Abstract
In temperate climates, respiratory syncytial virus (RSV) can occur outside the typical fall/winter/spring season. Using the Premier Healthcare Database, we identified and described hospital-based RSV episodes in the US prior to the COVID-19 pandemic (2017-2020), comparing narrowly-defined in-season (December–February) and off-season (June–August) episodes. We identified 141 427 in-season and 3756 off-season episodes among children (<18 years) and 41 735 in-season and 1243 off-season episodes among adults. Among pediatric episodes, most in- and off-season (60%, 59%) were infants. Black and Hispanic pediatric patients were over-represented in off- versus in-season episodes (Black: 22% vs 15%; Hispanic 30% vs 19%; P < .0001). Trends were similar for adults. A higher proportion of off-season pediatric episodes included antibiotics (54% vs 45%; P < .0001), and among infants and children 1–4 years, were more likely to be individuals with congenital or chronic conditions (P < .0001). Disparities due to race, ethnicity, and underlying conditions were identified among off-season cases.
Keywords: respiratory syncytial virus, RSV, seasonality
Hospital-based off-season respiratory syncytial virus (RSV) episodes, while less frequent in the off-season months, persist throughout the year, even prior to the shift observed during the pandemic era. Racial and ethnic disparities may exist among patients with RSV in the off-season.
Respiratory syncytial virus (RSV) is a common respiratory virus that typically causes mild, cold-like illness. However, among some individuals, such as infants, older adults, and patients with underlying medical conditions, RSV can cause serious illness. In temperate climates, RSV outbreaks typically follow an annual seasonal pattern, occurring during the fall, winter, and spring. Across the continental United States, RSV season onset (first two consecutive weeks when RSV PCR test positivity is ≥3%) typically occurs between mid-September and mid-November and offset (the last two consecutive weeks with ≥3% positivity [1]) typically occurs between mid-April and mid-May. However, annual and geographic variability in RSV season timing exists [2, 3]. Respiratory syncytial virus season onset generally occurs earliest in the South, and latest in the West, with considerable regional variation in onset, offset, and duration [4]. More recently, the COVID-19 pandemic and associated nonpharmaceutical interventions severely disrupted RSV seasonality, with historically low levels of RSV detection during the 2020–2021 season [5]. This was followed by a highly atypical off-season spike in RSV during the summer and early fall of 2021 [5, 6] and substantial RSV transmission in October and November of 2022 [7]. Respiratory syncytial virus circulation has showed signs of returning to prepandemic patterns.
With RSV preventive products (eg, maternal and adult immunization, long-acting monoclonal antibodies) finally approved and available [8, 9], understanding RSV epidemiology across the United States will become increasingly important for effective use of these public health interventions. Studies of RSV epidemiology tend to focus on data collected from seasonal surveillance; however, studies describing episodes of RSV identified outside of the historically defined season are needed, including data that may be able to identify characteristics associated with increased susceptibility or likelihood of experiencing off-season RSV illness. We sought to quantify and describe RSV episodes diagnosed outside of the typical RSV season, in comparison with RSV episodes diagnosed during a typical RSV season, across the United States, over time and among patients of all ages.
METHODS
Data Source
We used the Premier Healthcare Data (PHD), an electronic healthcare database that includes administrative, healthcare utilization, and financial data from patient encounters at selected US hospitals, including nonprofit, nongovernmental, and community and teaching hospitals and health systems [10]. The data do not include insurance claims or electronic health record data, but rather include the hospital administrative and billing information as submitted by the hospitals. The healthcare encounters in the PHD consist of inpatient admissions as well as outpatient (OP) visits, which include emergency departments (EDs), ambulatory surgery centers, and alternate sites of care. The PHD includes hospitals from the nine census divisions (see Supplementary Figure 1 for a map) [10]. We also used data from the Centers for Disease Control and Prevention’s (CDC's) National Respiratory and Enteric Virus Surveillance System (NREVSS) to compare trends of RSV diagnosis capture in PHD with trends reported in national surveillance. National Respiratory and Enteric Virus Surveillance System is a voluntary, laboratory-based surveillance system that monitors temporal and geographic trends in RSV and other viral infections via weekly reports. Participating laboratories include university and community hospitals, selected state and county public health departments, and commercial entities [11]. This study was exempt from human subjects review as the data are secondary in nature, deidentified, and HIPAA compliant.
Study Design and Population
We identified patients with an ICD-10-CM diagnosis code of RSV diagnosis (J12.1 [RSV pneumonia], J20.5 [acute bronchitis due to RSV], J21.0 [acute bronchiolitis due to RSV], or B97.4 [RSV as the cause of diseases classified elsewhere], in either primary or secondary position) during one or more inpatient hospitalizations or OP ambulatory encounters reported in the PHD during the prepandemic era from 1 January 2017 to 31 March 2020. We took an episode-level perspective, in which a single patient could have multiple eligible episodes of RSV over time. An RSV episode was defined as beginning with the earliest (index) date of the qualifying healthcare visit (admission date for inpatient visits, or date of the OP visit) and extending through the following 90 days. If a patient had two or more encounters with RSV diagnosis codes within 90 days of each other, the first encounter was used to define the index date of the RSV episode. When describing healthcare resource utilization (HCRU), we instead took a visit-level perspective so that we could fully describe all RSV-related visits.
