OUP user menu

Perceived health status and use of healthcare services among children and adolescents

Silvina Berra, Carme Borrell, Luis Rajmil, Maria-Dolors Estrada, Maica Rodríguez, Anne W. Riley, Christopher B. Forrest, Barbara Starfield
DOI: http://dx.doi.org/10.1093/eurpub/ckl055 405-414 First published online: 27 April 2006

Abstract

Background: The aim of the present study was to analyse the use of healthcare services according to health status in a population of children and adolescents, taking into account family socio-demographic characteristics and characteristics of the proxy respondent. Methods: A total of 836 interviews of proxy respondents for children aged 5–14 years from the Barcelona Health Interview Survey carried out in 2000 were included. Dependent variables were visits to a healthcare professional, visits to the emergency room, and hospitalization. Independent variables were: report of medical conditions, health status of the child measured by the Child Health and Illness Profile-Child Edition, Parent Report Form (CHIP-CE/PRF), the educational level of the head of household, social class, child's healthcare coverage, and proxy-related variables [mental health status by means of the General Health Questionnaire-12 items version (GHQ-12), and other]. Logistic regression analysis was used to estimate prevalence ratio (PR) to compare the use of healthcare services among different categories of independent variables. Results: Children having worse health status were more likely to have visited a healthcare professional [PR = 1.68; 95% confidence interval (95% CI) = 1.09–3.83], whereas children with a reported medical condition were more likely to have made a visit to the emergency service (PR = 1.47; 95% CI = 1.27–2.55) and were hospitalized more frequently (PR = 2.50; 95% CI = 1.12–5.57). Higher likelihood of visits to the emergency room was associated with children having both public and private coverage and a proxy respondent scoring 3 or higher on the GHQ-12. Conclusions: Use of healthcare services differed by health needs but not by social class. Double healthcare coverage and mental distress of the proxy respondent influenced the use of emergency services.

  • child health care
  • health services
  • health status
  • mental health

Introduction

The use of healthcare services varies according to factors such as age, gender, and health status, as well as perceptions of susceptibility, severity, and costs and benefits of seeking care. Socio-economic characteristics of the population, including income level, social class, or educational level may also have an impact on the use of services; this is also true of structural and functional aspects of the healthcare system, especially coverage, supply, access, and the quality of services provided.1

When considering children and adolescents, factors that have an impact on the utilization of healthcare services include: perceived health status; the mother's pattern of use; and certain family characteristics, including number of children in the household and the presence of conflicts in the form of physical violence among family members.2

The assessment of perceived health status, as well as health-related quality of life,3 is useful for detecting groups with greater needs, and identifying inequities in health and in the use of healthcare services. When dealing with child populations, information is usually gathered through proxy respondents (generally the mother) who are, moreover, those who decide to seek healthcare.2

In Spain, the state-run health service provides universal coverage, with free access to services, including primary, specialized, emergency, and hospital healthcare. However, 10% of people pay for private healthcare coverage in addition to their public coverage. Some socio-economic differences in the use of healthcare services might persist, however, as has been shown in the adult population.4 The Barcelona Health Interview Survey from 2000 (BHIS)5 included the Child Health and Illness Profile-Child Edition, Parent Report Form (CHIP-CE/PRF) as a perceived health status measure for children 5–14 years old, and also collected information regarding morbidity, selected health conditions, and families' socio-economic characteristics, as well as those characteristics of the proxy respondent that could potentially influence the results.

On the basis of equity defined as the absence of systematic differences in one or more aspects of health status across socially, demographically, or geographically defined population groups,6 we proposed to test a hypothesis of equity: there will be no differences in healthcare use by social class, but use will be higher among those with greater healthcare needs. Therefore, health status measured by means of perceived health and common medical conditions reported, was considered an indicator of health need through which we could identify differences among groups having worse health status or with selected illnesses, compared with those having a better or more desirable health status. Specifically, the aim of the study was to analyse the use of healthcare services according to health status in a population of children and adolescents, taking into account family socio-demographic characteristics and characteristics of the proxy respondent.

