Table 6.
Area-level studies for childhood burn, fall, poisoning, and drowning injuries: summary of methodological features and results (n = 17).
| Author & year country (city/region) | Outcome/s B, F, P, Da | Age group/s data sourceb | SES measure | Analysis covariates | Results: the level of 95% is used for all confidence intervals (CI) |
|---|---|---|---|---|---|
| Birken et al. 2006 Canada (urban areas) | FDeaths | 0–14 years R: death register | Household income for census tracts (quintiles) | Poisson regression Age, sex | For each unit change in income quintile, from highest to lowest, the risk of death from falls increased by 29% (CI 8%−54%). This did not change over time. |
| Durkin et al. 1994 United States (Northern Manhattan) | B, F Hospitalisations and deaths combined | 0–16 years R: injury surveillance system | Household income (3 levels), education (2 levels), unemployment (2 levels) for census tracts (quartiles) | Regression analysis, rate ratios with 95% CI None | Compared to children living in areas with few low-income households, those in areas with moderate and high numbers of low-income households are more likely to have burn injuries (RR = 1.4; CI 1.1–1.8 and RR = 1.6; CI 1.3–2.1, respectively) and fall injuries (RR = 1.5; CI1.3–1.8 and RR = 1.9; CI 1.5–2.2, resp.) |
| Edwards et al. 2008 United Kingdom (England) | F Serious hospitalised injuries | 0–15 years R: centralised inpatient registers | Index of Multiple Deprivation (deciles) | Negative binomial regressionEthnicity, % households with no car, % lone-parent families | The increased risk of falls with greater deprivation disappeared after adjustment (OR=0.57, CI 0.24–1.33 for most deprived decile compared to least deprived one) |
| Faelker et al. 2000Canada (Kingston) | F Injuries seen in emergency departments | 0–19 years R: population-based injury surveillance system | % people living below poverty line for enumeration areas (5 levels) | Poisson regression Age, sex, other SES variables | Gradient of increasing injury with decreasing income; RR = 1.42 (CI 1.21–1.68) for children in poorest quintile compared to those in richest quintile |
| Gagné & Hamel 2009 Canada (Québec province) | P, B, F All, and severe, hospitalised injuries; 6 subdiagnoses of falls | 0–14 years R: hospital administrative data system | Area material deprivation for census dissemination areas (quintiles) | Poisson regression Age, sex, residence location, area social deprivation | Hospitalizations were associated with deprivation, especially for severe injuries. Compared with children in the least deprived quintile, those in the most deprived quintile had higher hospitalisation rates for fire and burn (RR = 2.05; CI 1.5–2.7), and poisoning (RR = 1.68; CI 1.4–2.0) injuries. Associations only significant for particular types of falls |
| Groom et al. 2006 United Kingdom (East Midlands) | P Hospitalisations, 2 broad and 7 narrow subdiagnoses | 0–4 years R: hospital records | Townsend deprivation index of electoral wards (quintiles) | Negative binomial regressionPercentage males, ethnicity, rurality, distance from nearest hospital | Unintentional poisoning was higher among children in the most deprived wards than those in the least deprived. For all poisonings combined, RR = 2.28 (CI 1.78–2.91) for children in poorest quintile compared to those in richest quintile. Gradients were particularly steep for benzodiazepines, antidepressants, cough and cold remedies, and organic solvents |
| Hippisley-Cox et al. 2002 United Kingdom (Trent) | P, B, F Hospitalisations | 0–14 years R: regional admissions data | Townsend deprivation index of electoral wards (quintiles) | Poisson regressionPercentage males, ethnicity, rurality, distance from nearest hospital | Gradient of increasing injury admissions with increasing deprivation. Compared with children in the least deprived quintile, those in the most deprived quintile had a higher admission rate for poisoning (RR = 2.98; CI 2.7–3.3), burns and scalds (RR = 3.49; CI 2.8–4.3), and falls (RR = 1.53; CI 1.5–1.6) |
| Istre et al. 2002 United States (Dallas City) | B Residential fire-related injuries resulting in emergency medical treatment, hospitalisation or death | 0–19 years R: linkage of emergency medical services, hospital, medical examiner, and fire department records | Census tract median income (5 levels) | Chi squared for trend None | There was a marked gradient in the rate of fire-related injuries by income of census tracts. Injury rate in lowest income census tract group was 7.0, compared with 3.1, 1.2, 0, 0 for each successively higher median income grouping (P < .01 by χ 2 for trend) |
| Laflamme & Reimers 2006 Sweden (Stockholm County) | F Hospitalisations; 7 subdiagnoses; 2 severity levels | 0–5 and 6–15 years R: routine centralised inpatient registers | Socioeconomic circumstances index and SES index of parishes (3 levels of each) | Logistic regression None | Results varied by age, fall injury type and severity. Deprived socioeconomic circumstances and low SES typically associated with reduced risk, especially for 0–5 year olds (eg, for falls on the same level, OR=0.63, CI 0.5–0.7 for children living in poor as compared to high socioeconomic circumstances) |