The index date was used to categorize seasonality of the episode (in-season or off-season). Because of the variability in RSV season onsets, there is the potential of misclassifying the seasonality of an RSV episode. To minimize this misclassification, we narrowly defined off-season as June, July, and August and in-season as December, January, and February. These months are most typically associated with low RSV and peak RSV activity, respectively. Episodes occurring in the other months (March–May and September–November) were not included to allow cleaner comparisons because these are shoulder months where levels of RSV transmission are expected to be intermediate.
Outcomes
We analyzed patient demographic and baseline characteristics, including age, sex, race, ethnicity, presence of certain comorbid conditions, and payor type. Age was defined on the index date of the first RSV-related visit. Patients 0–17 years of age were considered to be pediatric, and patients ≥ 18 years of age were adults. The presence of selected comorbidities in each RSV episode was determined by assessing all healthcare encounters for that individual at the same hospital during the 12 months preceding and inclusive of the RSV index date. Comorbidities were selected based on conditions considered to be risk factors for serious illness from RSV per the American Academy of Pediatrics (children), or associated as risk factors for severe influenza (adults), including chronic pulmonary, cardiovascular, renal, hepatic, neurologic, hematologic, and metabolic disorders, obesity, and immunocompromising conditions (including conditions associated with immunosuppressive therapy) [12, 13]. For pediatric RSV episodes, comorbidity status was categorized into congenital or chronic medical condition, immunocompromised, or no risk factor. Categories for the adult comorbidities were the same as pediatrics, but excluded congenital conditions. Risk groups were collapsed into a dichotomous “any risk” versus “no risk” variable for the logistic regression models since patients could have both a chronic medical condition and an immunocompromised status. Diagnosis codes for pediatrics and adults are listed in Supplementary Tables 1A and 1B, respectively.
The following HCRU details were assessed: visit type (inpatient, ED visit, or OP clinic visit), ICU admission (yes/no, ICU LOS), inpatient antibiotic use (yes/no; inpatient setting only), hospital length of stay, and clinical outcomes (eg, 30-, 60-, and 90-day readmissions, inpatient mortality). When categorizing visit type, if a patient had an OP (ED or OP clinic) visit with an associated RSV diagnosis code and subsequently had a hospitalization with an associated RSV diagnosis code within 3 days of that OP visit, the event was categorized as an inpatient hospitalization with the index date being the visit date of the OP visit with the RSV diagnosis code.
Statistical Analysis
Baseline demographic and clinical characteristics of RSV episodes were summarized by in-season and off-season groups using descriptive statistics. For categorical variables, differences between in-season and out-of-season RSV episodes were evaluated using a Fisher's exact test. For continuous variables, differences between in-season and out-of-season RSV episodes were evaluated using an unpaired t test. A Bonferroni correction was used to adjust for multiple testing, and a P-value of < .0006 was considered significant.
We modeled the odds of an in- versus off-season RSV episode with logistic regression for the pediatric and adult populations to adjust for confounding variables, including baseline demographic (age, sex, race, ethnicity, payor, and geography) and clinical (presence of a risk condition, hospital teaching status, and population served) characteristics. The resulting odds ratios (ORs) describe the adjusted odds of being diagnosed with RSV in the in-season versus the off-season, with an OR less than one indicating a lower odds of RSV occurring in-season (and a higher odds of RSV occurring off-season) relative to the reference group. All analyses were carried out using SAS v9.4 software.
RESULTS
Table 1 compares baseline characteristics of RSV episodes reported in-season with those reported off-season. Figures 1A and 1B presents the results of the logistic regression models to estimate the odds of an RSV episode occurring in-season versus off-season adjusting for these baseline demographic and clinical profiles to these baseline variables, along with a summary of the presence of a risk condition (detailed description in Table 2).
Table 1.
