https://orcid.org/0000-0002-8949-1568
Abstract
Bronchiolitis is the commonest cause of respiratory related hospital admissions in young children. This study aimed to describe temporal trends in bronchiolitis admissions for children under 2 years of age in Scotland by patient characteristics, socioeconomic deprivation, and duration of admission.
The national hospital admissions database for Scotland was used to extract data on all bronchiolitis admissions (International Classification of Disease, Tenth Revision, code J21) in children <2 years of age from 2001 to 2016. Deprivation quintiles were classified using the 2011 Scottish Index of Multiple Deprivation.
Over the 15-year study period, admission rates for children under 2 years old increased 2.20-fold (95% confidence interval [CI], 1.4–3.6-fold) from 17.2 (15.9–18.5) to 37.7 (37.4–38.1) admissions per 1000 children per year. Admissions peaked in infants aged 1 month, and in those born in the 3 months preceding the peak bronchiolitis month—September, October, and November. Admissions from the most-deprived quintile had the highest overall rate of admission, at 40.5 per 1000 children per year (95% CI, 39.5–41.5) compared with the least-deprived quintile, at 23.0 admissions per 1000 children per year (22.1–23.9). The most-deprived quintile had the greatest increase in admissions over time, whereas the least-deprived quintile had the lowest increase. Zero-day admissions, defined as admission and discharge within the same calendar date, increased 5.3-fold (5.1–5.5) over the study period, with the highest increase in patients in the most-deprived quintile.
This study provides baseline epidemiological data to aid policy makers in the strategic planning of preventative interventions. With the majority of bronchiolitis caused by respiratory syncytial virus (RSV), and several RSV vaccines and monoclonal antibodies currently in clinical trials, understanding national trends in bronchiolitis admissions is an important proxy for determining potential RSV vaccination strategies.
Bronchiolitis is the most common cause of respiratory related hospital admissions for children <2 years old, with a prospective study in England suggesting 3.7% of infants are admitted within their first year of life [1–5]. In the United Kingdom, several trends in bronchiolitis admissions in England have been described [2, 4, 6]. First, infants <6 months of age are the most vulnerable to bronchiolitis and have the highest risk of severe disease leading to admission [1, 7]. Second, children born just before or during the winter season are at a higher risk of being admitted for bronchiolitis than those born in the summer or spring months [1, 8]. Third, the rate of bronchiolitis admission for infants <1 year old was estimated to increase by an average of 1.8% per year between 2004 and 2011, with suggestions of an accelerated increase of 5.2% per year between 2011 to 2016 [2, 5]. Finally, studies suggest that children from lower socioeconomic groups have a 30%–37% higher risk for both standard hospitalizations for acute lower respiratory tract infections and pediatric intensive care unit admissions for bronchiolitis [6, 9, 10]. However, bronchiolitis admissions in Scotland have not been described in such detail [10, 11].
Respiratory syncytial virus (RSV) is a major cause of bronchiolitis. In England, it has been estimated that RSV is the cause of approximately 78% (95% confidence interval [CI], 75%–83%) of bronchiolitis admissions in children <5 years of age [1, 10]. Although much of the severe bronchiolitis due to RSV infection can be prevented by treatment with the monoclonal antibody palivizumab, this is recommended for use only in certain high-risk patients, owing to its limited cost-effectiveness [12]. As such, for most of the population, there is no primary prevention targeted at RSV infection. However, several RSV vaccines and monoclonal antibodies are currently in clinical development and evaluation [10, 13]. Therefore, understanding the underlying national trends in admissions caused by RSV—particularly bronchiolitis—is crucial in determining potential future vaccine strategy and monitoring the impact of a potential future vaccine.
The current study aimed to describe bronchiolitis hospital admissions in children <2 years old in Scotland from 2001 to 2016, by admission and patient characteristics. The association between admission rates and socioeconomic deprivation was also considered, to further highlight target populations for preventive measures and interventions.
