Objectives. We examined associations between childhood intelligence and hospital admissions for injuries in adulthood.

Methods. Data were derived from a cohort study (=11103) involving individuals born in Aberdeen, Scotland, between 1950 and 1956.

Results. Overall, 1043 cohort members had at least 1 hospital admission resulting from an unintentional injury over 231152 person-years of risk. There were inverse linear associations between childhood intelligence assessed at the ages of 7, 9, and 11 years and having had a hospital admission stemming from an unintentional injury (gender-adjusted hazard ratio [HR] for a 1-standard-deviation increase in intelligence test score at age 7 years = 0.75; 95% confidence interval [CI]=0.70, 0.80). These associations were not markedly affected by adjustment for childhood socioeconomic status, maternal age or height, birthweight, or childhood growth. However, they were attenuated after adjustment for educational attainment (HR=0.85; 95% CI=0.78, 0.91).

Conclusions. Childhood intelligence is related to hospital admissions for injuries in adulthood, and this relationship is partly explained by educational attainment. The association between childhood intelligence and injury may contribute to the association between childhood intelligence and premature mortality demonstrated in several studies.

Unintentional injuries among adults are an important public health problem resulting in substantial morbidity, disability, and premature mortality.1,2 Injury risk has been shown to be related to educational attainment and socioeconomic position,1,2 but the extent to which this association reflects cognitive capabilities is unknown. Over the past few years, there has been increasing interest in the association between childhood cognitive ability and later health,3 with studies indicating that childhood intelligence is inversely related to all-cause mortality and other adverse health outcomes.47

Little is known about the association between cognitive ability early in life and later injury. To our knowledge, only 1 study has assessed this relationship. In the Australian Veterans Health Study, there was a strong inverse linear association between intelligence assessed in early adulthood and deaths resulting from motor vehicle accidents.8 Associations with other forms of injury or with non-fatal injuries were not assessed. For our study, we examined the association of childhood intelligence measured at 3 different ages (7, 9, and 11 years) with hospital admissions for unintentional injuries in adulthood.

We used data from the Aberdeen Children of the 1950s cohort study. The cohort, described in detail elsewhere,9 comprised participants in the Aberdeen Child Development Survey,10 which collected data on the parental and childhood characteristics of 14 938 children enrolled in primary schools in Aberdeen, Scotland, in 1962.10 Comprehensive information on the 12 150 of these children born in Aberdeen, including the course of their mother’s pregnancy and their physical characteristics at birth, was abstracted from the Aberdeen Maternity Neonatal Databank.10 These 12 150 individuals were born between 1950 and 1956 and were members of the Aberdeen Children of the 1950s cohort.9

In 1999, we began tracing study members through Scotland’s General Register Office; 97% were successfully traced.9 Traced participants were linked to the Scottish Morbidity Register, which provides information including diagnoses coded according to the International Classification of Diseases, Ninth and Tenth Revisions (ICD-9 and ICD-10),11,12 for all admissions to hospitals in Scotland. Because these data are complete only from January 1, 1981, onward, we began our follow-up period on that date. Participants have also been linked to the National Health Service Central Register, which provides detailed information on deaths and migration out of Scotland. We used ICD codes to define types of unintentional injuries (a table of the specific codes used to define types of unintentional injuries is available from the authors).

Throughout the 1950s in Scotland, tests of intelligence were routinely administered to children at ages 7, 9, and 11 years, and results for members of the Aberdeen Children of the 1950s cohort were linked to their study data.9 The tests used were the Moray House Picture Intelligence Test (number 1 or 2) at age 7 years, the Schonell and Adams Essential Intelligence Tests (form A or B) at age 9 years, and a battery of Moray House Tests—2 ability tests (verbal reasoning 1 and 2) and 2 attainment tests (arithmetic and English)—at age 11 years.9 Children took all of these intelligence tests within 6 months of their 7th, 9th, and 11th birthdays, respectively. Tests were age standardized using means of 100 and standard deviations of 15 for Scotland as a whole.