Baseline RSV Episode Characteristics, In-season (Narrowly Defined as December, January, and February) vs Off-season (Narrowly Defined as June, July, and August) 2017–2020
| Pediatric | Adult | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| In-season N = 141 427 |
Off-season N = 3756 |
P-Value | In-season N = 41 735 |
Off-season N = 1243 |
P-Value | |||||
| Age category (y) | ||||||||||
| <1 | 84 190 | 59.5% | 2202 | 58.6% | .27 | – | – | – | – | – |
| 1–4 | 52 351 | 37.0% | 1438 | 38.3% | .11 | – | – | – | – | – |
| 5–17 | 4886 | 3.5% | 116 | 3.1% | .22 | – | – | – | – | – |
| 18–49 | – | – | – | – | – | 6576 | 15.8% | 231 | 18.6% | .01 |
| 50–64 | – | – | – | – | – | 9523 | 22.8% | 312 | 25.1% | .06 |
| 65+ | – | – | – | – | – | 25 636 | 61.4% | 700 | 56.3% | .0003 |
| Sex | ||||||||||
| Male | 76 396 | 54.0% | 2062 | 54.9% | .29 | 16 409 | 39.3% | 464 | 37.3% | .16 |
| Female | 64 996 | 46.0% | 1694 | 45.1% | .30 | 25 324 | 60.7% | 779 | 62.7% | .16 |
| Unknown | 35 | 0.0% | 0 | 0.0% | .33 | 2 | 0.0% | 0 | 0.0% | .81 |
| Race | ||||||||||
| Black | 21 671 | 15.3% | 829 | 22.1% | <.0001 | 4752 | 11.4% | 182 | 14.6% | .0004 |
| White | 86 987 | 61.5% | 2046 | 54.5% | <.0001 | 31 490 | 75.5% | 907 | 73.0% | .05 |
| Asian | 2503 | 1.8% | 112 | 3.0% | <.0001 | 691 | 1.7% | 15 | 1.2% | .22 |
| Other | 23 528 | 16.6% | 566 | 15.1% | .01 | 3892 | 9.3% | 107 | 8.6% | .39 |
| Unknown | 6738 | 4.8% | 203 | 5.4% | .07 | 910 | 2.2% | 32 | 2.6% | .35 |
| Ethnicity | ||||||||||
| Hispanic | 27 199 | 19.2% | 1125 | 30.0% | <.0001 | 3770 | 9.0% | 302 | 24.3% | <.0001 |
| Not Hispanic | 88 673 | 62.7% | 2176 | 57.9% | <.0001 | 28 466 | 68.2% | 764 | 61.5% | <.0001 |
| Unknown | 25 555 | 18.1% | 455 | 12.1% | <.0001 | 9499 | 22.8% | 177 | 14.2% | <.0001 |
| Primary payora | ||||||||||
| Medicare | 353 | 0.2% | 5 | 0.1% | .16 | 27 956 | 67.0% | 764 | 61.5% | <.0001 |
| Medicaid | 94 080 | 66.5% | 2638 | 70.2% | <.0001 | 5312 | 12.7% | 153 | 12.3% | .66 |
| Commercial | 37 240 | 26.3% | 847 | 22.6% | <.0001 | 6365 | 15.3% | 243 | 19.5% | <.0001 |
| Uninsured | 5213 | 3.7% | 118 | 3.1% | .08 | 1257 | 3.0% | 64 | 5.1% | <.0001 |
| Other | 4541 | 3.2% | 148 | 3.9% | .01 | 845 | 2.0% | 19 | 1.5% | .22 |
| Geography of hospital | ||||||||||
| East North Central | 23 014 | 16.3% | 257 | 6.8% | <.0001 | 6681 | 16.0% | 92 | 7.4% | <.0001 |
| East South Central | 10 035 | 7.1% | 142 | 3.8% | <.0001 | 1757 | 4.2% | 51 | 4.1% | .85 |
| Middle Atlantic | 15 819 | 11.2% | 311 | 8.3% | <.0001 | 10 677 | 25.6% | 197 | 15.8% | <.0001 |
| Mountain | 9095 | 6.4% | 70 | 1.9% | <.0001 | 3113 | 7.5% | 36 | 2.9% | <.0001 |
| New England | 2578 | 1.8% | 12 | 0.3% | <.0001 | 1309 | 3.1% | 24 | 1.9% | .02 |
| Pacific | 10 485 | 7.4% | 291 | 7.7% | .44 | 2418 | 5.8% | 53 | 4.3% | .02 |
| South Atlantic | 34 371 | 24.3% | 1989 | 53.0% | <.0001 | 8971 | 21.5% | 654 | 52.6% | <.0001 |
| West North Central | 10 410 | 7.4% | 117 | 3.1% | <.0001 | 3016 | 7.2% | 52 | 4.2% | <.0001 |
| West South Central | 25 620 | 18.1% | 567 | 15.1% | <.0001 | 3793 | 9.1% | 84 | 6.8% | .00 |
| Hospital teaching status | ||||||||||
| Teaching | 67 083 | 47.4% | 1656 | 44.1% | <.0001 | 22 679 | 54.3% | 641 | 51.6% | .05 |
| Nonteaching | 74 344 | 52.6% | 2100 | 55.9% | <.0001 | 19 056 | 45.7% | 602 | 48.4% | .05 |
| Population served | ||||||||||
| Rural | 25 453 | 18.0% | 412 | 11.0% | <.0001 | 4410 | 10.6% | 84 | 6.8% | <.0001 |
| Urban | 115 974 | 82.0% | 3344 | 89.0% | <.0001 | 37 325 | 89.4% | 1159 | 93.2% | <.0001 |
Abbreviation: RSV, respiratory syncytial virus.
aMedicare and Medicaid are both government-sponsored healthcare programs in the United States. Medicare is a federal insurance program that primarily serves older adults (ages 65+ years) and individuals with disabilities, while Medicaid is a joint federal and state program that provides healthcare coverage to low-income individuals and families.
Figure 1.
Pediatric (A) and adult (B) logistic regression models: odds ratios of respiratory syncytial virus (RSV) in-season versus off-season.
Table 2.
Summary of Risk Factors for Severe RSV in Pediatric and Adult Episodes, In-season (Narrowly Defined as December, January, and February) vs Off-season, (Narrowly Defined as June, July, and August) 2017–2020
| In-season | Off-season | P-Value | |||
|---|---|---|---|---|---|
| N | % | N | % | ||
| Pediatric | |||||
| Overall pediatric | |||||
| No risk factor | 130 220 | 92.1% | 3314 | 88.2% | <.0001 |
| Congenital or chronic medical condition | 10 763 | 7.6% | 427 | 11.4% | <.0001 |
| Immunocompromised | 1270 | 0.9% | 44 | 1.2% | .08 |
| <1 y | |||||
| No risk factor | 78 547 | 93.3% | 1978 | 89.8% | .001 |
| Congenital or chronic medical condition | 5447 | 6.5% | 220 | 10.0% | <.0001 |
| Immunocompromised | 374 | 0.4% | 12 | 0.5% | .52 |
| 1–4 y | |||||
| No risk factor | 47 982 | 91.7% | 1257 | 87.4% | <.0001 |
| Congenital or chronic medical condition | 4148 | 7.9% | 170 | 11.8% | <.0001 |
| Immunocompromised | 590 | 1.1% | 26 | 1.8% | .01 |
| 5–17 y | |||||
| No risk factor | 3691 | 75.5% | 79 | 68.1% | .05 |
| Congenital or chronic medical condition | 1168 | 23.9% | 37 | 31.9% | .29 |
| Immunocompromised | 306 | 6.3% | 6 | 5.2% | .46 |
| Adult | |||||
| Overall adult | |||||
| No risk factor | 20 643 | 49.5% | 620 | 49.9% | .77 |
| Chronic medical condition | 20 635 | 49.4% | 607 | 48.8% | .67 |
| Immunocompromised | 6066 | 14.5% | 206 | 16.6% | .04 |
| 18–49 y | |||||
| No risk factor | 4067 | 61.8% | 147 | 63.6% | .02 |
| Chronic medical condition | 2380 | 36.2% | 80 | 34.6% | .27 |
| Immunocompromised | 939 | 14.3% | 34 | 14.7% | .26 |
| 50–64 y | |||||
| No risk factor | 4679 | 49.1% | 152 | 48.7% | .26 |
| Chronic medical condition | 4724 | 49.6% | 152 | 48.7% | .32 |
| Immunocompromised | 1656 | 17.4% | 63 | 20.2% | .05 |
| 65 + years | |||||
| No risk factor | 11 897 | 46.4% | 321 | 45.9% | .04 |
| Chronic medical condition | 13 531 | 52.8% | 375 | 53.6% | .09 |
| Immunocompromised | 3471 | 13.5% | 109 | 15.6% | .57 |
Patients could be counted in multiple risk factor categories.
Abbreviation: RSV, respiratory syncytial virus.
Among the pediatric RSV episodes, both in- and off-season RSV episodes were most common among infants <1 year of age (59.5% and 58.6%, respectively) and in children 1–4 years of age (37.0% and 38.3%, respectively). Slightly more than half of pediatric RSV episodes occurred in males, with no difference in this pattern in- or off-season (54.0% and 54.9%), respectively. A larger proportion of off-season episodes were among Black (22.1% off-season vs 15.3% in-season), Asian (3% vs 1.8%), and Hispanic (30.0% vs 19.2%) children (all P-values < .0001). These patterns persisted in the adjusted model: relative to White patients, Asian and Black patients had lower odds (OR = 0.47 and 0.66, respectively; P-values < .0001) of being diagnosed in-season. Similarly, Hispanic patients had lower odds of in- versus off-season episodes relative to non-Hispanic (OR = 0.59; P < .0001). In other words, these groups had higher odds of off-season RSV episodes. Compared with in-season episodes, a higher percentage of off-season involved Medicaid as payor (70.2 off-season vs 66.5% in-season; P < .0001), and a lower percentage of off-season episodes involved Commercial insurance as payor (22.6% off-season vs 26.3% in-season; P < .0001). Pediatric episodes involving uninsured were similar among off-season and in-season episodes. Differences among payors were attenuated in the adjusted model. Most (53.0%) of all off-season pediatric RSV episodes were reported in hospitals in the South Atlantic census division, with expected geographic variability across the 9 divisions (see Table 1 and Figure 1 for details). Off-season pediatric RSV episodes were reported more frequently among nonteaching hospitals (55.9% vs 52.6% for off- and in-season, respectively; P < .0001) and hospitals serving urban populations (89.0% vs 82.0% for urban and rural, respectively; P < .0001).
Among RSV episodes in adults (Table 1), the age distribution was similar in-season and off-season. Most (>56%) RSV episodes in adults were reported among adults ≥ 65 years of age, > 20% of episodes were reported among adults 50–64 years of age, and >15% were reported among adults 18–49 years of age in both in-season and off-season episodes. Females represented 60.7% and 62.7% of RSV episodes among adults during the in- and off-seasons, respectively. Most RSV episodes were reported among White adults, for both in-season (75.5%) and off-season (73.0%) episodes. A higher proportion of adult off-season RSV episodes than in-season episodes were reported among Black (14.6% vs 11.4%, P = .0004) and Hispanic patients (24.3% vs 9.0%, P < .0001). The pattern of relatively increased RSV episodes among Hispanic patients in the off-season remained significant in the adjusted model (OR = 0.33; P < .0001). Most episodes involved Medicare as the payor (67.0% and 61.5% for in- and off-season, respectively); a higher proportion of off-season than in-season episodes involved commercial insurance (19.5% vs 15.3%, P < .0001) or were uninsured (5.1% vs 3.0%, P < .0001). Most adult off-season RSV episodes (52.6%) were reported among hospitals in the South Atlantic census division, as compared to only 21.5% of in-season episodes. Off-season compared with in-season adult RSV episodes were reported more frequently among hospitals serving urban populations (93.2% vs 89.4%; P < .0001).