Methods

Design, study population, and sample

The study has a cross-sectional design. The study population consisted of non-institutionalized children, in Barcelona, 5–14 years of age. Data were obtained from the BHIS the sampling universe of which comprised the non-institutionalized resident population registered as living in the city of Barcelona according to the updated 1998 census, stratified by age, sex, and municipal district. The sample size was set at 1000 residents randomly selected from each of the city's 10 districts. This provides a margin of error close to ±1% for the entire sample, with a confidence level of 95.5%, as well as the possibility of establishing different kinds of territorial and social stratification, with adequate reliability.

Persons selected to participate each received two letters with information about the survey prior to the interviewers' visit to the household. The subjects who were not found after five attempts at different hours were replaced by another person of the same age group, sex, and district. In the children's group, an average of 66.9 interviews were conducted per 100 persons contacted. Some 14.1% of the persons contacted refused the interview, while other replacements were due to other reasons (persons who were not found, who had changed their address, etc.). Complete interviews were obtained on 836 children, 5–14 years old, and their mothers. The health survey was carried out over the course of an entire year, maintaining a random distribution, in order to avoid seasonal impact on morbidity processes and use of health services.

Procedures

The questionnaire was administered to a proxy respondent, preferably the principal caregiver, i.e. the person who usually looked after the child.

Dependent variables

Table 1 shows all the variables included in this study. Utilization of healthcare services included having one or more visits or consultations, including consultations by telephone, with a primary care or a specialist healthcare professional in the 15 days prior to the interview not involving administrative issues; visits to the emergency room over the previous 12 months; and hospitalization over the previous 12 months.

View this table:
Table 1

Variables included in the analysis and their categorization

Variables and instrumentsCategories
Utilization of healthcare services
    Visits to a healthcare professional in the 15 days prior to the interviewYes
Noa
    Visits to the emergency room over the previous 12 monthsYes
Noa
    Hospitalization over the previous 12 monthsYes
Noa
Characteristics of the child
    SexGirls
Boysa
    Age5–9 years
10–14 yearsa
    Percieved health status: CHIP-CE/PRF: Child Health and Illness Profile-Children Edition, Parent Report FormBest health profile: high scores in both, comfort and satisfaction with health; Intermediate health profilea: medium scores in both, comfort and satisfaction with health; and other combinations where categories of comfort and satisfaction with health's domains were different; Worst health profile: low scores in both, comfort and satisfaction with health
    Report of any of the following medical conditions: chronic allergy, asthma, repetitive bronchitis, diabetes, chronic constipation, repetitive otitis, chronic urinary tract infection, growth retardation, febrile convulsion, epilepsy, congenital malformationsYes
Noa
Family characteristics (head of household)
    Social class: Classification proposed by the Sociedad Española de Epidemiología (Spanish Society of Epidemiology)I + IIa: high-level civil servants or private-sector executives and university-level independent professionals III: administrative and service workers, self-employed workers, and supervisors of manual workers; IV + V: manual and unskilled workers
    Educational levelUniversity degreea
Secondary school
Primary school
    Health coveragePublica
Double or private
Characteristics of the proxy respondent
    SexFemalea
Male
    Relationship to childMothera
Father
Other (relative or non-relative)
    Educational levelUniversity degreea
Secondary school
Primary school
    Perceived general health statusGood health: include excellent, very good, and good health
Poor health: include poor or very poor health
    Mental health: GHQ-12: General Health Questionnaire-12 items0–2 pointsa
3–12 points: risk of psychiatric distress
  • a: Categories of reference for the statistical analysis

