The relation between family poverty and children’s health is well known.^1–6 However, few population level studies have examined the morbidity of toddlers in relation to parents’ poverty.^7–9 Epidemiological studies include children in the 0 to 17 year age range.^{1–6,10–14} Various studies have examined specific health problems that occur more frequently in poor children.^15–18 Some authors have also emphasised that children from poor families may accumulate several health problems at once.¹ In studies dealing with young children’s health, researchers have often focused on consequences of low birth weight or premature birth but without consideration to the family’s socioeconomic status.^19–22 Although extant research shows a relation between socioeconomic status and child health, there are still unresolved issues pertaining to which aspects of socioeconomic status are important to young children’s health. This study attempts to disentangle some of these issues.

Child poverty is characterised by several inter-related features including lack of material resources, low level of education in the mother, and younger parents. The relative influence of each of these features on the child’s health and the role of relative deprivation compared with absolute lack of material resources is currently the subject of debate.^23–32 According to some, lack of parental skills resulting from lower maternal education better explains disparities in children’s health than lack of material resources resulting from low family income.^26,33–35 Moreover, neonatal health problems and mother’s health habits are sometimes viewed as mediators of the relation between poverty and health of children. Premature births and intrauterine growth delays are more frequent among poor children. Children from poor families are less often breast fed³⁶ and more often have parents who smoke.³³ Given these influences, it is not clear whether or not financial difficulties play a unique part in undermining the child’s health.

Measuring poverty is as challenging as defining it.^2,28,37 Annual family income, from which low income thresholds are established, provides information on resources linked to income over the past year but does not incorporate information about other resources (social or material) on which a family may count in difficult times. An alternative to using family income as an indicator of material strain consists of assessing a lack of money for basic needs. This measure provides a combined indicator of financial difficulties and lack of resources that might better operationalise the notion of financial difficulties that might result in child health problems.

It is noteworthy to mention that although Quebec toddlers are likely to share much in common with American or British toddlers, they benefit from universal health insurance coverage and universal day care programmes that may result in differing cross national patterns. This study therefore examined the links between one indicator of financial difficulties (“a lack of money for basic needs”) and several indicators of health while controlling for maternal education and age, the child’s previous health status, neonatal health problems, his/her characteristics, and other factors associated with each of the health indicators in a cohort of children aged 29 months that is representative of singleton births in the province of Quebec. The goal is to ascertain the unique effect of serious financial difficulties (“a lack of money for basic need”). Children from the cohort have been tracked annually since their birth in the context of the Quebec longitudinal study of child development. Filling in the gap in knowledge regarding the role of the lack of material resources is important in establishing relevant social policies for child health.

METHODS

Analyses were carried out on data from Quebec’s longitudinal study of child development (QLSCD 1998–2002) conducted by Santé Québec, a division of the Institut de la statistique du Québec (ISQ) when children were aged on average 29 months (27 to 31 months). This study annually tracks a representative sample of Quebec children. The initial sample was recruited from the Quebec birth registry when babies were about 5 months old and is representative of singleton births in Quebec in 1997–1998, excluding babies born to mothers living in Northern Quebec, in Cree and Inuit territories as well as on First Nations reservations (2%).³⁸ All singleton births in the included territories were eligible except births of unknown gestational age and premature babies born before 24 gestational weeks.

The research was approved by the human research ethics committee of the Faculty of Medicine of Université de Montréal. Two data sources were used: data on neonatal health problems, weight and height measurements at birth, and mother’s height were drawn from hospital birth records. Other data came from interviews conducted in the home after having obtained written consent from parents of the babies. Questions concerning the mother’s and child’s characteristics, and child health problems were addressed to the mother in 98.8% of cases. Child height was measured by the interviewer following a standardised method at 29 months.

Dependent variables were (1)—mothers’ report of: (a)—presence, during the previous three months, of at least one acute health problems (respiratory tract infections, otitis media, gastroenteritis, or other infections), (b)—occurrence of at least one asthma episode in the past 12 months, (c)—hospitalisations for one or more nights during the previous 12 months, (2)—occurrence of a growth delay (height under the 10th centile of Z scores according to the CDC growth curve, 2000),³⁹ (3)—composite index of health problems at 29 months (presence of 0–1 versus 2–3 of the following problems: acute problems, asthma episode, growth delay).

