Understanding the incidence of hospitalizations caused by RSV can help to anticipate the potential impacts of upcoming RSV vaccine programs in different sub-populations. In addition, information about the differences in recording ratios between subpopulations may help increase awareness of providers about testing RSV in older adults and may lead to an improvement of RSV surveillance. We estimated a considerable burden of disease in young children (< 5 years of age) and in the elderly (≥ 65 years of age). The estimated incidence of hospitalizations due to RSV was notably higher among patients from low SES ZIP codes. RSV was less likely to be diagnosed in older adults. The percentage of hospitalizations for respiratory illnesses attributable to RSV was highest in infants and lowest among adults 20–64 years of age. With several vaccines and monoclonal antibodies against RSV under active development, these estimates can help to guide estimates of the impact of these interventions in different populations.
Our age-based estimates for the incidence and attributable percent of hospitalizations due to RSV infections are consistent with previous studies [18, 27, 28]. The statistical models we used in our analysis are an extension of the commonly employed time series models that estimate rates of respiratory virus-associated hospitalization [15, 29]. By modifying the model structure, our study provides new insights into the variation in RSV-associated respiratory hospitalizations by SES groups. Our estimates of the incidence of hospitalizations attributable to RSV infections among infants < 1 year of age are lower than the average incidences in published estimates in the more distant past, but comparable to the averages in recent years [18, 27].
Our estimates suggest that the incidence of hospitalizations caused by RSV has been under-recorded in the older adults. The substantial estimated incidence of hospitalizations attributable RSV among those aged 65 and older agrees with earlier observations in a cohort study and time-series study, both of which showed that the incidence of hospitalizations attributable RSV is heavily skewed toward the older adults [30] [27]. This age group should be considered as a potential target population for RSV vaccines due to the potential high case-fatality rate after RSV infection [30, 31]. The number of hospitalizations recorded as being caused by RSV increased over the study period in the patients 65 years of age and older. This trend may reflect changes in testing practices among older adults over time. To understand the actual incidence of hospitalizations attributable RSV in the elderly, more frequent testing for RSV infections is needed.
Our results indicate that children in low-SES communities suffer from a particularly high incidence of RSV-associated hospitalizations. There are a number of potential causes for this disparity, including factors that might influence risk of viral infection, such as family size and the number of contacts; exposure to tobacco smoke and other pollutants; high prevalence of underlying respiratory diseases like asthma and chronic lung disease from prematurity; and duration of breastfeeding [32, 33]. Additionally, decisions to admit patients could be influenced by the family’s SES and may vary based on factors such as the co-morbid illnesses of the patients, the reliability of follow-up, and the practices of individual clinicians.
Our results suggest that about 45% of hospitalizations for respiratory illnesses in infants < 1 year old are attributable to RSV infections (approximately 20 of 1000 infants/year). Monoclonal antibodies against RSV with an extended half-life [34], as well as vaccination of mothers and direct vaccination of infants using live-attenuated vaccines, might help to reduce the incidence of RSV infections [31].
There are several caveats to our results. First, we used hospitalizations due to RSV infections among children < 2 years of age as a proxy for the timing of RSV infections in the entire population. However, there may be differences in the timing of infections among the various age groups. In our preliminary analyses, we tested different time lags between the various age groups, but it did not improve our model fit. Since our model used monthly inpatient data, minor differences in timing between age groups were less likely to be a factor. Second, the cocirculation of other respiratory viruses may confound our estimates. We may overestimate hospitalizations attributable to RSV by not including infections due to respiratory viruses other than influenza and RSV as covariates. However, previous studies indicated that most other respiratory viruses do not have the same timing as epidemics of RSV infections [35–37]. Therefore, the cocirculation of other respiratory viruses should have a relatively small impact on our estimates of RSV-attributable hospitalizations. Third, the estimates of recording ratios rely on the validity of estimates of the incidence of RSV-attributable hospitalizations. Since there is no gold standard of estimates of RSV-attributable hospitalizations, it is hard to validate our estimates. Still, our age-based estimates of recording ratios are similar to those in a previous study conducted by the Centers for Disease Control and Prevention [27].
In conclusion, children in families residing in low-SES areas had the highest incidence of RSV-associated hospitalizations for respiratory illnesses. The incidence of hospitalizations for RSV infections in the older adults is greatly under-recorded. More comprehensive testing for RSV among older adults might help to better define this problem. Vaccines against RSV might provide substantial benefit to young children and to the older adults.