Independent variables

Characteristics of the child: The variables included were sex, age, report of a medical condition from a list of 11, and health status as measured by the CHIP-CE/PRF. The CHIP is a health measure developed in the US. Based on a broad concept of health, it makes it possible to establish different health profiles.7,8 The version for 12–18 year olds (Adolescent Edition, CHIP-AE) was adapted for use in Spain, with acceptable reliability and validity.9,10 For younger children, the questionnaire utilized in the BHIS used two of the CHIP-CE/PRF domains, including the comfort domain with its three subdomains (physical discomfort, emotional discomfort, and limitation of activities), and the subdomain of ‘satisfaction with health’. These were found to be significantly related to healthcare use in a previous study.11 These domains and subdomains achieved acceptable reliability and validity for this sample,12 similar to those obtained for the original US version.13 The score for each child on a given subdomain is a mean obtained from its items;14 the comfort domain is an average of its three subdomain scores. The averages were standardized at 50, with a standard deviation (SD) of 10, using the average of the sample itself. Then, two cut-off points were established (−0.6 and +0.6 SD of the standardized mean), which made it possible to categorize the dimensions into three levels: low, with a score of <44; medium, with a score of 44–55; and high, with a score of >55.7,15 The health status profiles were established based on nine possible combinations, with the ‘best health profile’ being assigned to those children whose score was high on both domains, and ‘worst health profile’ for those whose score was low on both; the remaining possible combinations were grouped under ‘intermediate health profile’.

Family characteristics: The socio-demographic variables analysed were the educational level of the head of household and social class, which was obtained from the current or previous job of the head of household, according to the classification of the Sociedad Española de Epidemiología (Spanish Society of Epidemiology)16; the six categories from this scale were reclassified into I + II, III and IVa + IVb + V (see description in table 1). The type of health coverage was categorized into public, private, or double, in the case of those reporting that they had both the state-run Catalan Health Service and private health insurance.

Characteristics of the proxy: Regarding the proxy respondent, the characteristics recorded were sex, kinship with the child, educational level, and perceived general health status and mental health status. Perceived health status was recorded as the answer to the question ‘How would you describe your general health?’ with five possible responses, grouped under ‘good health’ and ‘poor health’. Mental health status was assessed with the General Health Questionnaire-12 (GHQ-12),17 which detects the inability to continue to carry out normally the activities of a ‘healthy’ individual and new-onset distress in the general population, outpatients, and primary care users. It does not indicate severity, or the type of disorder; it is a first screening phase for psychiatric distress. Scores on the GHQ-12 are obtained from a summation of responses to the 12 questions on a Likert scale, with the first two response options scoring 0 and the last two scoring 1 (0-0-1-1). On the 0–12 point scale, a cut-off point was used that indicated risk of psychiatric distress with a score of 3 points or higher, which had been used in previous studies.18

Statistical analysis

In order to restore each district proportion in the whole population, a weight was applied during the analysis of data. Having one or more visits to a health professional, the emergency room, and hospitalization were calculated by age, sex, and other independent variables. Comparisons between age groups were assessed using X2 test of significance. The frequencies of different kinds of health coverage were calculated according to social class and educational level of the head of household.

Logistic regression models were fitted to assess the relationship between dependent and independent variables at bivariate as well as at multivariate level. Odds ratio (OR) and its 95% confidence intervals (95% CIs) were obtained. The multivariate analysis was carried out using those variables, which, in the bivariate analysis, were statistically significant (95% CI that did not include the value 1) and those, which, although not statistically significant, were of interest from a healthcare and epidemiological viewpoint (i.e. educational level). Variables were eliminated, one at each step, according to their significance in the model used (Wald statistic) and considering the model's goodness of fit with regard to the previous step (likelihood ratio test). The effects of interaction among the variables included in the final model—including health status and type of health coverage—were also examined.

Because the likelihood of having one or more visits to a healthcare professional in 2 weeks and emergency services in a year is >10%, the adjusted ORs were corrected to obtain an estimation closer to the prevalence ratio (PR),19 by means of the formula: PR = OR/(1 − P0) + (P0 × OR), P0 indicates the prevalence in the category of reference.20

Statistical analysis was carried out with the Statistical Package for Social Sciences (SPSS), version 11.0.