METHODS
Study Data
Data for this study were sourced from the national hospital admissions database for Scotland (SMR01), which routinely captures information on all general, acute inpatient and day case admissions to National Health Service (NHS) hospitals in Scotland. Data were extracted on calendar week of admission, sex, age (in months for infants up to 12 months old, after which age was recorded as age in years), birth month, Scottish Index of Multiple Deprivation (SIMD) quintile, and length of hospital stay. Length of hospital stay was calculated using admission and discharge dates, with zero-day admissions defined as an admission to a ward (either general, acute or day-bed unit) and discharge from hospital on the same calendar date. Accident and emergency presentations that did not result in an admission were not included in the study
SMR01 uses International Classification of Disease, Tenth Revision, codes to record diagnoses. Patients can have 1 primary diagnosis and up to 6 other diagnoses. The inclusion criteria for subjects were as follows: any diagnosis of bronchiolitis (as either the primary diagnosis or otherwise) as coded under International Classification of Disease, Tenth Revision, J21; age <24 months at admission; admission to any Scottish hospital between the beginning of calendar week 21 in 2001 to calendar week 20 in 2016. Patients who were readmitted to the hospital for bronchiolitis on the same day as discharge were classified as a single admission.
The 2011 SIMD definition of deprivation was used to classify subjects into deprivation quintiles, wherein 1 indicated the most-deprived and 5 the least-deprived population. The full SIMD criteria are described in the Supplementary Material but included factors on health (including low birth weight, emergency stays in hospital, comparative illness factor, and standardized mortality ratio), geographic access to services (including average drive and public transport time to general practice services), and household crowding [14]. SIMD does not include smoking prevalence, breastfeeding practices, or vaccination rates.
Data Analysis
Admissions fitting the inclusion criteria were described by calendar week and year of admission, sex, age, birth month and SIMD. Calendar weeks were defined as ISO week date standard (ISO-8601). Because bronchiolitis admissions follow a seasonal trend, peaking in the winter months, an epidemiological year was defined as the time between the beginning of week 21 of 1 year and the end of week 20 of the following year (eg, children admitted between week 21 of 2001 and week 20 of 2002 were categorized in the 2001–2002 year).
Admission rate per year and SIMD were calculated using imputed midwinter populations [15]. Denominator data on the Scottish population were derived from publicly accessible midyear estimates released by the National Records of Scotland which were stratified by sex, age in years, and SIMD. Denominator population estimates for SIMD were only available from 2011 onward [15]. The overall change in admission rate was calculated by comparing rates from the 2015–2016 and the 2001–2002 years. Average annual percentage increases in admission rates and 95% CIs were calculated overall and for each SIMD quintile.
Length of stay (in days from admission to discharge) was calculated for all admissions and compared over each year and by SIMD. Data were cleaned using STATA software, version 14 and analyzed using IBM SPSS Statistics 24 software.
RESULTS
Between 2001 and 2016, there were a total of 43 514 bronchiolitis admissions in children <2 years old. Of these, 25 904 patients (59.6%) were males (male-female ratio, 1.4 [95% CI, 1.4-1.4]). Bronchiolitis admission rates increased 2.2-fold (95% CI, 1.4–3.6) between 2001–2002 and 2015–2016, from 17.2 (95% CI, 15.9–18.5) to 37.7 (37.4–38.1) admissions per 1000 children per year. This equated to a yearly increase in admission rates of 1.3% (95% CI, 1.2%–1.4%) (Figure 1). Changes in admissions rates were not uniform among the 14 health boards, with NHS Ayrshire and Arran having the largest yearly increase, 3.6% (95%. CI 3.4–3.9), and NHS Grampian the smallest, at 0.3% (95% CI, .2–.3) (Supplementary Figure 1).
Weekly number of bronchiolitis admissions (solid line) and annual rate of bronchiolitis admissions per 1000 children (dotted line) in children <2 years in Scotland from 2001–2002 to 2015–2016.
Most admissions were in infants <3 months of age (12 660 of 43 514 [29.1%]) with those aged 3 to <6 months, 6 to <9 months, 9 to <12 months, and ≥12 months, making up 24.9% (10 817 of 43 514), 18.6% (2702 of 4514), 12.4% (1801 of 43 514), and 15.0% (6526 of 4514) of the overall cohort, respectively (Supplementary Table 1). Infants aged 1 month accounted for 12.6% (5464 of 43 514) of all admissions in children <2 years old. After age 1 month, the number of admissions decreased with increasing age (Supplementary Figure 2).
On average, admissions peaked in the month of December across the 15 years of study (Supplementary Figure 3). Infants born in the 3 months leading up to the peak bronchiolitis month—September, October, or November—had the highest number of admissions, making up 10.6% (3937 of 36 988), 11.4% (4215 of 36 988), and 11.7% (4327 of 36 988) of all admissions in children <1 year old (Supplementary Figure 3).
Children from the most-deprived backgrounds (SIMD quintile 1) contributed the highest proportion of admission, at 31.0% (13 473 of 43 128), compared with 24.5% of the general population classified in this quintile where data were available (Supplementary Table 1 and Supplementary Material). SIMD quintiles 2, 3, 4, and 5 made up 22.0% (9570 of 43 128), 17.1% (7457 of 43 128), 15.9% (6913 of 43 128), and 13.3% (5805 of 43 128) of the overall study cohort and 20.4%, 18.3%, 18.2% and 18.4% of the general population, respectively (Supplementary Table 1 and Supplementary Material). Rates of admission by deprivation could only be calculated for the years 2011–2012 to 2015–2016, which translated to 42% (18 537 of 43 662) of the overall cohort.
The national average rate of admission across the 5 years was 32.2 admissions per 1000 children per year (95% CI, 31.7–32.6). The most-deprived quintile had the highest overall rate of admission at 40.5 per 1000 children per year (95% CI, 39.5–41.5), and the least-deprived quintile had the lowest rate, at 23.0 per 1000 children per year (22.1–23.9) (Figure 2). Over the 5-year period for which admission rates could be calculated, the most-deprived quintile had seen the steepest increase in admission rates, with an average increase of 4.7% (95% CI, 4.6%–5.0%) per year from an admission rate (per 1000 children per year) of 31.5 (29.5–33.6) in 2011–2012 to 50.3 (47.7–52.8) in 2015–2016. In comparison, the least-deprived quintile had an increase in admission rates of 2.2% (95% CI, 2.1%–2.3%) over the same period, from 17.8 (16.0–19.5) to 26.9 (24.7–29.1) (Figure 2 and Supplementary Table 2)
Bronchiolitis admission rates per 1000 children per year in Scotland by Scottish Index of Multiple Deprivation (SIMD) quintile from 2011–2012 to 2015–2016. Most-deprived quintile, 1; least-deprived quintile, 5.
Since 2001–2002 to 2015–2016, there was a 5.3-fold (95% CI, 5.1–5.5-fold) increase in zero-day admissions ,from 319 to 1695 admissions. with a smaller increase of 2.4-fold (2.3–2.4) in 1-day admissions. from 442 to 1068 admissions (Supplementary Figure 4). The proportion of zero-day admissions rose from 18% (319 of 1801) to 40% (1695 of 4259) of all admissions from 2001–2002 to 2015–2016, whereas the proportion of 1-day admissions stayed the same, at 25% (442 of 1801 in 2001–2002 and 1068 of 4259 in 2015–2016) and the proportion of >1-day admissions decreased, from 58% (1040 of 1801) to 35% (1496 of 4259) over the same period (Supplementary Figure 4). Trends in the number of zero-day admission by SIMD suggest that across the 15 years, SIMD quintile 1 had seen the largest increase in admissions, with an estimated 5.5-fold increase, from 102 in 2001–2002 to 571 in 2015–2016, whereas SIMD quintile 5 increased 3.9-fold in the same period (fro 50 to 195 cases) (Supplementary Figure 5A–5C).
DISCUSSION
Between 2001 to 2016, the rate of bronchiolitis hospital admissions in Scotland rose 2.2-fold (95% CI, 1.4–3.6-fold) from 17.2 (95% CI, 15.9 – 18.5) to 37.7 (37.4 – 38.1) admissions per 1000 children per year in children <2 years old. The high increase seems to have been primarily due to an increase in short-stay admissions. Almost a third of admissions in the cohort were in infants aged ≤3 months old, with a peak in 1-month-old infants. Seasonality of bronchiolitis admissions was consistent across the 15 years of study, with December the peak admission month and infants born within the preceding 3 months most likely to be admitted. The highest admission rates were among the most-deprived patients, who also accounted for the highest proportion of zero-day admissions.
The current study found an average yearly increase in bronchiolitis hospital admissions of 1.3% (95% CI, 1.2%–1.4%,) for children <2 years old, with an overall increase in admission rate of 2.2-fold (1.4–3.6-fold) across the 15 years. Trends in admission were not the same across health boards, which may reflect differing characteristics, such as population density, resource allocation, and referral practices (whereby smaller boards are more likely to send patients to larger secondary or tertiary care centers in another region) [16]. Prevalences of prematurity and low birth weight, key risk factors for bronchiolitis, have been relatively stable in Scotland during the study period [17, 18]. However, general pediatric emergency admissions rates for the same age group have increased during the study period by 3.1% per year [3]. This increase corresponds with English data showing a rise in bronchiolitis admission in children <2 years of 1.6%–1.8% per year between 2004 and 2011, and 5.2% per year between 2011 to 2016 [2, 5]. Global trends in bronchiolitis admission rates vary greatly, however, with data from Portugal and Spain reporting increases over time, Australia reporting relatively stable rates, and the United States reporting decreasing trends; these variations are likely to reflect differences in health policies and healthcare systems [4, 19–22].
Our study suggests that the increase in admissions was largely driven by a 5.3-fold (95% CI, 5.1–5.5-fold) increase in zero-day admissions, reflecting a wider UK trend in increasing zero-day admissions over the past 2 decades [3, 23–25]. The Scottish Intercollegiate Guidelines Network (SIGN) updated diagnostic criteria for bronchiolitis in 2015, the penultimate year of our study period, and therefore changes in diagnostic criteria cannot account for the changes over time seen in our study [26]. Factors contributing to an increase in bronchiolitis admissions may be similar to the wider factors contributing to a general increase in all-cause pediatric admissions in Scotland: daycare attendance at a younger age increasing early exposure to pathogens, changes in parental expectations of treatment, and the decrease in availability of out-of-hours primary care owing to changes in general practice contracts [3, 25, 27]. Coupled with increased pressures on emergency departments to see patients within 4 hours and the focus on routine laboratory testing and oxygen saturation monitoring, this may explain the increase in short-stay admissions [3, 25, 28, 29]. Although we did not directly measure clinical outcomes, such as intensive care unit admission or use of additional oxygen supplementation, the relatively unchanged number of >1-day admissions across the 15 years does not suggest an increase in disease severity over the study period.
Our study suggests that the risk for hospital admission was highest for infants >3 months old, and especially those 1 month old, with data from England corroborating this finding [30]. Because RSV has been estimated to cause 78% of bronchiolitis cases in children <5 years old [1, 4, 31], and because there are a number of potential vaccine candidates in development, a universal RSV immunization scheme could potentially significantly reduce the burden of bronchiolitis in secondary care [1, 6, 8, 32]. We found that bronchiolitis admissions were highest in children born in the months preceding the RSV season, in line with findings from other studies in the United Kingdom and temperate countries [8, 33, 34]. Cost-effectiveness models based on English data suggesting the most cost-effective RSV vaccine strategy would target infants born between the months of August and January [8]. Therefore, a similar approach could potentially be adopted for Scotland.
The consistently higher rate of admissions in the least-deprived SIMD quintile is consistent with recent studies showing a strong correlation between deprivation in both general pediatric acute lower respiratory tract infections and emergency admissions [3, 4]. A recent study reported an incidence risk ratio of 1.3 (95% CI, 1.2–1.4) for hospitalization between the most- and least-deprived deciles for bronchiolitis in infants <1 year old in Scotland [4]. Our study found that over 15 years there was a larger increase in the number of zero-day admissions in the most-deprived than in the least-deprived quintile (5.6-fold vs 3.9-fold increase, respectively). These results are also in line with wider Scottish and English admission trends, with data on general pediatric emergency admissions suggesting an 80% difference in admission rates between quintiles in 2013 [3, 6, 23, 24]. These differences may be partly explained by the distribution of bronchiolitis risk factors, such as having a household member who smokes, low birth weight, and lack of exclusive breast feeding.
A Scotland-wide survey in 2010 suggested that 65% of infants in the most-deprived quintile had received some form of breast milk during their first 12 weeks of life, compared with 86% in the least-deprived group [18, 35–37]. Data revision in 2017 showed little change in breastfeeding habits in the least-deprived group, but in the most-deprived group breastfeeding had risen by 5% [38]. The Scottish Government’s Growing Up in Scotland cohort study and NHS Information Services Division data found that infants in the most-deprived quintile had a 43% risk of being exposed to maternal or household smoking during pregnancy, compared with only 9% in the least-deprived quintile [17, 18]. However, these data also show that the nationwide maternal smoking has fallen from 20.2% in 2007 to 14.0% in 2016 and suggests that the largest fall has been in the most-deprived quintile. As such, although these risk factors have a significant role in bronchiolitis infections and maintaining the inequality gap, other factors are likely to be driving the increase over time [38]. Areas for further consideration may include the impact of general childhood poverty, which has been increasing since 2010 and currently affects 18% of the UK population, and housing conditions, including household overcrowding [39].
