We conducted a cross-sectional study between March and April of 2022. The study universe comprised all school-age children and adolescents (age 6–14 years) with active asthma managed by the public health care system of the region of Castilla y León.

We selected patients by stratified random sampling. First, we established 3 strata defined by the 3 tertiles of the “at risk of poverty rate” (ARPR) distribution by census section in Castilla y León. The ARPR is the percentage of the population aged less than 18 years residing in households with incomes of less than 60% of the median equivalent income (income per consumption unit).21 We obtained the ARPR distribution of each census section from the Household Income Distribution Map developed by the national institute of statistics of Spain, the Instituto Nacional de Estadística (INE),22 which combines data from 2 sources23: (1) fiscal data from the Spanish tax administration agency (Agencia Estatal de Administración Tributaria) and (2) data from the INE on the population residing in the area as of January 1st the year after the period covered by the income tax data. In our study, we used the most recently published data (2018 fiscal year).

To identify patients with asthma in each stratum, we revised the health records of the primary care centres (PCCs) that provide health care services in each census section. We established the geographical boundaries of each census section through the Cartociudad24 application (Ministry of Transport, Mobility and Urban Planning of Spain) and determined how they corresponded to the catchment area of each PCC using the Health Care Organization Guide of the Department of Health of Castilla y León.25 We randomly selected PCCs within each stratum (ARPR) in proportion to the population of children aged 6–14 years of each province in the region (data from the INE as of July 2020). We excluded PCCs that included census sections with different ARPR levels. After, we invited the paediatricians who worked in the selected PCCs to participate in the study. If they did not respond or declined to participate, they were replaced by other paediatricians in the same province and ARPR stratum, who were also selected at random. The collaborating paediatricians identified all children managed in their respective PCCs who met the inclusion criteria, contacted their parents and invited them to participate in the study.

The inclusion criteria were: documentation in the electronic health record, as of 1/1/2022, of an asthma-related condition for which the patient had received health services in 2021 (at the PCC or hospital level) or prescription of any asthma medication (inhaled corticosteroids or leukotriene receptor antagonists) in order to include only cases of currently or recently active asthma. We excluded children whose families had language barriers precluding effective written communication.

After consenting to participation, the main caregivers of the patients completed questionnaires from which data for the variables of interest were retrieved, which concerned:

• 1

  Sociodemographic characteristics. Age, sex, residential setting (urban/rural municipality, defined by the threshold of 20 000 inhabitants), number of members in the household, exposure to tobacco smoke (smoker in the household).

• 2

  Asthma control. We assessed the components of impairment and risk. Impairment, defined by the current level of control (at the time of the survey in 2022), was assessed with the parent proxy version of the CAN questionnaire (Control del Asma en Niños, or asthma control in children).26 This instrument is used to assess asthma control in the past 4 weeks and comprises 9 items rated on a Likert scale ranging from 0 to 4 points (total possible score, 0–36 points), with higher scores indicating poorer asthma control. Poorly controlled asthma was defined as a CAN score of 8 or higher.26 We assessed risk through 3 variables in reference to year 2021: on-demand or urgent visits to PCC or hospital, having received at least 1 course of oral corticosteroids, hospital admission due to asthma.

• 3

  Parent/caregiver asthma-related quality of life. This was assessed with the IFABI-R (revised version of the Impacto Familiar del Asma Bronquial Infantil, or Family Impact of Childhood Bronchial Asthma questionnaire), specifically designed to measure the impact on the quality of life of the primary caregivers of children with asthma in the past 3 months.27 It consists of 15 items rated on a Likert scale exploring the functional, emotional and socio-occupational dimensions of quality of life. The scores range from 1 to 4, and higher scores represent a greater impact of asthma and a poorer quality of life in caregivers. The minimum clinically important difference is of 0.3 points.27

• 4

  Health care-related variables. Use (or not) of daily maintenance treatment, and number of drugs used, physician chiefly responsible for managing asthma in the child (only PC physician, involvement of allergologist), previous allergy testing (in vivo or in vitro), previous lung function testing with spirometry (ever and in 2021), verification by a health care professional of correct inhaler technique (ever and in 2021), having a written asthma action plan for management of exacerbations, having received guidance to avoid tobacco exposure, scheduled follow-up visit in 2021.

• 5

  Other social and family-related aspects: parental educational attainment (having completed at least secondary education or a higher education degree) and whether the parents lived with the child.

We calculated the sample size necessary to estimate the prevalence of poorly controlled asthma, assuming a prevalence of asthma of 7% and a prevalence of poorly controlled asthma of 30%, with a precision of 5% and a confidence level of 95%, and found that we needed 313 patients evenly distributed through the 3 ARPR strata.

