the incidence of varicella. Although vaccination rates were slow at first, they reached 89% in 2006 in children aged 19 to 35 months.8 Observational studies suggest that individuals vaccinated for varicella are at a decreased risk of developing HZ.3 However, studies examining the effect on the overall population rates in the United States with observation periods ranging from 1992 through 2010 have been inconclusive.3-6,9
A live-attenuated vaccine for HZ has been available in the United States since 2006. The vaccine reduces the likelihood of developing HZ by 51.3% in individuals aged 60 years or older, and vaccinated individuals who develop HZ usually have less severe disease. In 2008, the Advisory Committee on Immunization Practices (ACIP) recommended zoster vaccine live for the prevention of HZ and its complications among adults aged 60 years or older.2 Vaccination rates in this population increased from 6.7% in 2008 to 24.2% in 2013.10,11
To our knowledge, there are no data on nationwide emergency department (ED) utilization, hospitalization, and costs related to HZ and limited data on trends of disease since routine varicella vaccination was implemented in 1995 for children, as well as both before and after the ACIP recommendation of HZ vaccination in 2008. We hypothesized that, despite the increased incidence of HZ, the introduction of the HZ vaccine in 2008 led to attenuation of severe disease and reduction of ED utilization for patients aged 60 years or older with HZ. We additionally hypothesized that routine varicella vaccination decreased ED utilization for HZ in patients aged less than 20 years.
In this study, we used the Nationwide Emergency Department Sample (NEDS) to evaluate trends in ED utilization and costs to assess the potential impact of routine varicella vaccination and HZ vaccination.
We conducted a US population–based, descriptive epidemiologic study of HZ-related ED visits from 2006 through 2013 using the NEDS, a nationally representative all-payer database that includes a 20% stratified sample of US community hospitals and contains data on approximately 130 million ED visits per year when weighted.12 Cases of HZ were identified by a primary International Classification of Diseases, Ninth Revision–Clinical Modification (ICD-9-CM) code of 053.x, which has been shown to detect HZ with 98% sensitivity and a positive predictive value between 85% and 100%.13-15 The Partners Healthcare Institutional Review Board approved the study and granted an exemption for use of deidentified data from the NEDS database.
We estimated the number of population-based total and HZ-related ED visits using sampling weights, clusters, and strata provided in the NEDS databases for each year. We stratified patients into the following age groups: less than 20 years, 20 to 59 years, and 60 years or older. Population-based incidence rates were calculated using US Census Bureau American FactFinder.16 Emergency department charges were adjusted to 2016 dollars using the Consumer Price Index adjustments for medical care.17 Separate logistic regression models for each age group were used to estimate the change in the log odds of a HZ visit per 1-year increase in the survey, and significant coefficients were interpreted as a trend in the probability of an HZ visit over the survey time period. We used the Rao-Scott χ2 test to compare the proportion of visits for HZ from 2006-2007 with those from 2008-2013 among patients 60 years or older. P values <.05 were considered statistically significant. All tests were 2-tailed except for the Rao-Scott χ2 test, which is 1-tailed. All group comparisons were unpaired. Analyses were performed with survey procedures in SAS, version 9.4 (SAS Institute Inc).
From 2006 to 2013, there were 1 350 957 visits to the ED for HZ, which represented 0.13% of all ED visits in the United States (Table 1). When stratified by age groups, visits by patients 60 years or older represented 42.3% of all ED visits for HZ compared with representing 21.3% of ED visits without a primary diagnosis of HZ. Most patients were women (58.3%) and there was a slightly higher number of visits for HZ in the summer months (June through August). Overall, 7.2% of patients who presented to the ED with a primary diagnosis of HZ were admitted, with a mean (SE) length of stay of 4.72 (0.04) days (Table 1). The percentage of patients admitted declined from 2006 to 2013 for all age groups. We found that the overall percentage of HZ-related ED visits (Figure 1), population-based incidence rates (eTable in the Supplement), percentage admitted, and mean ED and inpatient charges, rose with increasing age (Table 2).
From 2006 to 2013, the overall population-based rate for HZ-related ED visits increased from 51.57 to 59.44 per 100 000 persons per year (15.2%). The overall percentage of HZ-related ED visits also increased from 0.13% to 0.14% (8.3%). After stratifying the sample into age groups (<20, 20-59, and ≥60 years), the percentage of ED HZ visits from 2006 to 2013 increased for patients 20 to 59 years from 0.12% to 0.14% (22.8%) (Figure 1). The percentage of ED HZ visits decreased for the <20 years and 60 years or older groups, from 0.03% to 0.02% (−39.6%) and from 0.28% to 0.25% (−10.9%), respectively (P < .001 for the change in odds of an HZ visit by year for all age groups).
For all age groups, there was an increase from 2006 to 2013 in overall adjusted total (from $92.83 to $202.47 million) and mean charges (from $763 to $1262) for HZ-related ED visits (Figure 2 and Figure 3). For patients 60 years or older, although incidence rates declined, total charges increased by 110%. This rise was due to an increase in that population, as well as a greater mean charge per visit. Adjusted mean charges per HZ-related ED visit increased the most for patients 60 years or older (75.5%) from 2006 to 2013 compared with those 20 to 59 years (57.3%) and less than 20 years (58.6%) (Figure 3). The increases in adjusted mean charges per ED visit for HZ outpaced those seen for all ED visits for any indication for patients 60 years or older and those less than 20 years, which increased by 72.8% and 45.9%, respectively, but were similar for the 20- to 59-year age group (59.5%).
