text in 
English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
INTRODUCTION
Providing comprehensive services to mothers and children from birth means a fair start in life. Education and a good early psychomotor development are known to have an impact on health during a person’s lifetime.(1) To achieve health equity, the WHO Commission on Social Determinants of Health (CSDH) proposed improving the living conditions into which children are born to promote early childhood development, combat the unequal distribution of social and economic resources, and measure health problems, analyzing them and evaluating the interventions made to improve.(2)
In the region under study, 291 children below the age of 6 died in 2013, meaning a rate of 10.6 per 1000 live births according to Colombia’s National Administrative Department of Statistics (Departamento Administrativo Nacional de Estadística -DANE); 85% of these deaths occurred in the first year of life, and 63% of the cases occurred from preventable causes.(3) Seeking to reduce morbidity and mortality rates in the general population, with an emphasis on children in the city of Cali, Colombia, the Ladera Public Health Network E.S.E. implemented the Family and Community Health Strategy (FCHS), aimed at providing institutional services to the most vulnerable families of Community 20 (Siloé neighborhood), given the prevailing conditions of poverty in the area. By means of home visits, this program sought to raise parents’ awareness of their families’ health needs and their important role as caregivers in their children’s health.
One of the objectives of the strategy was to reduce morbidity and mortality in the population under 6 years of age. To this end, the local health network created extramural teams formed by a physician, a nursing assistant, a health promoter, and an oral hygienist, who provided primary health care activities on health promotion and disease prevention at the families’ homes. They also referred patients for complementary health care in the Public Health Network.(4) To date, the family health strategy (FHS) has covered 1000 low-income families.
This study aimed to identify the explanatory variables for the health care services provided for ADD, ARI, and dental cavities in children under 6 years of age belonging to the families visited in 2011.(5)
METHODS
The Ladera Public Health Network includes six communities in the western region of the city of Cali (communities 1, 3, 17, 18, 19, and 20). 90% of the population living in this area is classified as strata 1 and 2, which are the poorest according to Colombian standards, with a significant portion of the population living in difficult social and economic conditions, which places them at high social and environmental vulnerability. A multilevel study was used to identify the contribution of characteristics, at both the individual and the contextual levels, associated with the production of services provided by the FHS for some prevalent illnesses in early childhood.
The sample size was estimated based on the method proposed by Jos Twisk and using the census tracts created by the DANE. There is a substantial amount of reports in the literature on sample size calculations in multilevel studies in general, in order to calculate the number of subjects needed in a multilevel study. First, a standard sample size calculation must be performed, and a correction factor must later be applied to it. There are two correction factors, each leading to a different sample size. The first correction factor is as follows: m x n = N x [1 + (n - 1) ρ], where N is the number of study subjects according to the standard sample size calculations; m = number of clusters, n = number of study subjects for each cluster, and ρ = intraclass correlation coefficient (ICC). This factor is known as the design effect, and was used in this study. However, it is important to have in mind that all sample size estimations are based on lots of guesswork.(7) A sample of 90 families was estimated from those visited by the health teams in 2011, and were selected through a systematic random sample procedure. The families were later distributed taking into account the correction factors, the number of neighborhoods in each census tract, the number of observations in each cluster, and the intraclass correlation coefficient; the end result was that 10 clusters of 9 children each should be selected(6,7,Table 1).
