Introduction

Residential mobility, defined as the frequent change of residence, either within the same city, or between cities, states or communities, is common among American households. Between 2008 and 2009, more than 10 % of households with children between 6 and 17 years of age changed location of their primary residence [1]. Further, residential mobility is more prevalent among low income households [2]. Frequent residential mobility in childhood has been associated with numerous adverse and long-term educational [3–5], behavioral [3, 6–10], emotional and mental health issues [5, 9–15], and physical health outcomes [12, 15–17]. Adverse physical health outcomes ranged from increased incidence and severity of unintentional burns [18] to increased risk of attempted suicide [14] and poor self-assessment of overall health [12] to increased mortality [16, 19]. Children and youth who frequently move also have been found to have poorly coordinated, non-continuous healthcare [15, 17, 20–22].

While there is some research [12, 15–17] to support an association between frequent residential mobility and markers of child health and well-being, these studies have been conducted with small, specialized study populations which are not necessarily generalizable to children aged 6–17 years across the US. In addition, many of these studies [12, 15–17] did not control for child, family, household, and/or environmental factors that could influence purported associations. Further, the characteristics of US children who frequently move and their markers of health and well-being have not been examined or well described in large, nationally representative samples of children. In these analyses, we describe the characteristics of children who frequently move residences and explore the association between the degree of residential mobility and markers of health and well-being.

Methods

We conducted an analysis of cross-sectional data from the 2007 National Survey of Children’s Health (2007 NSCH) using a subpopulation of children, aged 6–17 years, among whom information on residential mobility was collected (N = 63,131). Children <6 years of age were excluded from analysis because some 2007 NSCH variables are not collected among this subpopulation. We conducted descriptive, bivariate and multivariable analyses on this subset of data to describe the association of key markers of health and well-being with residential mobility while adjusting for potential confounding variables.

Outcome Variables

The primary outcome variables are described below and were included in the analysis based on prior research and the availability of child health and well-being variables in the 2007 NSCH.

Child’s overall health was measured based on the question: “In general, how would you describe [child’s name]’s health?” Responses were grouped into three categories: excellent/very good, good, and fair/poor for bivariate analyses and collapsed into two categories (excellent/very good and good/fair/poor) for multivariate analyses.

Preventive medical care was measured based on the question: “During the past 12 months, how many times did child see a doctor, nurse, or other health care provider for preventive medical care such as a physical exam or well-child checkup?” Responses were categorized into one or more preventive visits and no preventive visits.

Severity of current chronic conditions was measured by a variable comprised of responses to 3 questions about 16 different chronic health conditions. Parents were asked to rate the severity of the child’s chronic conditions as mild, moderate, or severe. Responses were grouped into three categories: no chronic condition, mild current chronic conditions, and one or more moderate/severe current chronic conditions for bivariate analyses and collapsed into two categories (moderate/severe chronic conditions and none or only mild conditions) for multivariate analyses.

Child’s overall oral health was based on the question: “In general, how would you describe the child’s teeth?” Responses were grouped into three categories: excellent/very good, good, and fair/poor for bivariate analyses and collapsed into two categories (excellent/very good and good/fair/poor) for multivariate analyses.

Preventive dental care was measured based on the question: “During the past 12 months, how many times did child see a dentist for preventive dental care such as check-ups and dental cleanings?” Responses were categorized into one or more preventive visits and no preventive visits.

Consistency of health insurance coverage was measured using responses to the question: “During the past 12 months was there any time when he/she was not covered by any health insurance?” Responses were grouped into currently uninsured or periods of no coverage during the past year and consistently insured during the past year.

We measured whether the child had a medical home using an indicator variable, which takes into account the six component American Academy of Pediatrics’ (AAP) Medical Home definition (accessible, continuous, comprehensive, family-centered, coordinated, and compassionate) [23]. In order to have a medical home the child must have a usual source of care and a healthcare provider who is considered a personal doctor or nurse. Additionally, if the child had needed health services in the past 12 months, they must have (1) received family-centered, compassionate, culturally effective care from all of the child’s doctors and other healthcare providers, (2) reported no problems getting referrals, if needed, and, (3) if needed, effective care coordination [24]. Responses were categorized into care meets medical home criteria and care does not meet medical home criteria. Elements of a medical home were also explored by indicator variables that measured the presence or absence of a personal doctor or nurse, a usual source of sick and well care, family-centered care, referrals, and effective care coordination.

Explanatory Variables

The primary independent variable of interest was residential mobility. Residential mobility was measured by the question “How many times has the child ever moved to a new address?” [25, 26] The continuous variable was re-classified into a categorical variable: no lifetime moves, 1–2 lifetime moves, and 3 or more lifetime moves.

Other explanatory variables, described below, were included in the analysis based on prior research and the availability of variables plausibly-associated with the outcome variables. Age was re-classified into four levels: 6–8 years, 9–11 years, 12–14 years, and 15–17 years for bivariate analyses, and left as a continuous variable for logistic regression. Race and ethnicity were re-classified into a single variable: Hispanic; black, non-Hispanic; white, non-Hispanic; and, multi-racial/other, non-Hispanic. Family household structure was classified into four levels: two-parent households (two biological or adoptive parents); two-parent households with one step parent; one-parent households (mother only, no father figure); and, all other family household compositions. Household education (highest level of parental education) was classified into three levels: no parent has greater than high school education; at least one parent has more than high school education; and both parents have more than a high school education. Family poverty level was based on the imputed variable for family incomes: 0–99 % of the federal poverty level (FPL); 100–199 % of the FPL; 200–399 % of the FPL; and ≥400 % of the FPL [25].

Statistical Analyses

We created and recoded variables using SAS version 9.2. Analyses were carried out using SAS-callable SUDAAN version 10 to appropriately weight estimates and adjust for the survey’s complex sampling design [25].

We estimated the prevalence and 95 % confidence intervals (CI) of children whose parents reported 0, 1–2, and 3 or more lifetime moves. In bivariate analyses, crude odds ratios (ORs) and 95 % CI were estimated to examine associations between residential mobility and potentially confounding independent variables, e.g., household structure, education, and FPL. Logistic regression models [27] were constructed to examine associations between markers of child health and well-being and residential mobility, while controlling for individual, family, and household characteristics. Children who had never moved served as the referent group. The following explanatory variables were retained in all final adjusted models: age, race/ethnicity, family structure, parental education, and poverty level, health insurance status (except in the insurance model) and presence of special healthcare need (except in the severe conditions model).

Results

There were 63,131 children in the study population with data on the number of lifetime moves. More than 35 % had moved ≥3 times during their lifetimes, 40.9 % had moved 1–2 times, and 23.7 % had no reported lifetime moves.

Table 1 shows the characteristics of children aged 6–17 years and their families by degree of residential mobility. As expected, ≥3 lifetime moves was more common among older children with 42.7 % of 15–17 year olds having ≥3 lifetime moves compared to 26.6 % of children 6–8 years. White, non-Hispanic children were more likely to have zero lifetime moves than children of other racial/ethnic backgrounds. Children living within ‘non-traditional’ family structures (i.e., two-parent step families, households headed by single mothers, and other family compositions) were more likely to have ≥3 lifetime moves than children living within ‘traditional’ two-parent biological/adoptive families. For example, ≥3 lifetime moves were more frequently reported among children of two-parent step families (62.5 %), single mother households (47.6 %), and other family compositions (44.5 %) than for children of two parent biological/adoptive families (26.2 %). Low levels of parental education were also associated with ≥3 lifetime moves. Lastly, children living in families at 0–99 % of the FPL were more likely to move frequently (47.6 %) than children living at 200–399 % or at or above 400 % of the FPL (32.3 and 26.4 % respectively).

Table 1 Characteristics of children aged 6–17 years and their families and households by degree of residential mobility
Full size table

Table 2 presents the markers of health and well-being of the children in the study population by degree of residential mobility. Children who had ≥3 residential lifetime moves had the poorest overall physical health, oral health, and ≥1 current moderate/severe chronic conditions (51.6, 46.6, and 47.1 %, respectively) compared with children who moved less frequently. Children who had ≥3 residential moves were also more likely to be uninsured or experience gaps in coverage (44.3 %). Children who had ≥1 residential lifetime move were more likely to experience fragmented services (lack of a personal doctor or nurse, usual source for sick or well care, family-centered care, needed referrals, and effective care coordination).

Table 2 Markers of health and well-being of children aged 6–17 years by degree of residential mobility
Full size table

Table 3 presents the crude and adjusted odds ratios for residential mobility and markers of child health and well-being. The odds ratios are adjusted for individual, family and household characteristics. Compared to children who had no lifetime moves, children who moved ≥3 times were more likely to have good/fair/poor reported general health status (AOR 1.21 [95 % CI: 1.01–1.46]) and good/fair/poor condition of teeth (AOR 1.31 [95 % CI: 1.15–1.49]). Residential mobility was associated with dental preventive care visits, but not medical preventive care visits. Children who moved ≥3 times were more likely to lack dental preventive care visits (AOR 1.35 [95 % CI: 1.09–1.67]) than children who had no lifetime moves. Compared to children who had no lifetime moves, children who had ≥3 lifetime moves were more likely to have ≥1 current chronic conditions that were rated moderate or severe (AOR 1.40 [95 % CI: 1.19–1.65]). Children who moved ≥3 times were also more likely to be uninsured/have gaps in coverage during the previous year (AOR 1.35 [95 % CI: 0.98–1.87]) compared to children who had no lifetime moves. Lastly, children who moved ≥3 times were more likely to lack a medical home than children who had no lifetime moves (AOR 1.16 [95 % CI: 1.04–1.31]). These associations were only significant for children ≥3 lifetime moves. No associations were found for children with 1–2 moves.

Table 3 Crude and adjusted odds ratios of children’s health outcomes and residential mobility
Full size table

Discussion

We found that even after adjusting for potential confounders, residential mobility was significantly associated with reported poor health, lack of a sufficient medical home, and consistent health care coverage. With the exception of medical preventive care visits (which was not significantly associated with residential mobility) the odds ratios of residential mobility were strikingly similar among the markers of child health and well-being. Additionally, we did not observe any statistically significant effect on markers of child health and well-being with 1–2 moves, indicating an apparent threshold effect with regard to the number of lifetime moves at ≥3 moves. This relationship has been previously reported [17].

Several studies have reported associations between residential mobility and emotional/behavioral outcomes [3, 5–15], particularly in the educational outcome research. This study adds to the smaller body of literature reporting associations between physical health and residential mobility. This study also supports and strengthens the previously limited findings related to markers of well-being by examining several health outcomes while adjusting for individual, family, and household characteristics. While research has established that children living in poverty are more likely to have high residential mobility as well as poor health outcomes, few studies have had the large, nationally representative survey data that would allow for results more generalizable to the US populations and ability to simultaneously adjust for characteristics like household structure, parental education, and FPL.

It was not surprising that a child’s frequent residential mobility was associated with a lower probability of having a medical home or consistent health coverage. Residential mobility is likely both a marker for and the result of chaotic or disrupted family life arising out of a number of social determinants of health, including young maternal age, poverty, lack of safe and stable housing [28], poor employment conditions and opportunities, and lack of or decreased availability of employer-based healthcare coverage. This notion of chaotic and disrupted family life has been particularly evident over the past couple of years during the economic recession and housing crisis [29]. These 2007 NSCH data were collected in the year prior to the economic recession and likely reflect the residential mobility of the nation’s impoverished and working poor households who had limited employment and income opportunities, unstable housing, and limited means to maintain a medical home and health coverage for their children [29–31]. For these families, urgent care clinics, emergency rooms, and public health clinics may serve as their sole or primary source of care.

The association between high residential mobility and preventive dental care also was not a surprise. Lack of preventive dental care may be a reflection of loss of family employment, coverage, or a reduction in health benefits as well as the loss of available income to pay for preventive visits and better quality foods. In turn, poor prevention manifests in poor tooth conditions and increased likelihood of childhood caries [32, 33]. Changes in dietary conditions also may increase the risk of poor oral health.

Children with high levels of residential mobility were more likely to have a moderate or severe chronic condition and reported poor overall physical and oral health. This association may be due in part to a lack of or access to regular medical and dental care. Stress responses caused by frequent mobility may lead to allostasis. These cumulative responses may result in continuously elevated cortisol levels and a cascade of related adverse physiological responses, which may aggravate certain existing chronic conditions [34]. However, whether mobility serves as a stressor for the child may depend on the circumstances surrounding the nature of the residential move. Children may benefit from moves associated with positive situations (e.g., promotion of a parent or moving to be closer to extended family). However, frequent moves associated with negative events (e.g., foreclosure, eviction, divorce, or death) may have profound, compounding effects. Residential mobility is considered one of several stressful life events [35], disrupting daily routines, impacting the development of and ability to maintain friendships/social networks, and negatively impacting classroom learning.

There are several limitations to this research. These data are from a cross-sectional survey and, because both exposure and outcome data were collected simultaneously, temporality between residential moves and health outcomes cannot be assessed. A longitudinal analysis of children could be more informative as to the impact of residential mobility and its specific role in child health. Additionally, parents were asked to remember the number of times the child had ever moved, introducing the possibility of both recall and detection bias. This bias could be more pronounced among older children and children in foster care. Residual confounding is also a limitation. Moving may coincide with other stressful life events, like eviction or divorce, which are not measured in the 2007 NSCH. Thus, the effects associated with frequent mobility in this study could be associated with the other stressful life events rather than residential mobility. Previous research has found residential mobility to be associated with adverse childhood experiences [9]. In addition, it is difficult to separate the specific contribution of residential mobility to the outcomes due to the complex relationships between social disadvantage, housing conditions, family characteristics, and neighborhood characteristics [17]. The study was also limited by the information available around residential mobility. The NSCH does not assess the reasons for moving, the length of time since last move/duration of current residency, whether the residence is rented or owned, or the distance moved. Lastly, because the health effects are reported by the subject child’s parent, misclassification, under/over reporting, and/or recall bias are possible.

In spite of the limitations, this study has several strengths. This study was based on a recent, large nationally representative sample of children to examine the associations between residential mobility and child health and sentinel measures of well-being. To our knowledge, this is the first study that investigates both the association and degree of impact of residential mobility on overall physical and oral health, presence and severity of chronic health conditions, as well as key measures of child well-being such as preventive medical and dental care, consistent health care coverage, and medical home. Using this study as a baseline, it will be interesting to compare these data with NSCH data collected during the present economic recession and housing crisis to determine further impacts of residential mobility and its association with child health and markers of well-being.

Conclusion

This study confirms previous findings regarding the association of residential mobility with the overall physical health of children. The findings suggest that residential mobility—even after controlling for individual child, family and household characteristics—is an important factor in the severity of chronic disease conditions, oral health, continuity of health insurance coverage, and presence of a medical home. Residential mobility may be a potential social determinant of health that warrants further attention. Longitudinal studies of family mobility may be necessary to fully tease apart the relationships between coincident stressful life events, confounding factors, and residential mobility.

Healthcare providers—clinical, public health, and school-based—need to be aware that children who move frequently may lack stable medical homes and consistent coverage increasing their risk of poor health outcomes and aggravation of mild or underlying chronic conditions. Providers can help all families by endorsing system-level changes that would result in universal coverage, mandatory medical homes, and comprehensive electronic medical records. Providers should encourage families to establish a medical home and work with them to maintain and plan transitions in a medical home. Public health systems could provide the necessary link between parents and clinicians to ensure that continuous, coordinated care is established for children who move frequently and experience coverage gaps.