Table 2 summarizes the presence of risk factors for serious RSV illness by pediatric and adult patients. Note that the chronic medical and immunocompromising conditions groups are not mutually exclusive. Among RSV episodes <1 year of age, fewer (88.2%) off-season episodes had no risk factor for serious RSV illness (vs 92.1% of in-season episodes; P < .0001). Among RSV episodes <1 year of age or 1–4 years of age, fewer in-season episodes than off-season episodes had a congenital or chronic medical condition (<1 year of age: 6.5% vs 10.0%, P < .0001; 1–4 years of age: 7.9% vs 11.8%). Among adult RSV episodes, there were no differences in the presence of risk factors among in-season and off-season RSV episodes, either overall, or stratified by age group (18–49, 50–64, and ≥65 years). In the adjusted results (Figure 1), the summarized variable for the presence of any risk factors (vs none) was statistically significant in the pediatric population (OR = 0.66 [95% CI: 0.60–0.73]), suggesting increased odds for off-season RSV among pediatric patients with risk conditions, but not for the adults.
Table 3 summarizes healthcare utilization among pediatric and adult RSV visits by in-season and off-season. The OP setting was defined as either an office/clinic visit, or an ED visit that did not result in an inpatient admission. Note that the PHD is primarily a hospital-based data source, so the majority of OP visits are ED visits. Among pediatric RSV visits in the OP setting, a lower proportion of in-season versus off-season cases occurred in the office/clinic setting (8.8% vs 13.4%, P < .0001). Conversely, within the OP setting, 91.2% of the pediatric in-season visits versus 86.6% of the off-season were treated in the ED without a subsequent hospital admission (P < .0001). Among the overall pediatric visits, a higher proportion of off-season than in-season pediatric cases involved an inpatient visit (33.2% vs 27.1%, P < .0001). Within these pediatric inpatient admissions, a higher proportion of off-season inpatients were admitted from the ED (66.8% vs 65.2%, P < .0001). The median inpatient length of stay was 3 days (interquartile range [IQR]: 2, 5) for both in-season and off-season pediatric RSV inpatient visits. A higher proportion of inpatient pediatric RSV visits in the off-season involved an ICU stay (25.9% off-season vs 23.3% in-season, P < .0001) or antibiotic use (54.3% off-season vs 45.4% in-season, P < .0001). Most inpatient pediatric RSV visits resulted in discharge to home, and 30-, 60-, and 90-day readmission rates were similar among in-season and off-season visits.
Table 3.
Healthcare Resource Utilization in Pediatric and Adult RSV Episodes, In-season (Narrowly Defined as December, January, and February) vs Off-season, (Narrowly Defined as June, July, and August) 2017–2020
| Pediatric | Adult | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| In-Season | Off-Season | P-Value | In-Season | Off-Season | P-Value | |||||
| Total index visits | 144 659 | 100% | 3992 | 100% | N/A | 41 918 | 100% | 1326 | 100% | N/A |
| Outpatient visits | 105 498 | 72.9% | 2668 | 66.8% | <.0001 | 8967 | 21.4% | 362 | 27.3% | <.0001 |
| ED only | 96 223 | 91.2% | 2311 | 86.6% | <.0001 | 7855 | 87.6% | 227 | 62.7% | .14 |
| Office/clinic | 9275 | 8.8% | 357 | 13.4% | <.0001 | 1112 | 12.4% | 135 | 37.3% | <.0001 |
| Inpatient visits | 39 161 | 27.1% | 1324 | 33.2% | <.0001 | 32 951 | 78.6% | 964 | 72.7% | <.0001 |
| Admitted from ED | 25 544 | 65.2% | 885 | 66.8% | <.0001 | 27 924 | 84.7% | 789 | 81.8% | <.0001 |
| LOS | ||||||||||
| Mean (SD) | 4.1 (6.7) | … | 8.3 (24.4) | … | <.0001 | 6.7 (7.5) | … | 7.3 (11.6) | … | .02 |
| Min, max | (0, 294) | … | (0, 304) | … | (0, 574) | … | (0, 271) | … | ||
| Med (IQR) | 3 (2, 5) | … | 3 (2, 5) | … | 5 (3, 8) | … | 5 (3, 8) | … | ||
| ICU stay | 9123 | 23.3% | 343 | 25.9% | <.0001 | 6308 | 19.1% | 169 | 17.5% | .02 |
| Mean (SD) | 4.8 (10.1) | … | 18.9 (43.2) | … | <.0001 | 4.8 (9.5) | … | 6.1 (14.5) | … | .32 |
| Min, max | (1, 216) | … | (1, 303) | … | (1, 553) | … | (1, 159) | … | ||
| Med (IQR) | 2 (1, 5) | … | 2 (1, 8) | … | 2 (1, 6) | … | 3 (1, 6) | … | ||
| Antibiotics | 17 791 | 45.4% | 719 | 54.3% | <.0001 | 28 062 | 85.2% | 856 | 88.8% | .07 |
| Discharge statusa | ||||||||||
| Elevated care | 1171 | 3.0% | 34 | 2.6% | .77 | 8426 | 25.6% | 190 | 19.7% | <.0001 |
| Home | 37 874 | 96.7% | 1279 | 96.6% | <.0001 | 23 078 | 70.0% | 729 | 75.6% | .96 |
| IP mortality | 53 | 0.1% | 3 | 0.2% | .23 | 1339 | 4.1% | 42 | 4.4% | .96 |
| Other/unknown | 63 | 0.2% | 8 | 0.6% | <.0001 | 108 | 0.3% | 3 | 0.3% | .82 |
| Readmission | ||||||||||
| 30 d | 2282 | 5.8% | 66 | 5.0% | .70 | 3031 | 9.2% | 79 | 8.2% | .08 |
| 60 d | 3051 | 7.8% | 91 | 6.9% | .46 | 5176 | 15.7% | 138 | 14.3% | .03 |
| 90 d | 3571 | 9.1% | 116 | 8.8% | .08 | 6604 | 20.0% | 179 | 18.6% | .03 |
Abbreviations: ED, emergency department; IQR, interquartile range; LOS, length of stay; RSV, respiratory syncytial virus.
a“Elevated care” is defined as a discharge status indicating a transfer to another healthcare facility, hospice care, or planned inpatient readmission. “Other/unknown” is a miscellaneous category that includes discharges where information is not available, or discharge/transfer to law enforcement or psychiatric hospitalizations.
Most adult RSV visits were inpatient admissions (78.6% of in-season and 72.7% of off-season). The median length of stay was 5 days (IQR: 3, 8) for both in-season and off-season adult RSV inpatient visits. There were no differences in ICU admission or antibiotic use among off-season or in-season adult RSV inpatient visits. A higher proportion of inpatient adult RSV visits in the in-season involved elevated care at discharge (25.6% vs 19.7%, P < .0001). Readmission rates at 30-, 60-, and 90-days were similar among in-season and off-season adult RSV visits.
To assess the seasonal representativeness of RSV visits captured in PHD, the distribution of RSV visit counts identified in PHD was plotted alongside RSV case counts from NREVSS. Figure 2 shows the overall distribution of RSV diagnoses as captured in PHD and positive RSV tests as reported in NREVSS, by month, from January 2017 to March 2020. The curves for positive RSV tests reported to NREVSS and RSV diagnoses captured in PHD were generally similar. Across the analysis timeframe, positive RSV tests were reported to NREVSS, even during the typical off-season months; this pattern was consistent across census divisions (Supplementary Figure 2A–I).
Figure 2.
Overall monthly respiratory syncytial virus (RSV) case counts—Premier Healthcare Data (PHD) and National Respiratory and Enteric Viral Surveillance System (NREVSS), January 2017–March 2020.
DISCUSSION
This descriptive analysis provides insights into the characteristics of RSV episodes occurring during a typical US RSV season (eg, December, January, and February) compared with those occurring during a typical off-season (eg, June, July, and August); episodes occurring the “shoulder” seasons were not included. Among both pediatric and adult RSV episodes, this analysis showed differences in race or ethnicity among off-season versus in-season RSV episodes. We attempted to adjust for confounders with a multivariable logistic regression. However, it is unknown whether the differences in race/ethnicity seen across all ages represent actual disparities, or result from unmeasured confounding reflecting other differences (variable testing practices).
This study showed that pediatric patients were more frequently admitted to the ICU and antibiotic use was higher among off-season RSV visits. Increased ICU admissions among off-season RSV visits may reflect the presence of underlying chronic conditions, which we did observe in the adjusted model, or increased off-season RSV testing among high acuity visits. The increase in antibiotic use during the off-season may reflect treatment patterns that occur prior to receipt of the positive test results, when physicians are less likely to expect RSV episodes, and more likely to proactively prescribe antibiotics. Prevention of such episodes might decrease the inappropriate use of antibiotics, as well as the negative downstream effects on the patients’ microbiomes.
In terms of the results for the adult population of RSV episodes, aside from the observed disparity in race and ethnicity of episodes, demographic and clinical profiles for the in-season versus off-season cohorts were generally similar. The results indicated a difference in urbanicity (patients receiving care at hospitals serving urban communities were more frequent in the off-season 93.2% vs 89.4% in-season). This could potentially be due to urban hospitals having more resources for testing in the off-season when RSV is less common. Further, it was noted that more adult visits in the in-season required elevated care postdischarge (25.6%) as compared to off-season (19.7%). Additionally, this analysis showed no substantial difference in the presence of preexisting comorbid conditions among in-season and off-season RSV episodes across the adult age groups, with approximately half of adult episodes both in and out of season having no underlying conditions or risk factors for severe RSV, highlighting the potential for RSV to cause medical attended illness even among individuals without underlying conditions and when RSV is not in high circulation.
This analysis has several limitations. As with all real-world data, there are limitations inherent with the PHD that could lead to several biases. Our analysis is vulnerable to misclassification bias since we do not have lab result data on all patients, so the RSV episode definition is based on ICD-10-CM diagnosis codes, which generally underestimate RSV diagnoses when applied to retrospective database analyses. Therefore, this analysis likely underestimates actual RSV episodes in both in-season and off-season [14, 15], with a potential for increased misclassification during the off-season when clinicians are even less likely to suspect RSV and potentially less likely to test for it. Respiratory syncytial virus diagnosis depends on clinical testing, and clinical testing patterns can vary by setting, facility, provider, age group, and time of year [16, 17]. Standard-of-care testing rates are particularly low among adults [17], and RSV cases may be missed by single specimen testing, particularly among adults due to lower viral loads among adults [18, 19] versus children [20]. Thus, it is possible that adult RSV cases are underreported by standard-of-care testing and cases that were identified by standard-of-care testing might not be representative of all RSV cases among adults. Moreover, although we attempted to accurately capture and categorize episodes as in- or off-season, there is the potential for misclassification of seasonality. Because definitions of RSV seasons used in this analysis were applied across all geographic regions even though RSV seasonality can vary by region, there could be misclassification of RSV episodes as in-season when the episode occurred off-season for that region or vice versa; hence, we attempted to control for this source of bias by focusing our analyses on 3-month periods of historical peak and trough RSV detection, while excluding “shoulder” months.
Similarly, comorbidities were identified via an opportunistic lookback period of 1 year prior to, and inclusive of the qualifying RSV visit. If a patient had a comorbid condition but did not receive care at the same hospital in the year prior to their index visit or the diagnosis was not coded on the index visit, that patient would be misclassified as not having that condition.
Other potential biases involve a potential lack of generalizability and lack of precision. Although the PHD represents one in every four discharges in the United States, it is not designed to be nationally representative. Study definitions were sometimes limited by the available data. Conclusions about pediatric RSV would be more robust if age in months was available from PHD. Furthermore, while PHD captures data on race, the categories are limited. For example, it does not capture American Indian or Alaskan Native (AI/AN) as a category, and AI/AN infants and children have been shown to experience higher rates of severe RSV including hospitalization. Further, the PHD is comprised of hospital-based data, whereas NREVSS is comprised of laboratory-based data; thus, exact comparisons of seasonality patterns in PHD and those reported by CDC cannot be directly compared. Both the Premier hospital data and the NREVSS laboratory-based surveillance data are large, national data sources. However, the underlying populations are likely to differ, and neither is designed to be comprehensively representative of RSV circulation. Premier data represent hospitalized patients, while NREVSS data are sourced from public health and clinical laboratories, including collaborating university and community hospital laboratories, selected states and county public health laboratories, and commercial laboratories. The underlying source populations are similar, but not identical, and could drive some variability in RSV detection across the two datasets.
Finally, this study was descriptive in nature only; we did not conduct multivariable analysis to examine the potential for confounding or effect modifying factors that could impact the observed differences between in- and off-season episodes. Patients may have been included in more than one season and were treated as independent episodes. Although we took a conservative approach in adjusting the P-value of significance for multiple testing, these results should be interpreted as hypothesis generating and used to inform future studies. Lastly, differences cited were based on statistical significance and may not necessarily reflect clinically meaningful differences.
CONCLUSIONS
Although more medically attended RSV episodes occur when RSV is in high circulation, medically attended RSV episodes do occur off-season. Our results highlight both the similarities in demographics and clinical characteristics of episodes occurring in-season and off-season as well as the need to further assess the presence of racial and ethnic disparities in medically attended RSV. Future studies should examine differences in race and ethnicity among RSV-related visits throughout the year, ideally in settings where RSV testing is implemented routinely and year-round. Understanding the impact of testing patterns will be crucial to interpreting results. Increasing awareness of the potential of RSV to cause serious illness in certain populations is important to understanding, characterizing, and quantifying the burden of RSV illness and implementing RSV preventive strategies in the years to come.
Supplementary Material
Notes
Acknowledgments. The authors wish to thank Dr Tamuno Alfred for her statistical guidance and Jingying Zhou for her programming support.
Author contributions. K. Y. and J. J. analyzed the data. All authors participated in the analysis design, interpretation of results, and development of this manuscript, had full access to the data, gave final approval of the manuscript before submission, and read and approved the final manuscript.
Data availability. The Premier Data Use Agreement prohibits the sharing of raw data; aggregated findings may be available upon request. NREVSS data may be requested from CDC at nrevss@cdc.gov.
Patient consent. This study does not include factors necessitating patient consent.
Financial support. Pfizer was the funding source in all stages of this study and funded all costs associated with the development and publishing of the present manuscript.
Contributor Information
Jennifer Judy, Pfizer Global Respiratory Vaccines and Antivirals, Real World and Core Evidence, Remote, MD, USA.