Results

Characteristics of the sample

Table 2 shows the characteristics of the children, their families, and proxy respondents, by age groups. The proportion of children having the best health status was similar in both age groups (18.8% and 18.7%), but for those having the worst health status it was 7.7% for 10–14 year olds and 5.5% for younger children (P = 0.44). Of the 11 medical conditions listed, at least one was reported for 26.4% of the children, with the proportion being slightly higher among the adolescents (28.4%) than among the 5–9 year olds (24.1%).

View this table:
Table 2

Descriptive characteristics of the sample, according to age groups. BHIS-2000

Characteristics of the sample5–9 year olds (n = 382)10–14 year olds (n = 454)Total (n = 836)
n%n%n%
Characteristics of the child
    Sex
        Boy20052.422950.442951.3
        Girl18247.622549.640748.7
        Not knownb000
    Reported common medical conditions
        Yes9224.112928.422126.4
        No29075.932571.661573.6
        Not knownb000
    Health status (CHIP-CE/PRF)
        Best health profile7218.88518.715718.8
        Intermediate health profile28975.733473.662374.5
        Worst health profile215.5357.7566.6
        Not knownb000
Family characteristics (head of household)
    Social class
        I + II12935.512829.425732.1
        III12032.814332.626332.9
        IV + V11531.716537.928035.0
        Not knownb181836
    Educational level*
        University15240.316236.431438.2
        Secondary13235.213630.626832.7
        Primary9224.514733.023929.1
        Not knownb6915
    Health coverage*
        Public23962.532371.156167.1
        Double or private14337.513128.927432.9
        Not knownb000
Characteristics of the proxy respondent
    Informant's sex
        Man7920.79721.417621.1
        Woman30379.335778.666078.9
        Not knownb000
    Relationship to child
        Mother28073.232471.460472.2
        Father7519.78618.916119.3
        Other277.1449.7718.5
        Not knownb000
    Educational level*
        University12733.512627.925330.5
        Secondary14638.316235.830836.9
        Primary10728.216436.327132.6
        Not knownb224
    Perceived health status
        Excellent4612.17616.912214.7
        Very good11129.112026.723127.8
        Good19049.621447.440448.4
        Fair297.6286.2576.9
        Poor61.6122.7182.2
        Not knownb044
    Mental health (GHQ-12)a *
        0–2 points33391.539691.072991.2
        3–12 points318.5399.0708.8
        Not knownb181937
  • *P < 0.05: significant difference between age groups

  • a: GHQ-12: General Health Questionnaire, 12-item version

  • b: Values not known were not included in the percentages

Children from the higher social classes (I + II) made up 31% of the sample, with a higher percentage among the families of the 5–9 year olds (table 2). A similar phenomenon could be seen in the distribution according to educational level, where the proportion of heads of household having finished only primary studies or lower was 24.5% for the under-10 group, and 33% for the adolescents (P < 0.05). Two-thirds of the sample reported having only public health coverage (Catalan Health Service); 30.8% had double healthcare coverage, and 2.0% declared having only private coverage. These proportions were markedly different according to the head of household's social class and educational level. The lower social classes, and the group with a lower educational level, reported having mostly public health coverage (figure 1). The proportion with double health coverage was higher among the 5–9 year olds (37.5%) than among the 10–14 year olds (28.9%), even after adjusting for the head of household's educational level and other characteristics of the children (PR = 1.36; 95% CI = 1.00–1.85) (data not shown).

Figure 1

Health coverage by social class and educational level of the head of household

Women constituted 78.9% of the proxy respondents, and in 72.2% of the cases the woman respondent was the child's mother; 90.9% reported an excellent, very good, or good self-perceived health status, and only 8.8% scored 3 points or higher on the GHQ.

Visits to health professionals in the last 15 days

One-fifth (20.5%) of the sample had visited a healthcare professional during the 15 days prior to the interview. Percentages by age and sex are shown in table 3. The younger group were more likely to have had a visit, girls more so than boys. However, the differences observed between genders for each age group were not statistically significant. Table 4 shows the results of the bivariate and multivariate logistic regression models. The adjusted PR of a visit was associated with the child's health status, since the children having worse health status were more likely to have a visit to health professionals than those having intermediate or better health status (PR = 1.68; 95% CI = 1.09–3.83). No differences were found in the likelihood of visits by social class.

View this table:
Table 3

Utilization of healthcare services by age and sex

Age groups15 previous daysDuring the previous year
Health professional visit or consultationVisit to emergency servicesHospitalization
n%n%n%
5–9 year olds
    Girls18225.818230.81822.2
    Boys20021.519930.72005.5
10–14 year olds
    Girls22220.322231.52224.5
    Boys23015.722935.42303.5
Total83420.583232.28343.8
Not knowna242
  • Percentages (%). BHIS-2000

  • a: Values not known were not included in the percentages

View this table:
Table 4

Utilization of healthcare services by children (5–14 year olds)

15 previous daysDuring the previous year
Health professional visit or consultationVisit to emergency servicesHospitalization
%PRa95% CIPRb95% CI%PRa95% CIPRa,b95% CI%PRua95% CIPRb95% CI
Characteristics of the child
    Sex
        Girl22.81.320.94–1.851.210.89–1.8233.20.910.68–1.220.940.67–1.264.20.800.39–1.620.860.39–1.92
        Boyc18.41131.2113.511
    Age
        5–9 years23.6*1.421.02–1.991.270.94–1.9330.70.880.66–1.180.860.58–1.103.91.020.50–2.080.740.33–1.68
        10–14 yearsc17.9*1133.5113.811
    Health status (CHIP-CE/PRF)d
        Best health profile14.6**0.660.41–1.070.730.41–1.1323.1**0.600.40–0.900.740.42–1.000.60.150.12–0.990.170.02–1.46
        Intermediate health profilec20.5**1133.3**14.711
        Worst health profile38.2**2.361.33–4.211.681.09–3.8345.5**1.670.96–2.901.150.68–2.293.60.970.26–3.670.660.13–3.42
    Report of a medical condition
        Noc18.91128.2**112.3**11
        Yes25.01.430.99–2.061.220.86–1.9043.4**1.961.42–2.701.471.27–2.558.2**3.601.76–7.352.501.12–5.57
Social characteristics (head of household)
    Social class
        I + IIc23.0130.613.91
        III20.10.850.56–1.2930.40.990.68–1.443.50.910.37–2.25
        IV + V18.90.790.52–1.2033.91.160.81–1.674.31.040.44–2.46
    Educational level
        Universityc19.61129.6114.811
        Secondary22.41.180.79–1.771.190.82–1.9031.01.060.75–1.521.040.73–1.553.40.710.31–1.650.950.38–2.39
        Primary20.11.040.68–1.581.070.69–1.7437.21.400.98–2.001.341.05–2.342.90.640.26–1.580.720.25–2.07
    Health coverage
        Publicc19.11130.0113.611
        Double or private23.81.320.93–1.881.230.89–1.9036.81.361.00–1.851.371.16–2.304.41.361.00–1.851.210.51–2.85
Characteristics of the proxy respondent
    Relationship to child
        Motherc21.1133.7*113.81
        Father21.11.000.65–1.5333.3*0.980.68–1.421.130.82–1.794.31.120.47–2.69
        Other14.30.640.32–1.2717.1*0.410.22–0.780.490.19–0.794.20.940.25–3.57
    Mental health (GHQ-12)e
        0–2 pointsc19.51130.0*113.211
        3–12 points28.61.680.97–2.911.480.95–2.9441.4*1.630.99–2.681.441.04–3.017.02.270.82–6.302.280.79–6.58
  • Prevalences, unadjusted prevalence ratios (PRc), adjusted prevalence ratios (PRa) and 95% confidence intervals (95% CI), adjusted through logistic regression analysis, BHIS-2000

  • *P < 0.05; **P < 0.01

  • a: PRu, unadjusted prevalence ratio

  • b: PRa, prevalence ratio adjusted through multivariate logistic regression, incorporating simultaneously into the model all variables whose results are presented in the table; the number of cases that could be included in the multivariate analysis was 801 for the visits with a health professional model, 785 for visits to the emergency service, and 801 for hospitalization