We identified increasing utilization and costs associated with treatment of HZ in US EDs. Although the overall proportion of ED visits for HZ is increasing marginally (0.13% to 0.14% from 2006 to 2013), stratification by age revealed that this growth is driven by patients between ages 20 and 59 years, while younger and older patients have had fewer ED visits for HZ over time. We found that a higher percentage of women presented to the ED with HZ (58.3%), which is consistent with the results of almost all studies identifying a greater disease burden among women.2 However, women were also higher utilizers of the ED for any indication (Table 1).
The trends in ED utilization may reflect a combination of influences. Patients seen in the ED for HZ must (1) develop HZ and (2) choose the ED as their venue of care. The trends that we observed thus could be attributable to trends in the incidence of HZ and/or ED utilization among different age groups. We found that, although the overall incidence rate of ED visits for all indications for patients 60 years or older and those less than 20 years remained stable from 2006 to 2013, there was a decrease in the rate of ED visits for HZ for these age groups. This finding suggests that the observed decrease may be due to a reduced disease incidence among individuals vaccinated for varicella (<20 years) and those who qualify for HZ vaccination (≥60 years), potentially providing further insight into the population-level impact of vaccination.
Although we found a significant difference in the proportion of ED visits related to HZ before and after the recommended use of the HZ vaccine in 2008 (P < .001) in the 60 years or older age group, we were unable to perform a similar analysis examining the effect of the varicella vaccination (which was introduced into the vaccination schedule in 1995) on HZ incidence in the younger than 20–year age group since the NEDS database started in 2006. Thus, the decline seen in this population could have been associated with other factors. However, the decrease in HZ incidence in patients aged less than 20 years is consistent with prior studies in the outpatient setting that found a stabilization or decrease in age-specific incidence of HZ in the varicella-vaccinated population before and after 1996.4,5 Among patients aged 20 to 59 years, there was an increase in the rate of HZ-related ED visits from 2006 to 2013, which may have been partially due to an elevated rate of ED visits for all indications in this age group. However, HZ-related ED visits represented an increasing percentage of all ED visits over this time period, which suggests a rising disease incidence in this age group.
Other studies have found an increasing incidence of HZ in the United States and other countries, consistent with our findings.2,18-21 The cause for this elevation in disease burden is unclear and is likely multifactorial. The age-specific increased rate of HZ-related ED visits in patients aged 20 to 59 years may be explained in part by the Hope-Simpson hypothesis that increased varicella zoster virus vaccination has led to a rise in overall HZ incidence.7 However, other studies have found that the rate of HZ was rising before introduction of routine varicella zoster virus vaccination and that this increased rate did not accelerate after routine vaccination was started.19 Unfortunately, the NEDS data collection began in 2006 and we could not examine prior trends.
The substantial increase in costs is also likely to be multifactorial. The NEDS database does not provide insight into the treatments administered at each visit, and it is unclear whether changes in practice (additional testing or imaging) or increased costs for similar therapeutics are the culprit.
These findings suggest that, although vaccination rates may be correlated with a broader effect on HZ incidence and downstream utilization of ED services, reduction in disease incidence or ED presentation may be insufficient to reduce overall health care expense. Further research is needed to identify drivers of increased costs and implement interventions to streamline treatment and evaluation of patients with HZ. Currently, there is a subunit vaccine containing varicella zoster virus glycoprotein E and the AS01B adjuvant system (called HZ/su) being evaluated for the prevention of HZ. Because it is not a live vaccine, the HZ subunit vaccine is considered safe for individuals who are immunocompromised and has been found to have an efficacy rate of 97.2% in adults 50 years or older in a phase 3 trial. This vaccine may provide a means by which we can reverse the disease burden and associated costs of HZ in this population.22
These results must be interpreted in the context of our study design. Although we present these changes in ED utilization over a period marked by the introduction of vaccines, this study does not allow us to directly link change in utilization to vaccination. This limitation is especially true in the context of the substantial change in the American health care system that occurred during the study period—the real driver of changes in ED presentation and cost may be larger systems issues. Aside from primary payer type, there are no indicators of changing access to care or health care coverage, and we were not able to adjust for this in our analysis.
Although the NEDS database provides nationally representative estimates for ED utilization, there are limitations to these data. Because this is a publicly available, deidentified data set, validation by medical records review of HZ cases identified by ICD-9-CM codes was not possible. However, prior studies have found that the algorithm used in the present study detected HZ with a positive predictive value of 85% or more.13-15 In addition, because our study was limited to ED visits, the results may not be generalizable to the outpatient setting.
Our findings should be corroborated in other data sets, especially those including outpatient visits for HZ, that allow for evaluation of the impact of patient characteristics and validation of our findings in other settings outside of the ED.
Our study found an increase in total ED visits associated with HZ between 2006 and 2013 due to an increased number of visits by patients aged 20 to 59 years. Despite decreased utilization in patients aged less than 20 years and 60 years or older, we found increased total adjusted charges in these populations. Our findings suggest that vaccination may be associated with a reduction of ED utilization. Further research is necessary to identify the drivers of increased costs.
Accepted for Publication: April 6, 2017.
Corresponding Author: Arash Mostaghimi, MD, MPA, MPH, Department of Dermatology, Brigham and Women’s Hospital, 15 Francis St, PBB-B 421, Boston, MA 02115 (firstname.lastname@example.org).
Correction: This article was corrected on August 2, 2017, to fix an age range error repeated in the Abstract, Key Points, Introduction, and caption to Figure 2.
Published Online: June 21, 2017. doi:10.1001/jamadermatol.2017.1546
Author Contributions: Drs Joyce and Mostaghimi had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Dommasch, Mostaghimi.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: All authors.
Administrative, technical, or material support: Mostaghimi.
Supervision: Dommasch, Mostaghimi.
Conflict of Interest Disclosures: None reported.
Create a personal account or sign in to: