Study population
This prospective cohort study was performed on a convenience sample of 69 family pediatricians (FPs) working for the Italian National Health System in Lombardy, which provides free medical care to all children in Italy. The FPs were enrolled by the Italian Primary Care Pediatrics Society (SicuPP) and the members of the Italian Pediatric Society (Lombardy Section). All FPs joined the study as volunteers.
The recruitment of FPs took place from September 1 to September 20, 2019. At the end of September 2019, all the recruited FPs attended a meeting to discuss the study procedure. Each FP was requested to recruit up to 10 children aged 0–5 years between October 2019 and January 2020. Healthy Caucasian preschool children of both sexes from 1 month to 5 years were eligible for the study. Children with conditions or diseases predisposing them to recurrent respiratory infections such as prematurity (gestational age < 37 weeks), congenital abnormalities of the respiratory tract, congenital or acquired immunodeficiency including cystic fibrosis, cardiovascular, renal, and hematological diseases, and Down syndrome were excluded.
For the purposes of the study, the recruited population was divided into 6 groups (children aged 0, 1, 2, 3, 4, and 5 years) according to age. Infants were defined as children aged between 1 and 12 months.
Study protocol
At recruitment, each FP completed a case report form (CRF) which included socio-demographic information about each child. Parents of enrolled children were instructed to contact the FP if their child developed a fever (defined as an axillary temperature > 38 °C) or became ill [15]. The FP carried out a full clinical examination and, if an ARTI was diagnosed, a CRF was completed with the specific diagnosis in accordance with the international and statistical classification of diseases and related health problems [16]. Based on the site infection, a common cold, tonsillitis, pharyngitis, and otitis media were considered upper respiratory tract infections (URTIs), whereas infections below the epiglottis such as laryngitis, bronchiolitis, acute bronchitis, and pneumonia were considered lower respiratory tract infection (LRTIs) [17, 18]. The diagnostic criteria for each disease were discussed in an in-presence meeting and agreed upon by FPs prior to enrolling the first child.
Upon enrollment, FPs were instructed to perform follow-up phone calls to parents every 15 days to remind them of the study procedures and to monitor participants’ adherence to the protocol. Each child was followed from the day of recruitment up to March 30, 2021 (end of the study). The flow of patients during the study is described in the Results Section.
A single episode was taken into consideration if at least 2 medical visits within 14 days reported the same ARTIs diagnosis during continuous respiratory symptoms.
For this analysis, the pre-pandemic period of COVID-19 was defined as October 1st, 2019, up to February 28th, 2020, and the pandemic period as March 1, 2020 up to March 30th, 2021 (end of the study), based on national epidemiological data characterizing COVID-19 as a pandemic [19]. Italy was placed on a nationwide lockdown from March 9th, to May 3rd, 2020.
The winter period is defined as November to February.
During the COVID-19 pandemic, children with possible COVID-19 signs and symptoms were swabbed for SARS-COV-2 infection, as recommended by the National Health Authorities [20]. If the child was tested negative for COVID-19 and had symptoms of ARTIs the full clinical examination was performed and CRF was completed, and the diagnosis was recorded. Confirmation of COVID-19 was defined as the detection of SARS-CoV-2 in nasopharyngeal samples through RT–PCR.
The study was conducted in accordance with the Declaration of Helsinki and was approved in September 2019 by the Ethical Committee of the Fondazione IRCCS Ca′ Granda Ospedale Maggiore Policlinico, Milan, Italy (752–2019). Each child’s parents gave their written consent. Physicians did not receive any payment incentives for the study. The study procedures were free for all patients.
Outcomes
The primary outcome was the estimation of the incidence of ARTIs in children aged 0–5 years (see “Statistical analysis” for further details). The secondary outcomes were the estimation of the incidence of ARTIs in different age groups, and the effect of the COVID-19 pandemic on these estimates.
Statistical analysis
The sample size was calculated as follows. By recruiting 70 children for each of the 6 annual age groups (420 children in total) and for each of the strata identified for each month and each age group, we expected an average of 2100 children x days of attendance. Assuming that 5 to 50% of the children are afflicted by ARTIs across the strata, we estimated an incidence of 5 cases/100 children × 30 days (95% CI: 1.2–13.6) to 50 cases/100 children × 30 days (95% CI: 34.8–69.5) within each stratum (3).
The descriptive analysis of the sample of recruited children was carried out using the following criteria: a) for the numerical variables, the median value and the interquartile range, that is, the difference between the values of the third and first quartile; and b) for categorical variables, the absolute and relative frequencies for each of their values. The incidence rate was calculated as the number of events observed (episodes reported by pediatricians for each of the age groups at the time of the episode, in completed years, and for each of the months analyzed), divided by the time at risk of event for each observation period. Each child could experience more than one episode in the same calendar month. The ratios obtained in this manner were normalized as the number of episodes reported per 100 children × 30 days, and this analysis was repeated for episodes of infection affecting both the upper and lower respiratory tracts. The Poisson distribution was used to calculate the 95% confidence intervals (CIs) for incidence rates [21, 22]. The weighted average of the incidence rate in the two winter periods (November, December, January, and February of the year 2019–20 versus the same period the following year 2020–21) were then calculated, as well as the 95% confidence intervals for incidence rate differences [23].