Table 1 Distribution of children from Community 20 by sex, age, social stratum, and type of affiliation. Cali, Colombia, 2011
| VARIABLE | Intervened Commune (%) | |
|---|---|---|
| SEX | ||
| Male | 51 (54) | |
| Female | 45 (46) | |
| Age (years) | ||
| 0 | 2 (2) | |
| 1 | 13 (14) | |
| 2 | 18 (19) | |
| 3 | 16 (17) | |
| 4 | 20 (21) | |
| 5 | 27 (27) | |
| STRATUM | ||
| 1 | 88 (91) | |
| 2 | 8 (9) | |
| KIND OF FAMILY | ||
| Nuclear | 25 (26) | |
| Extensive | 20 (21) | |
| Monoparental | 47 (49) | |
| Single parent | 1 (1) | |
| Reconstructed | 3 (3) | |
| No links | 0 (0) | |
| TYPE OF INSURANCE COVERAGE | ||
| Not covered | 21 (22) | |
| Subsidized | 57 (58) | |
| Contributing | 18 (20) | |
| TOTAL | 96* (100) | |
* 90 families from the initial 96 were randomly selected to set up the 10 clusters
The response variable was the number of health care consultations for ADD, ARI, and dental cavities in children under 6 years of age, registered in 2011. The independent variables were as follows: sex; the child’s age; socioeconomic status; type of population (regular, displaced, indigenous, or indigent); type of family (nuclear, extended, single-parent, etc.); health insurance (subsidized, contributing, uninsured, etc.); home ownership (owned, rented, etc.); type of work; materials of house roof; type of home flooring (according to DANE); parents’ educational level; family income; recycling at home; access to public services; the presence of pets at home; humidity in the home; home ventilation conditions; overcrowding; garbage disposal; alcohol consumption, and the use of illegal and legal drugs (tobacco, alcohol) by any parent.
Data of the 90 families were collected through interviews performed at the participants’ homes, using the same questionnaire used by health workers to characterize families. The interviewees signed an informed consent form, thus guaranteeing confidentiality of information.
To answer the research questions, several analytical methods were used. The assumptions of the normal distribution were verified, examining the presence of outliers and multicollinearity among the independent variables that could affect the measurement of the effect between the individual and the contextual variables and the response variable.
A multilevel multivariate linear regression procedure was used to assess whether the contextual variables associated with the effects of the FHS occurred as a function of the characteristics of children residing in the area (a composition effect, fixed) or if they were associated to a higher-order social process (a contextual effect, random). When the same variable was included in both levels, the goal was to test whether a correlation existed between the contextual factor of interest and those that operated at the individual level. Statistical significance was estimated comparing the -2-log likelihood of the reduced model and the full model. To do so, the -2-log likelihood of the model with the random intercept must be compared to the -2-log likelihood of the model with both a random intercept and a slope. The difference between the -2-log likelihoods of the two models follows a Chi square distribution, and the number of degree of freedom is equal to the difference in the number of parameters to be estimated in the two models.
A multilevel analysis was conducted using a multilevel multivariate or multi-hierarchical linear regression. The analysis was done using Stata version 11.2, using the XTMELOGIC command, and the results were compared with the MIXED procedure performed in SAS version 9.3.
RESULTS
ADD: In the analysis of health care visits for ADD as a response variable, the variables of the child’s age and sex were included at the first level. The second level included variables such as health insurance, parents’ educational level, garbage disposal, having pets at home, and census tract. The child’s age and sex were factors associated with health care consultations for ADD; age contributed with 6.3% and sex with 5.3% of the explained variance. At the second level, the variable of having pets at home was included, with a contribution of 9% of the explained variance.
ARI: Regarding health care consultations for ARI, the variables included at the first level were the child’s age and sex. At the second level, the included variables were health insurance, home ownership (owned or leased), home humidity, garbage disposal, parents’ educational level, the ventilation conditions at home, and tobacco and alcohol use by any family member. None of the variables at the first level was associated with health care visits for ARI. At the second level, the associated variables were: parents’ educational level, with a contribution of 3.5% of the explained variance; materials of house roof, with 20.7%, and type of population with 33%. The combined contribution was 57.2%.
Cavities: Concerning health care consultations for dental cavities, the child’s age at the first level and overcrowding and tobacco use at the second level were significantly associated. The child’s age explained 9.4% of the productivity of the FHS in providing care for cavities, whereas overcrowding and tobacco use at home explained 32.1% and 21.9%, respectively, for a 54.0% total.