The main outcomes under study were asthma control (poor control in 2022 based on the CAN questionnaire and 3 risk indicators in reference to 2021: on demand/urgent visits, hospital admission, use of oral corticosteroids) and caregiver quality of life in 2022. We assessed their association with the ARPR stratum, other sociodemographic variables (urban/rural setting, parental educational attainment, exposure to tobacco smoke at home) and with health care-related variables (management exclusively at PC level) with bivariate tests (χ2 test, Mantel–Haenzsel statistic, Pearson correlation, Student t test and analysis of variance, as applicable) and multivariate regression methods (logistic regression and multiple regression). We adjusted the multivariate models for age, sex, ARPR stratum, urban/rural setting, exclusive PC management, presence of smoker in the household and paternal and maternal educational attainment. The models of asthma control in 2022 and of caregiver quality of life were also adjusted for the history of on demand/urgent visits in 2021. We defined statistical significance as a probability of α error of less than 0.05.

The study was approved by the Clinical Research Ethics Committee of the Hospital Universitario Río Hortega. The parents of the participants received written information about the study and signed an informed consent form.

In this study, we did not find evidence that lower SES is associated with differences in asthma control in children in our population nor with a poorer quality of life in caregivers or poorer quality of the care received. A higher parental educational attainment seems to be associated with better asthma control.

There are several limitations to this study that may have reduced the observed effect of SEI:

• 1

  The identification of the income level based on the place of residence is considered an efficient method with a low risk of bias that generates results in the same direction as measures at the individual level.28,29 However, there is a risk of incorrect classification, which could have diluted the actual impact of SEI.

• 2

  The fact that census sections and the catchment areas of PCCs do not match may have resulted in contamination bias. Although we excluded PCCs with catchment areas across census sections in different ARPR categories, we cannot rule out the possibility of contamination, which would also reduce the apparent effect of SES.

• 3

  The census section income data correspond to 2018, but we collected data on asthma control variables in 2022. If there were changes in the income distribution, they would have hindered the detection of an association between income and asthma control.

• 4

  The COVID pandemic affected both the health care services provided to children with asthma as well as asthma control at the global level,30 which may have also affected the association between SEI and asthma control.

• 5

  Spain is very diverse in certain socioeconomic and health care aspects, so our results cannot be generalised without carrying out similar studies in other regions. Our region is sparsely urbanised, there are no large cities and no large economically depressed areas, which are the main targets of studies conducted in countries such as the United States. Thus, the impact of urban poverty, which may be substantial, is not adequately represented in our sample.

• 6

  We used instruments (CAN, IFABI-R) validated in Spanish, so we excluded families with language barriers. As a result, our study is not representative of the immigrant population with language barriers, which may be particularly socioeconomically vulnerable.

The Beveridge health care system model (National Health Service model, such as the systems found in Spain, United Kingdom, Denmark, Portugal, Sweden and other European countries) aims at providing universal coverage for health care services in order to decrease the impact of SEI on health. It is important to confirm that this objective is achieved, but in Spain the evidence on socioeconomic health determinants is scarce, especially in relation child health.31 As for asthma, most of the evidence comes from countries with other types of health care systems (especially the United States) and studies in adults. A recent metanalysis (61 studies, most conducted in the USA, only 2 including paediatric participants), found that a lower SES was associated with adverse events during asthma follow-up (emergency visits, hospital admission, readmission, need of ventilation).32 However, there is also recent evidence in Europe of an adverse impact of SEI on asthma control in children and adolescents.33 Our findings were not in agreement with those results, which could be due to several reasons:

• 1

  Some studies found an association between lower SES and less frequent use of asthma medication,13 which could contribute to poorer asthma control. In our study, however, we found a higher frequency of polypharmacy in patients who resided in low-income areas.

• 2

  The impact of SEI on asthma-related outcomes is chiefly mediated by factors like social and family support, caregiver stress and self-management education, while factors such as access to the health care system, individual behaviour and lifestyle habits are less important.18,34 This may explain why standardised management has been found not to eliminate the effect of SEI in asthma control in the United States,14 while interventions aimed at reducing social risks do achieve this goal.35 Social and family-related variables and stress may differ significantly in our population compared to the populations of these other studies.

• 3

  Particularities of the Spanish health care system, with a high number of paediatricians on staff in the PC network, which is not the case of other European countries.36 Furthermore, while other studies found that patients of lower SES were less frequently managed by specialists,37 the opposite was observed in our sample: exclusive follow-up in the PC system was less frequent with decreasing SES. We found that management exclusively at the PC level was associated with lower adherence to certain care quality standards, but at the same time was associated with better asthma control and caregiver quality of life, which we attributed to the lower severity of asthma cases managed exclusively at the PC level.

In conclusion, we found that in the population of our region, SEI was not associated with relevant outcomes of asthma in children, nor with differences in health care quality. We consider that this evinces the effectiveness of the National Health System in reducing the health care disparities that may result from SEI. Nevertheless, the quality of health care should improve, especially at the PC level.