Kari Yacisin, Pfizer Vaccines, US Medical Affairs, Collegeville, PA, USA.
Jessica E Atwell, Pfizer Vaccines, Global Medical Affairs, Remote, AZ, USA.
Elizabeth Begier, Pfizer Vaccines, Global Medical Affairs, Dublin, Ireland.
Daniel Curcio, Pfizer Vaccines, Global Medical Affairs, Madrid, Spain.
Alejandro Cané, Pfizer Vaccines, US Medical Affairs, Collegeville, PA, USA.
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
References
- 1. Hamid S, Winn A, Parikh R, et al. Seasonality of respiratory syncytial virus—United States, 2017–2023. MMWR Morb Mortal Wkly Rep 2023; 72:355–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Rose EB, Wheatley A, Langley G, Gerber S, Haynes A. Respiratory syncytial virus seasonality—United States, 2014–2017. MMWR Morb Mortal Wkly Rep 2018; 67:71–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mullins JA, Lamonte AC, Bresee JS, Anderson LJ. Substantial variability in community respiratory syncytial virus season timing. Pediatr Infect Dis J 2003; 22:857–62. [DOI] [PubMed] [Google Scholar]
- 4. McGuiness CB, Boron ML, Saunders B, Edelman L, Kumar VR, Rabon-Stith KM. Respiratory syncytial virus surveillance in the United States, 2007–2012: results from a National Surveillance System. Pediatr Infect Dis J 2014; 33:589–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Olsen SJ, Winn AK, Budd AP, et al. Changes in influenza and other respiratory virus activity during the COVID-19 pandemic—United States, 2020–2021. MMWR Morb Mortal Wkly Rep 2021; 70:1013–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. U.S. Centers for Disease Control and Prevention . Increased interseasonal Respiratory Synyctial Virus (RSV) activity in parts of the Southern United States. 2021. Available at: https://emergency.cdc.gov/han/2021/han00443.asp. Accessed 1 June 2024.
- 7. Hamid S, Winn A, Parikh R, et al. Seasonality of respiratory syncytial virus—United States, 2017–2023. MMWR Morb Mortal Wkly Rep 2023; 72:355–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kampmann B, Madhi SA, Munjal I, et al. Bivalent prefusion F vaccine in pregnancy to prevent RSV illness in infants. N Engl J Med 2023; 388:1451–64. [DOI] [PubMed] [Google Scholar]
- 9. Hammitt LL, Dagan R, Yuan Y, et al. Nirsevimab for prevention of RSV in healthy late-preterm and term infants. N Engl J Med 2022; 386:837–46. [DOI] [PubMed] [Google Scholar]
- 10. Premier Applied Sciences, Premier Inc . Premier Healthcare Database White Paper: Data that informs and performs. 2020. September 7, 2020.
- 11. U.S. Centers for Disease Control and Prevention . The National Respiratory and Enteric Virus Surveillance System (NREVSS). [updated September 28, 2023]. Available at: https://www.cdc.gov/surveillance/nrevss/index.html. Accessed 1 June 2024.
- 12. Grohskopf LA, Alyanak E, Ferdinands JM, et al. Prevention and control of seasonal influenza with vaccines: recommendations of the advisory committee on immunization practices, United States, 2021–22 influenza season. MMWR Recomm Rep 2021; 70:1–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Melgar M, Britton A, Roper LE, et al. Use of Respiratory Syncytial Virus Vaccines in Older Adults: Recommendations of the Advisory Committee on Immunization Practices — United States, 2023. MMWR Morb Mortal Wkly Rep 2023; 72:793–801. doi: 10.15585/mmwr.mm7229a4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Cai W, Tolksdorf K, Hirve S, et al. Evaluation of using ICD-10 code data for respiratory syncytial virus surveillance. Influenza Other Respir Viruses 2020; 14:630–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Makari D, Staat MA, Henrickson KJ, Wu X, Ambrose CS. The underrecognized burden of respiratory syncytial virus among infants presenting to US Emergency Departments. Clin Pediatr (Phila) 2015; 54:594–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kujawski SA, Whitaker M, Ritchey MD, et al. Rates of respiratory syncytial virus (RSV)-associated hospitalization among adults with congestive heart failure-United States, 2015–2017. PLoS One 2022; 17:e0264890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Rozenbaum MH, Judy J, Tran D, Yacisin K, Kurosky SK, Begier E. Low levels of RSV testing among adults hospitalized for lower respiratory tract infection in the United States. Infect Dis Ther 2023; 12:677–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Ramirez J, Carrico R, Wilde A, et al. Diagnosis of respiratory syncytial virus in adults substantially increases when adding sputum, saliva, and serology testing to nasopharyngeal swab RT-PCR. Infect Dis Ther 2023; 12:1593–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Onwuchekwa C, Moreo LM, Menon S, et al. Underascertainment of respiratory syncytial virus infection in adults due to diagnostic testing limitations: a systematic literature review and meta-analysis. J Infect Dis 2023; 228:173–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Onwuchekwa C, Atwell J, Moreo LM, et al. Pediatric RSV diagnostic testing performance: a systematic review and meta-analysis. J Infect Dis 2023; 228:1516–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.