Table 2 Contribution of variables, according to level, for ADD, ARI, and dental cavities. ESE Ladera. Cali, 2011
| Variable | Explained variance (%) | Level | 95% IC |
|---|---|---|---|
| ADD | |||
| Age of child | 6.30% | 1 | (6.26-6.34) |
| Sex | 5.30% | 1 | (5.18-5.41) |
| Having pets at home | 9.0% | 2 | (8.88-9.11) |
| ARI | |||
| Parents’ educational level | 3.50% | 2 | (3.44-3.55) |
| Materials of house roof | 20.71% | 2 | (20.53-20.86) |
| Type of population | 33.03% | 2 | (32.67-33.32) |
| Dental Cavities | |||
| Age of child | 9.41% | 1 | (8.88-9.45) |
| Overcrowding | 32.1% | 2 | (30.6-32.2) |
| Tobacco use | 21.92% | 2 | (21.69-22.10) |
Source: Prepared by the authors
DISCUSSION
There is consensus to consider the influence of the socioeconomic status of families on their members’ health. Some state that there is no direct effect, but rather an indirect effect mediated by other proxy factors such as the quality of the home.
This study found the same factor associated with health care consultations for ARI, with a 20.7% of contribution to the explained variance. Individual factors such as child’s age and sex were also associated with health care consultations for ADD, with 6.3% and 5.3% of contribution, respectively. In addition, having pets at home contributed with 9%. Regarding dental cavities, environmental factors such as overcrowding and tobacco use contributed with 31.1% and 21.9%, respectively.
Public health strategies that aim to impact prevalent health diseases in early childhood, such as ADD, ARI, and dental cavities, should consider a wide range of interventions from the health sector combined with intersectorial interventions to achieve a greater social impact in reducing the different levels of social exposures and vulnerability associated with family living conditions.(9,10) The use of a multilevel analysis helps identify factors that are beyond the biological aspect, which have been widely studied but tend to explain very little about the onset of health problems in people, such as children in early childhood. Therefore, interventions aimed exclusively at proxy or biological factors will not produce the expected impact, as there are other factors that contribute to or play a greater role on the occurrence of prevalent health problems in early childhood.(11) The results of this study also point to the important contribution of extramural strategies based on the principles of Primary Health Care (PHC) against prevailing childhood morbidity, which are highlighted in the case of health care for dental cavities. This type of care would otherwise not be possible for families in this social, economic, and environmental vulnerability. In summary, this study contributes to better understand the importance of social, economic, and environmental determinants and their interactions with human biology to cause health damage in early childhood.(2)
Other researchers have emphasized the importance of public policies focused on the social protection of families in conditions of extreme vulnerability and the reduction of behavioral risk factors such as tobacco use and harmful alcohol consumption to continuously improve the health of families and kids in early childhood.(13,14,15,16,17,18,19,20) A direct association has been reported between type of family and children oral health. Parents’ mental health and behavioral problems have also been found associated to children’s poor oral health.(21)
A multilevel study conducted in Colombia in 2001 documented the existence of health inequities in health care provision, and the extent of these inequities in the population under 5 years of age. The resulting variables analyzed were the presence of diarrhea, cough, and fever in recent weeks, finding out that the variables showing a significant effect were the presence of a cough with short and rapid breathing and, to a lesser extent, the presence of bloody diarrhea. Nonetheless, including these variables in the multilevel model proved to be of low significance, which, according to the authors, implies that the contribution of these diseases is minimal regarding inequities. This conclusion is similar to the findings of the present study, in which the variables in the first level had a small contribution in the multilevel explanatory model, while variables of the second level were more important. This result highlights the importance of reinforcing policies aimed at improving the families living conditions when trying to truly reduce the prevailing health problems in childhood.(22)
The results of the present study substantiate the need to work on factors beyond the merely biological aspects and to take into account other health determinants that play a crucial role in the survival of the most vulnerable children during their first years of life. Regarding health improvement among the poorest and most vulnerable populations, the promotion of universal access to basic health services such as those provided by the Primary Health Care strategy seems to be the only option for these population groups to receive health care, and therefore to reduce the health gaps between the most and the least privileged in a society like ours.
While this study succeeded in incorporating variables at the individual and family levels, it did not have access to information at higher levels; this can be interpreted as a limitation of the study and should be the subject of future research on identifying structural determinants and the way they act in producing health problems of particular interest.
The results of this study emphasize the need to strengthen alliances that focus on early childhood, including more actors from the civil society to close the health and social gaps, as a political and social objective that benefits early childhood.(3)