Socioeconomic status early in life has been shown to be related to childhood intelligence13 as well as occurrence of unintentional injuries1,2 and may confound any association between childhood intelligence and unintentional injuries. As a result, we adjusted for a range of indicators of socioeconomic status and other potential confounders from early in life—paternal occupational social class, maternal age at delivery, birthweight, and gestational age—in our analyses. Data on birthweight, gestational age, maternal height (to the nearest inch), paternal occupational social class at the time of the participant’s birth, gravidity, and maternal age at delivery (in 5-year age categories) were abstracted from the Aberdeen Maternal and Neonatal Database.9

During the 1950s and 1960s in Scotland, all children underwent a medical examination at the time of their entry into primary school and had their height and weight measured as part of this assessment. Data from this physical examination were abstracted from the school medical examinations for all participants. We further adjusted for educational attainment to determine the extent to which it mediated any association between childhood intelligence and adult injury.

Between 2000 and 2002 a questionnaire survey was mailed to 11282 surviving cohort members; 7183 (63.7%) responded.14 Respondents were more likely to be women, to have been members of more affluent families in childhood, and to have had higher intelligence test scores as children. We derived data on educational attainment from these questionnaire responses.14 Although these data were collected after most of the injury events assessed here had occurred, most of the participants had completed their formal education before their mid-20s and, therefore, before the occurrence of any outcome events.

We used Cox proportional hazards regression models, with participants’ age as the time axis, to analyze our data. As mentioned, the follow-up period began on January 1, 1981, when complete Scottish Morbidity Register data on hospital admissions became available. Participants were omitted from the analyses if they had died (n=116; 10 of these cases involved unintentional injuries), emigrated to anywhere outside Scotland (n=927), or had a record of a hospital admission for trauma during the period before January 1, 1981, when case ascertainment was incomplete as the register of hospital admissions was being established (n=4). After these exclusions, 11103 (91%) of the original cohort members remained in the analyses.

Contributions to risk were censored at the earliest of the following: (1) first episode of the outcome of interest (if individuals had repeated hospital admissions for the main analyses, they were censored at the first event); (2) emigration date (including emigration to England and Wales, where it was not possible to link information to hospital admissions data); (3) death (there were no deaths resulting from unintentional injuries during the follow-up period); or (4) December 31, 2003, the end of follow-up. We assessed proportionality assumptions by inspecting cumulative incident plots; there was no evidence of any violation. Because the follow-up period began in 1981, our analyses assessed hospital admissions for unintentional injuries that occurred when cohort members were between the ages of 25 and 54 years.

To compare effect magnitudes between intelligence scores at each of the 3 different ages, and to be consistent with other studies on the association of intelligence with health-related outcomes, we divided each of the intelligence measures by its standard deviation and estimated hazard ratios for unintentional injury per standard deviation of intelligence score at each age. Birthweight was standardized according to gestational age and gender. Thus, the hazard ratio for birthweight was represented by a 1-standard-deviation increase in birthweight standardized for differences in birthweight owing to differences in gestational age and gender; as such, it represented intrauterine growth.

There were small amounts of missing data for childhood intelligence scores, gestational age, father’s occupational social class at child’s birth, and childhood height and weight (Table 1). By contrast, a substantial percentage of data on educational attainment (42%) were missing because information on education was obtained from the questionnaire survey conducted in 2000 and was therefore affected by nonresponse (Table 1).

We used multiple multivariate imputation,15 including all other covariates, the log of survival time, and the censoring indicator, to impute a distribution of missing values for variables without complete data.15 We used switching regression in Stata version 9.2 (Stata-Corp LP, College Station, Tex), as described by Royston.15 We carried out 20 cycles of regression switching and generated 10 imputation data sets. There was no evidence of statistical heterogeneity between the data sets generated. We also conducted all analyses on the subsample with complete data (n=5572; 50.2%); the results from these complete data subset analyses were essentially the same as those presented here but were less precisely estimated.

Within the cohort as a whole, there were 9422 families; 5048 (41.5%) of the participants had at least 1 other sibling in the cohort. The usual method of computing standard errors in Cox regression analyses assumes that all study participants are independent (e.g., the intelligence in one individual will not be related to the intelligence of another individual other than by chance). It is likely that within families, this assumption does not hold because siblings’ intelligence will be related to each other through genetic and family environment factors. For this reason, we used robust standard errors (which take into account nonindependence of siblings) to estimate 95% CIs and P values in this study. Stata was used in conducting all analyses.

Table 1 shows the characteristics of the cohort. At the start of the follow-up period (1981), 11103 members of the cohort were alive and believed to be residing in Scotland. Over the follow-up period, they contributed a total of 231152 person-years of risk for injury. Among these individuals, 1043 had at least 1 hospital admission categorized as resulting from an unintentional injury, a rate of 45.3 injuries per 10000 person-years (95% CI=42.6, 48.1). Consistent with previous studies of injury-related mortality,1,2 men in the cohort (69.4 per 10000 person-years; 95% CI=64.7, 74.4) were more likely than were women in the cohort (45.3 per 10000 person-years; 95% CI=42.6, 48.1) to have had at least 1 hospital admission classified as resulting from an unintentional injury.

There were no gender differences in the effects of intelligence or other covariates on injury risk (all interaction P s>.5). Table 2 shows gender-adjusted associations of childhood intelligence and other early-life characteristics with hospital admissions categorized as resulting from unintentional injuries. Childhood intelligence test scores at the ages of 7, 9, and 11 years were inversely associated with adult hospital injury admissions. Childhood social class, maternal height, childhood height, and educational attainment were also associated with injury frequencies.

Gender-adjusted analyses showed a linear association across the childhood intelligence score distribution (P<.001; Figure 1a). Adjustment for childhood socioeconomic position resulted in attenuation of the association (Table 3, model 2), whereas adjustment for birthweight and childhood height and weight had little impact (models 3 and 4). Adjustment for educational attainment resulted in the greatest amount of attenuation (models 5 and 6). Childhood intelligence remained inversely associated with injury after adjustment for all of the study covariates (model 6).

Attenuation of the intelligence–injury association after adjustment for education was most pronounced among individuals with childhood intelligence test scores of 100 or above (Figure 1b), and we undertook post hoc analyses to further investigate this finding (Table 3). We did so by fitting a spline model with a knot (a prespecified position where the regression of exposure on outcome is allowed to change magnitude or direction) at the test score of 100 (because the outcome of Figure 1b suggested this number as the appropriate location). In the fully adjusted association, this spline model provided the best fit to the data (P=.008, with an inverse association for scores below 100 and a null association for those above 100 [Table 3]). However, in the case of the gender-adjusted association, a linear model provided the best fit to the data (spline model P=.6).

Similar findings were obtained when the knot was placed at the test score of 110, but there was no strong statistical evidence to support a spline with a knot at 90 (P=.1). In addition, there was no strong statistical evidence of a quadratic association between childhood intelligence test scores and unintentional injury frequencies in either the fully adjusted (P=.1) or gender-adjusted (P=.5) models.

Because low intelligence test scores were associated with questionnaire nonresponse, and because we imputed educational attainment data for nonresponders, we further explored whether any of our results might have been biased because of missing data. Results showed that the basic gender- and age-adjusted associations between childhood intelligence and unintentional injury were the same among those who did and did not respond to the questionnaire (P=.7 for difference in effect between responders and nonresponders). The results were essentially the same using intelligence measured at age 9 or 11 or the average of the 2 measures. Specifically, an inverse association remained among those with childhood intelligence test scores below 100 but not among those with higher scores after full adjustment (P=.04 for spline model with knot at 100).

The 3 intelligence test scores were highly correlated with each other (all pairwise correlation coefficients were 0.75 or above), and we obtained results similar to those presented for intelligence measured at the age of 7 years for associations between injury and childhood intelligence measured at the ages of 9 and 11 years; moreover, findings were similar when we calculated mean values for all 3 intelligence test scores. When we examined associations with injury subtypes (road traffic and rail accidents, falls, unintended poisonings, and medical accidents), the patterns of association in all of the models were similar to those for injuries overall. Associations between childhood intelligence and injury were also similar for injuries resulting in short hospital stays (less than 3 days) and those resulting in longer stays (3 or more days; all Ps>.4).

Of the 1043 individuals with at least 1 hospital admission for any type of injury, 272 had 2 or more injury admissions over the follow-up period, and 93 had 3 or more such admissions. Childhood intelligence was associated with repeated admissions for injury; that is, an increase of 1 standard deviation in childhood intelligence at the age of 7 years was associated with a gender-adjusted hazard ratio of 0.67 (95% CI=0.59, 0.75) for 2 or more (vs 1 or none) hospital admissions for any form of injury. This hazard ratio attenuated to 0.73 (95% CI=0.63, 0.83) with further adjustment for childhood socioeconomic position and to 0.84 (95% CI=0.72, 0.99) with adjustment for educational attainment.