The main strength of this study was the ability to track disease trends across a large population over 15 years using routine hospital data, with collection methods and diagnostic criteria seeing little change over this period. To our knowledge, this was the first study in Scotland to investigate the trends in bronchiolitis admissions over time in such detail, including the correlation between deprivation and trends in admission lengths [3, 4, 11, 40]. However, there were also limitations to this study. SMR01 uses retrospectively collected data, meaning that the quality of collected information was dependent on the quality of physician reports and diagnostic coding. Our extract was not linked to maternal or birth records; therefore, we did not assess comorbid conditions or prematurity, because these are likely underreported in SMR01 [41]. In addition, the study did not look at readmissions unless they occurred on the same day as discharge. Finally, because this data set was limited to hospital inpatient admissions only, emergency department visits and the primary care burden could not be analyzed.
In summary, the current study has demonstrated a significant increase in bronchiolitis admissions in Scotland over time, particularly in the most-deprived populations, largely driven by an increase in zero-day admissions. Future studies could investigate the main drivers behind the increase in zero-day admissions and how these could potentially be prevented. Because the greatest rise in admissions has been in the most-deprived quintile, interventions should be more strongly targeted to this population. These could include increasing the accessibility of primary care and home care services, as well as further targeting of prevention strategies, such as smoking cessation programs, and the promotion of breastfeeding practices [3, 18, 23]. This study provides detailed baseline epidemiological data to aid policy makers in the strategic planning of preventative interventions and has added weight to the argument for introducing a routine RSV immunization program.
Supplementary Data
Supplementary materials are available at The Journal of 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.
Notes
RESCEU investigators. Rachel M. Reeves, You Li, Harry Campbell and Harish Nair (University of Edinburgh, Scotland); Maarten van Wijhe, Thea Kølsen Fischer, Lone Simonsen, and Ramona Trebbien (Statens Serum Institut, Denmark); Sabine Tong, Mathieu Bangert, and Clarisse Demont (Sanofi Pasteur); Toni Lehtonen (Finnish Institute for Health and Welfare, Turku University Hospital, Finland); Terho Heikkinen (Turku University Hospital, Finland); Anne Teirlinck, Michiel van Boven, Wim van der Hoek, Nicoline van der Maas, and Adam Meijer (National Institute for Public Health and the Environment, the Netherlands); Liliana Vazquez Fernandez, Håkon Bøas, Terese Bekkevold, and Elmira Flem (Norwegian Institute of Public Health); Luca Stona, Irene Speltra, and Carlo Giaquinto (Penta, Italy); Arnaud Cheret (Janssen); Amanda Leach and Sonia Stoszek (GlaxoSmithKline); Philippe Beutels (University of Antwerp, Belgium); Louis Bont (University Medical Centre Utrecht, the Netherlands); Andrew Pollard (University of Oxford, United Kingdom); Peter Openshaw (Imperial College, United Kingdom); Michael Abram (AstraZeneca); Kena Swanson (Pfizer); Brian Rosen (Novavax); and Eva Molero (Synapse Research Management Partners).
Financial support. This study was supported by the Respiratory Syncytial Virus Consortium in Europe (RESCEU), which has received funding from the Innovative Medicines Initiative 2 Joint Undertaking (grant 116019); the Joint Undertaking receives support from the EU’s Horizon 2020 research and innovation program and the European Federation of Pharmaceutical Industries and Associations.
Supplement sponsorship. This supplement is sponsored by RESCEU (REspiratory Syncytial Virus Consortium in EUrope).
Potential conflicts of interest. R. M. R. was supported by grants from the Innovative Medicines Initiative, during the conduct of the study. H. N. reports grants from the Innovative Medicines Initiative, during the conduct of the study; and grants and nonfinancial support from the World Health Organization, grants and personal fees from Sanofi, personal fees from AbbVie and Janssen, and grants from the Bill & Melinda Gates Foundation, outside the submitted work. H. C. reports grants from the Innovative Medicines Initiative, during the conduct of the study; and grants, personal fees, and nonfinancial support from the World Health Organization, grants and personal fees from Sanofi, and grants from the Bill & Melinda Gates Foundation and from the UK National Institute for Health Research, outside the submitted work. A. C. reports no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
Author notes
Members of the study group are listed at the end of the text.