  • c: Categories of reference

  • d: CHIP-CE/PRF, Child Health and Illness Profile-Children Edition, Parent Report Form

  • e: GHQ-12, General Health Questionnaire-12 items

Visits to the emergency room

In the city of Barcelona, one out of three children between the ages of 5–14 had visited the emergency room in the year prior to the interview. Male adolescents presented a slightly higher frequency of visits than females (35.4% versus 31.5%; P = 0.08) (table 3).

In the multivariate model, factors associated with a higher likelihood of visits to the emergency room (table 4) were: having reported a chronic condition (PR = 1.47; 95% CI = 1.27–2.55), a head of household with only primary-level studies (PR = 1.34; 95% CI = 1.05–2.34), double or private healthcare coverage (PR = 1.37; 95% CI = 1.16–2.30), and a score of 3 or higher on the proxy's GHQ-12 (PR = 1.44; 95% CI = 1.04–3.01). The odds of a visit to the emergency service were lower in children having a better health status profile (PR = 0.74; 95% CI = 0.42–1.00), and when the informant was a person other than the mother or father (PR = 0.49; 95% CI = 0.19–0.79).

Hospitalizations

During the year prior to the interview, 3.8% of the sample had been hospitalized. Boys in the 5–9 age group were more likely to have been hospitalized (5.5%) than girls of the same age (2.2%), although the difference was not statistically significant owing to the small number of cases (11 boys and 4 girls). However, in the 10–14 age group, girls were slightly more likely (4.5%) than boys (3.5%) (table 3) to have been hospitalized.

The only factor associated with hospitalization over the previous year at multivariate level was that of reporting a medical condition (PR = 2.50; 95% CI = 1.12–5.57), although the odds of hospitalization was also higher in children with private or double healthcare coverage at bivariate level, and in those cases in which the proxy respondent scored 3 or higher on the GHQ-12; however, this association was not statistically significant (table 4).

Discussion

The present study analyses the utilization of healthcare services in the population aged 5–14 years in the city of Barcelona. In general, no inequities were found in the likelihood of utilizing healthcare services, since differences were not found by social class, whereas the likelihood of utilizing healthcare services was higher in those with greater healthcare needs. Children having worse health status were more likely to have visited a healthcare professional in the 15 days prior to the interview. Children with a reported medical condition were more likely to have visited the emergency service, and to have been hospitalized. Those with a better health profile had visited the emergency services less often over the previous year. Younger children, as expected, were more likely to have visited a health professional than adolescents.

These findings indicate the general success of Spanish health system reforms, which are directed at improving the primary care infrastructure of the health services system to better achieve equity in receipt of health services according to health needs. A prior analysis of 1994 data for children in Catalonia (the region in Northeast Spain of which Barcelona is the capital) revealed inequities in some aspects of services received, such as more waiting time for the last visit and in some services (such as dentistry) that are not covered by public health.21 We found a higher likelihood of utilizing emergency services among children having double coverage, which enables access to more sources of care in addition to those that are publicly available.

The odds of a visit to emergency services were also higher in groups with a head of household with primary studies or lower. Perhaps their precarious working conditions22 make it more difficult for them to access primary care services by appointment and redirect demand towards the emergency services. It seems that people in lower social classes are more likely to attend without an appointment than higher social classes.1

When the proxy presented a higher probability of psychiatric distress, visits to the emergency service were more likely, whereas the informant's perceived health status assessment was not associated with the child's use of healthcare services. The presence of mood disorders (the mental health condition most frequently detected by the GHQ-12) in the person responsible for the child could lead to the need for the child to receive urgent healthcare, independently of the motive for the consultation. It would be interesting to determine whether the higher odds of emergency department visits associated with the informant's mental health problems were due to inappropriate use, i.e. if these were health conditions that could be dealt with on a primary care level. Moreover, and given the cross-sectional nature of the present study, it is not possible to make causal assumptions, such as to determine whether psychiatric distress is caused by the child's illness. Some studies have also shown that mother's mental health has repercussions on the reporting of her child's emotional problems,23 and in seeking emergency healthcare;24 data that support the findings of our study. Other studies in adults have shown an association between the number of chronic physical illnesses reported and the level of psychiatric distress.18 Prospective designs could also contribute to establishing the association between proxies' or children's perceived health and their future healthcare utilization, as in the study by Forrest et al.11