| Lyons et al. 2003 United Kingdom (Wales) | P, B, FHospitalisations; burns including scalds | 0–14 years R: routine centralised inpatient register | Townsend deprivation index of electoral tract (quintiles) | Standardised admission rates, standardised hospitalisation ratios (95% CIs) | Admission rates are significantly higher in more deprived quintiles for each cause. For poisoning, burns, and falls, respectively, rates in the most deprived quintile were 663.6 (CI 622.7–704.5), 81.1 (CI 66.6–95.6), and 1384.0 (CI 1326.3–1441.6) compared to rates in the least deprived quintiles 341.3 (CI 299.3–383.4), 34.9 (CI 21.2–48.6), and 953.9 (CI 889.3–1018.4) |
| Poulos et al. 2007 Australia (New South Wales) | P, B, FHospitalisations; 2 subdiagnoses of falls | 0–14 years R: inpatient register | Index of Relative Socioeconomic Disadvantage of statistical local areas (quintiles) | Negative binomial regression Age, sex | Children in the most disadvantaged quintile were more likely than the least disadvantaged quintile to be hospitalized for poisoning (IRR = 1.52; CI 1.4–1.7) and fire and burn (IRR = 1.95; CI 1.7–2.3) injuries. Children in the most disadvantaged quintile at reduced risk of falls (IRR = 0.78; CI 0.7–0.8) |
| Reimers et al. 2008 Sweden (Stockholm county) | F Hospitalisations, stratified by sex, age and time period (1993–95; 2003–05) | 10–14 and 15–19 years R: regional inpatient register | Socioeconomic deprivation index of parishes (quintiles) | Poisson regression None | For boys, greater deprivation was associated with increased risk of injury only in the first time period and only for the most deprived (ages 10–14years RR = 1.62; CI 1.0–2.6) and intermediately deprived (ages 15–19 years RR = 1.69; CI 1.0–2.8) quintiles. Significant results were present only for girls aged 15–19 years—in the first time period, there was a protective effect of deprivation (RR = 0.65; CI 0.4–1.0 for most deprived), in the second time period, an aggravating effect (RR = 2.62; CI 1.3–5.5 for most deprived) |
| Reimers & Laflamme 2005 Sweden (Stockholm county) | P, B, F Hospitalisations | 0–15 years R: regional inpatient register | Deprivation index, SES index of parishes (3 levels of each) | Rate ratios None | Compared to high SES areas, areas with a greater concentration of people with low SES increased the risk of burn (RR = 2.30; CI 1.5–3.4) and poisoning (RR = 1.65; CI 1.2–2.3) but did not impact on the risk of fall injuries. Moderate, compared to low, deprivation was associated with reduced risk of burn injuries (RR = 0.36; CI 0.2–0.6) |
| Reimers & Laflamme 2004 Sweden (Stockholm county) | F Hospitalisations, 4 subdiagnoses, stratified by sex | 10–19 years R: routine centralised inpatient register | Material deprivation, SES, and multi-ethnicity indices for parishes (3 levels of each) | Logistic regression None | Results varied by sex, fall injury type and index, associations were both aggravating and protective (eg, for falls on the same level OR= 1.22; CI 1.1–1.4 for high, as compared to low, deprivation for boys; but OR= 0.82; CI 0.7–1.0 for girls) |
| Shai & Lupinacci 2003 United States (Philadelphia) | B Deaths from residential fires | 0–14 years R: fire department data | Education level and household income of census tracts (2 levels each) | Logistic regression % children aged under 15; age of house, single-parent households | Low-income tracts had higher odds of experiencing at least one fatal fire-related death (OR=3.18; CI 1.6–6.5) |
| Silversides et al. 2005 Ireland (North and West Belfast) | P, B, F Injuries seen in emergency department, 2 subdiagnoses of falls, burns including scalds | 0–12 years R: emergency department register | The Noble economic deprivation index of enumeration districts (2 levels - most vs. least deprived areas) | Student's t-test None | Although burn, fall and poisoning injuries were considerably higher in the most, as compared to the least, deprived areas, the difference in rates only reached significance for falls <1 metre (RR = 1.90; P < .02) |
| Van Niekerk et al. 2006 South Africa (Cape Town) | B Hospitalisations | 0–12 years R: hospital records | Housing conditions, socioeconomic barriers, and child dependency indices for residential areas (3 levels of each) | Logistic regression None | Children living in residential areas with poor (OR=2.39; CI 2.1–2.8) or impoverished (OR=3.33; CI 2.8–3.9) housing conditions; with medium (OR=1.94; CI 1.6–2.3) or severe (OR=3.61; CI 3.0–4.3) socioeconomic conditions; and with high (OR=1.80; CI 1.4–2.3) child dependency had greater odds of burn injuries than those living in areas with the most favourable levels of these dimensions |
Note aB=burns, F=falls, P=Poisoning, D=Drowning; bR=register; I= interview, Q=self-administered questionnaire.