The main independent variable was a score resulting from an ordinal scale indicating occurrence of a lack of money for basic needs (unable to pay bills for rent, heat, electricity, prescription drugs, children’s clothing, and other necessities over the past 12 months) and its degree of severity as reported by the mother. Scores varied from 0 to 12 and were re-categorised following an analysis of agreement: score 0 =  never been short of money for basic needs; score 1–2 indicating a lack of money for basic needs, and score 3 and over indicating a serious lack of money for basic needs. The undimensional scale had a Cronbach α of 0.79 and was correlated with level of family income sufficiency defined by Statistic Canada. The scale also correlated with chronic poverty defined as having had a low family income at each of the three study periods in 1998, 1999, and 2000. A prior validation study of the scale also showed a Cronbach α of 0.79 (unpublished data).

Main control variables were the mother’s level of education (1—partial or completed college-university studies; 2—high school or trade school diploma; 3—no high school diploma) and neonatal health problems (preterm birth, small for gestational age, congenital abnormality). Fathers’ level of education and occupation were not taken into account because of too many missing data points. Other control variables were child’s health status at 17 months, child’s age, sex, rank in the family, and duration of exclusive breastfeeding, mother’s age, presence of a partner, immigration status, height, parents’ smoking at home, type of day care in the past 12 months, and presence of a house pet at 5 months.

Analyses were conducted using sampling weights. The Direction Santé Québec developed sampling weights based on the sampling design and response rates (see: http://www.stat.gouv.qc.ca/publications/sante/bebe_v2no1_an.htm) to allow for accurate establishment of prevalence of health problems. After descriptive analyses, we estimated unadjusted odds ratios for each health outcome by lack of money for basic needs. Then, multivariate models of the relation between the score for lack of money and the presence of health problems using logistic regression were performed for each dependent variable controlling for confounding variables. Firstly, we controlled for previous health problems and therefore explored the occurrence of health problems since age 17 months. We verified the assumption of homogeneity of slopes and observed no violations. Secondly, we did not control for previous health problems and therefore explored the predictive value of variables on health outcomes at 29 months.

Multivariate models were performed using SPSS logistic regression. All hypothesis tests and confidence intervals were re-calculated taking into account the design effect using sudaan software. Full models are available upon request.

RESULTS

Response rate at 29 months was 94.2%. There was no difference in participation according to presence of an earlier health problem. After listwise deletion of missing data, analyses included 1946 children (97.5% of respondents at 29 months). Tables 1 and 2 show distributions of child, maternal, and environmental characteristics.

A large proportion of mothers stated having experienced a lack of money for rent or mortgage (12%), electricity or fuel to heat (14%), clothing or transportation (13%). Overall, about 10% of mothers had a score of 3 or more on the lack of money for basic needs variable indicating a serious lack of money (table 3).

As table 4 shows, 61.2% of children had had at least one acute health problem during the previous three months; 7.2% had had at least one asthma episode in the previous 12 months; 9.6% had a growth delay (under the 10th centile according to the Z scores on the CDC growth curve, 2000³⁹); 11.2% of the children presented two or more health problems during the preceding year, and 7.6% had been hospitalised at least one night in the year preceding the survey. The most frequent reason for hospitalisation was respiratory problems, which may include asthma.

All indicators of child health were significantly linked to having a score of 3 or more on the scale for lack of money for basic needs (table 4). Those children from families with a serious lack of money for basic needs were more likely to be identified by the mother as having suffered an acute health problem, an asthma episode, a growth delay, or a combination of two or more of these problems compared with those children from families with no lack of money for basic needs. In addition, toddlers living in families with financial difficulties had greater odds of reporting a hospital admission in the previous 12 months. No gradient was observed.

Table 5 presents multivariate models of the relation between score for lack of money for basic needs and child health problems at 29 months. The odds for acute health problems, asthma episodes, and the index of health problems decreased slightly after controlling for the presence of health problems at 17 months and mothers’ level of education, neonatal health problems, and child’s age and sex. As expected, children for whom a health problem had been reported at 17 months had a greater probability of having this same problem at 29 months. However, living in a family with lack of money for basic needs increased the likelihood of having a reported new incidence of acute health problems, a growth delay problem, or cumulating two or more of these problems and possibly asthma episodes. The odds of reporting hospitalisations between 17 and 29 months for those having a lack of money to cover basic needs remained higher even after controlling for all confounding variables in the model. Results of analyses concerning child health status without controlling for health status at 17 months are essentially the same as those above. Odds ratios for lack of money were, however, slightly higher.

No multiplicative or additive interactions between mother’s level of education and score for lack of money were observed for indicators of child health. Therefore, current data do not support the notion that lack of money for basic needs is more strongly associated with health effects among children of low educated women than among children of mothers with higher education. In other words, lack of money for basic needs is equally pervasive for all children regardless of mother’s level of education.