The fact that when the proxy respondents were not the child's mother or father, a lower likelihood of visits to the emergency service were reported could be due to information bias and also be associated with the study's cross-sectional design. Grandparents and others were less knowledgeable about children's use of healthcare services. The sample of proxy respondents was not randomly selected, and children with better health status could be in the care of family members who were neither their mother nor their father.

Whereas the use of proxy responders may be one major limitation of this study, the information gathered regarding the proxy is one of the present study's strengths, since it is not often found in analyses of utilization of healthcare services in child populations, despite the fact that health interview surveys for children are usually answered by a proxy respondent. This study was important in controlling for the characteristics of the proxy respondent. Their influence on the results has been controlled for by means of multivariate models. Moreover, it is necessary to take into account that Health Interview Surveys done in children are usually answered by a proxy respondent.

We were hindered by a limited number of cases in certain categories; for example, the gender differences in hospitalizations could attain statistical significance in a larger population. Separate models for girls and boys showed that girls had a higher likelihood of visits when declaring double or private healthcare coverage, and boys had a higher likelihood of hospitalization when the proxy respondent scored 3 or higher on the GHQ-12. However, these findings need confirmation in a larger sample.

The health survey's design made it impossible to gather information directly from the children themselves, a crucial aspect, and one inherent to perceived health. However, the use of CHIP-CE/PRF dimensions included in the study represents an improvement on the method by which this type of information is gathered in general health surveys. More importantly, from a needs assessment standpoint, it aimed to identify those dimensions for which the system has resources it can use to respond in order to reduce or eliminate differences among groups.11

The finding of no social inequities in the use of health services in the child population, when perceived health status is the measure of need, has potential international implications. A prior study of hospitalizations for ambulatory care sensitive conditions in Spain also found no socio-economic differences, whereas differences in socio-economic status have been found in countries lacking a strong primary care foundation.25 Our finding that double healthcare coverage (as well as psychological distress of the proxy respondent) influences the utilization of services may also have relevance in other countries. Since health interview surveys are one of the most important sources of population data on healthcare services in many countries, and children are not capable of answering some of its items, we emphasize the importance of collecting data on proxy's health and proxy–child relationship. Perhaps, in countries with universal and easier access to healthcare services, parental characteristics have more influence on paediatric care utilization. Also, future health interview surveys should obtain information from children themselves, including self-administered versions, such as the CHIP-Child Edition/Child Report Form.26 These measures would make it possible to identify needs for specific interventions, as well as how to assess the results of their implementation, in order to provide more useful information to decision and policy makers in healthcare services.

Key Points

  • A goal of public-funded universal coverage healthcare systems is to promote the use of services among those children with greater needs.

  • Worse health status perception and reporting a medical condition were the main factors associated with children's use of services.

  • Inequities were not found by social class, although children having double coverage were more likely to visit emergency services.

  • Children were more likely to visit emergency services when the proxy respondent presented a high probability of suffering from psychiatric disorders.

  • Health interview surveys should collect child self-perceived health and proxy respondent characteristics in order to provide useful information to decision and policy makers in healthcare services.

Acknowledgments

The authors are grateful to Delories Dunn and Maite Solans for their contributions to the preparation of the manuscript. This research was partially funded by the Fondo de Investigaciones Sanitarias (contract number 01/0420) and partially supported by funds from Instituto de Salud Carlos III (Network of Excellence IRYSS G03/202). Silvina Berra is PhD Student at the Pompeu Fabra University (Barcelona).

References

View Abstract