Data sources
Data were extracted from the French national hospital discharge database (Programme de Médicalisation des Systèmes d’Information – PMSI), covering the whole population [13]. PMSI provides medico-administrative information on secondary and tertiary care from the public and private sectors as described previously [14]. Study was carried out in compliance with the French regulatory [15].
-
a)
Hospitalization data
All RSV-associated hospital stays in general medicine, surgery and obstetric medical facilities (PMSI-MSO) as well as all-cause hospitalization from 2010 through 2018 in children aged < 5 years were selected for this analysis. RSV-associated hospitalization were identified using the following International Classification of Diseases 10th revision (ICD-10) codes as the primary diagnosis (PD): J210 (acute bronchiolitis due to RSV), J219 (acute bronchiolitis, unspecified), J121 (pneumonia due to RSV), J205 (bronchitis due to RSV), J45 (asthma) and R062 (wheezing).
Hospital admissions with an ICD-10 code of bronchiolitis (J210 or J219 ICD-10 codes) recorded as an associated diagnosis (AD) were also selected when the primary diagnosis was a “disorder of the respiratory system” (J00-J99 ICD-10 codes) or “certain conditions originating in the perinatal period” (P00-P96 ICD-10 codes), assuming that RSV-associated bronchiolitis contributed to the hospitalization despite not being listed as primary diagnosis.
-
b)
Demographic data
Monthly births recorded and published by the National Institute of Statistics and Economic Studies (INSEE) were used to estimate the incidence of RSV in France for the 2010–2018 period [16].
Statistical analyses
A descriptive analysis of RSV-associated hospitalizations during RSV seasons (defined from October to March the following year) was carried out over the period 2010–2018 [17].
Analyses were stratified by age group, defined by chronological ages (< 1 year old vs. ≥1 year old) and further stratified by month categories and by gestational weeks at birth (weeks’ gestational age, wGA).
wGA was determined based on mother’s information on the number of weeks of amenorrhea, directly available in PMSI database. Data on mother’s amenorrhea was not consistently available, especially for records of children born before 2009. Full term birth was defined as birth after 36 weeks of amenorrhea, otherwise it was considered as preterm birth, for which subcategories were defined: extremely preterm birth (< 29 weeks), very preterm birth (29–32 weeks) and moderate preterm birth (33–35 weeks). The stratification was based on the palivizumab recommendations of the healthcare authorities and the society of neonatology in France [7].
-
c)
Epidemiological analysis
RSV-associated hospitalization incidence was estimated by the number of RSV-associated hospitalizations per 1000 person-months for each RSV season. Total time of exposure was calculated from demographic data, corresponding to the number of months during which children in specific age subgroups were at risk for RSV, during the season. Incidence was also estimated for the respiratory year, defined from July to June the following year to ensure that an entire RSV season was covered.
We assessed if the month of birth has an impact on the risk of RSV hospitalization. For children born during a given month, the number of RSV-associated hospitalizations was computed from birth to the end of their first RSV season. The RSV-associated hospitalization risk across the 8 RSV seasons was determined by dividing the total number of RSV-associated hospitalisations from the total number of children born during the same month. For this risk estimation, a more restrictive definition of RSV-associated hospitalizations was considered excluding J45 (asthma) and R062 (wheezing).
Children characteristics were described considering the age, the gender and the known risk factors for RSV-associated hospitalization: presence or absence of congenital heart defects (PD, related diagnosis (RD) or AD ICD-10 Q20-Q26 in hospital discharge), bronchopulmonary dysplasia (PD, RD or AD ICD-10 P27.1), Down syndrome (PD, RD or AD ICD-10 Q90), cystic fibrosis with pulmonary manifestations (PD, RD or AD ICD-10 Q90, E84).
All-cause hospital admissions occurring within 3 months following the initial RSV-associated hospitalization were identified. The number and proportion of patients affected by re-hospitalization and the average number of re-hospitalizations per patient were estimated.
-
d)
Economic analysis
Direct cost of RSV burden was calculated based on the resources used during RSV-associated hospitalizations recorded in PMSI database. Total cost and average cost per stay per patient were calculated by RSV season from a collective perspective, considering the French national health insurance reimbursement, medical professionals costs and running costs of the medical unit (e.g. cleaning and laundry), public complementary and private insurance and out-of-pocket charges.
These costs were calculated for all RSV seasons according to the DRG (Diagnosis Related Group) based payment system for reimbursement of acute care (MSO). The costs were adjusted to 2018 euro value by applying the price index of health products in mainland France [16].