DISCUSSION

Although many studies have shown a relation between low family socioeconomic status and children’s health, the role of different dimensions of low socioeconomic status, and in particular of financial difficulties in the family, has not been thoroughly investigated. This study examined the unique contribution of serious financial difficulties, as measured by a lack of money for basic needs, on the occurrence of health problems in a representative birth cohort of children from Quebec aged between 1.5 and 2.5 years. Results showed that a high score for lack of money for basic needs is linked to occurrence of health problems in toddlers. Children raised in a family with a serious lack of money for basic needs during the preceding year are more likely to present acute health problems, a growth delay, or to cumulate two or more health problems, and to have been hospitalised for the first time within the past few months as compared with children living in a family that does not experience a lack of money for basic needs. This is true regardless of mother’s level of education, age, marital status, immigration status, her health habits, and regardless of the child’s neonatal health conditions. There is no gradient of poor health with increasing score for lack of money for basic needs suggesting that the effect is not attributable to occasionally having economic problems but rather to long term financial difficulties.

These results show that financial difficulties during early childhood have negative effects on the health of children and lead to the occurrence of new health problems in young children between the ages of 1.5 and 2.5 years. In a longitudinal analysis of the same cohort of children, longstanding poverty since birth had a negative impact on these health indicators including a first asthma episode between 17 and 29 months while current poverty had no effect.⁴¹ Cross sectional analyses may thus underestimate the real influence of poverty on children’s health.

When mother’s level of education is integrated into the models, there is little variation in the relation between score for lack of money and different indicators of health. Thus, mother’s level of education does not confound the relation between financial difficulties reported by the mother and occurrence of toddler’s health problems. Mother’s education is not an effect modifier as financial difficulties have a similar effect on children’s health regardless of mother’s education. Family financial difficulties have a unique link with toddlers’ health.

There are two main limitations of this study. Firstly, there is an absence of information on mother’s social resources such as economic and emotional support from family and friends that would allow for testing of the hypothesis of a differential effect of lack of money for basic needs according to the availability of economic help to get out of a crisis. Another limitation pertains to the self report nature of health indicators although literature suggests information on child health reported by the mother is valid.^42,43 Although there was no attempt to define asthma to respondents, the prevalence observed here is similar to prevalence observed in other studies suggesting that the measure did not over-report or under-report this health problem. The score for lack of money may also be a function of the mother’s perception and could vary to her level of education. However, we did not observe any multiplicative, additive, or confounding effects with maternal education level. Lastly, a lack of statistical power is a possibility because the odds of having new asthma episodes became non-significant in the final multivariate model.

Among the study’s strengths, we note that the sample is representative of 98% of singleton births in the province of Quebec. Furthermore, study follow up is excellent with minimal participant attrition. Finally, information is gathered through interviews conducted in participants’ homes and several health indicators were analysed.

Data support the presence of a robust main effect of a lack of money for basic needs on young children’s health. Thus, even in industrialised countries with universal medical care coverage and other wealth redistribution policies, children’s health can be negatively influenced by lack of material resources: infant and child mortality thus seem more closely linked to “…families’ capacity for meeting basic needs than to relative position within a state’s economic hierarchy.”²³

A lack of material resources can affect children’s health directly through under nutrition^44,45 as well as through poverty elicited longstanding stress for both parents and children.^46–48 The presence of more frequent health problems among poor children underscores the fact that the same children often present several health problems. This result suggests a greater overall vulnerability of poor children that may result from impairment of their immune system after long term stress.^46,48

Prevention of health problems in poor children is critical because of the association between current health and adult health and cognitive function.^49–53 Moreover, recent studies show the existence of links between growth delay among poor children and motor and cognitive development^54–56 and cardiovascular disease in adulthood.⁵⁷

Given these and other findings,^13,58–60 social policies must ensure that all families with a young child have sufficient income. Reducing inequalities in health among young children should be a public priority in Canada.⁶¹ Several Scandinavian and European countries such as Sweden, France, Belgium, and Great Britain have succeeded in reducing childhood poverty by “implementing family focused social policies.”⁶² The utility for Canadian children of an alternative strategy that “…stresses the need for a coherent and efficient approach to human capital investment” has also been acknowledged.⁶³ According to these authors, a diversified set of policies may be required and might include income support with measures such as universal child benefits graded according to age, work income credit, higher maternity leave and parental leave benefits, and maternity allowances and tax credit for child care expenses. In-kind transfers should also be offered such as full time public kindergarten for 4 and 5 year old children and early childhood interventions for infants and toddlers in at risk families.⁶³ These findings and recommendations could guide policy makers interested in children’s wellbeing. In the long term, reduction of social inequalities among children will benefit society as a whole.