To our knowledge, this is the first study to examine the relationship between childhood intelligence and risk of nonfatal injury in adulthood. We found inverse linear associations between childhood intelligence and all types of unintentional injuries as well as with specific types of injuries. Childhood intelligence was associated with repeated hospital admissions for injury in adulthood, and similar associations were found for admissions of less than 3 days and those of 3 days or more, suggesting that childhood intelligence is associated with both minor injuries requiring short-stay admissions and more serious injuries requiring longer-length admissions.

Whereas inverse linear associations remained after adjustment for a range of potential confounding and mediating factors, educational attainment resulted in marked attenuation toward the null in the overall association, suggesting that educational attainment may be an important mediator of the relationship between childhood intelligence and risk of injury in adulthood. Attenuation with adjustment for education appeared to be most marked among individuals with childhood intelligence test scores above 100, suggesting that educational attainment may have less impact on the association between childhood intelligence and adult injury among individuals with lower childhood intelligence scores. However, it should be noted that we had no a priori hypotheses in terms of whether education has different effects depending on level of childhood intelligence, and therefore our post hoc results should be treated with caution until they are replicated in other studies.

Study Limitations

A weakness of our study is that reliable information on hospital admissions was available only from 1981 onward. Thus, we were able to examine associations of childhood intelligence with injuries occurring between the ages of 25 and 54 years only, and our findings may have been affected by survival bias if childhood intelligence is related to deaths stemming from injuries occurring in childhood. However, in the original cohort, only 116 (1%) deaths occurred before 1981, 10 as a result of unintentional injuries; therefore, survival bias is unlikely to have affected our results.

Although there were substantial amounts of missing data on educational attainment, we attempted to minimize selection bias via multiple imputation.16 In this strategy, data are assumed to be missing at random, which means that whether an individual has a missing value for a particular variable does not depend on the value of that variable after adjustment for other observed variables. In our present study, data were not missing completely at random because individuals with missing data (largely as a result of questionnaire nonresponse) were more likely to be of low socioeconomic status, to be men, and to have had lower psychometric intelligence in childhood.

However, the missing at random assumption would be met if there was no association between one’s actual educational attainment and the likelihood of not having data for educational attainment (i.e., not responding to the questionnaire asking about education) after adjustment for other variables (such as socioeconomic status, childhood intelligence, and gender) that we have for all or most of the study participants. Although it is never possible to determine that this is the case from the data available in a given study, we cannot think of any reason why it would not be so in the present study. Results of analyses restricted to the subset of individuals with complete data were similar to the results presented here based on 10 multiple imputation data sets. The similar association in the complete data subset supports the idea that data were missing in a random pattern.

The only assessment of adult characteristics available in this cohort occurred after hospital admissions for injuries, and given that injuries may affect people’s socioeconomic position and behaviors, it would have been inappropriate for us to adjust for these characteristics. Thus, we were unable to adjust for indicators of adult socioeconomic position or adult behaviors and lifestyle that might, in addition to education, be mediators of the association between childhood psychometric intelligence and adult injuries.

We used ICD codes routinely applied to hospital admission discharge summaries, but there may have been inaccuracies in these diagnostic codes. For example, there is often concern that some suicides or attempted suicides are coded as unintentional injuries. This is a problem that affects all research in which routine data sources are used, including those related to death certificates.

In general, broad disease categories (in this case, unintentional injuries) are less prone to misclassification than subcategories. As such, we would emphasize as our primary results the associations we found with injuries overall. However, the analyses focusing on injury subtypes did not suggest that any specific subtype drove the inverse association we found with injuries overall. Specifically, there was no evidence that poisonings, which would be the subtype most likely to include suicide attempts, particularly drove this association.

Possible Mechanisms and Study Implications

Several (not mutually exclusive) mechanisms might explain the association between childhood psychometric intelligence and adult injury. First, children with lower intelligence are likely to be at greater risk of nonfatal injuries in childhood. Childhood head injury is associated with increased risk of further injury in adulthood; however, we were unable to assess this possible pathway because information on hospital admissions for injuries during childhood was not available. Second, lower childhood intelligence may be related to a reduced ability to process and use information that could provide protection against environmental risks.6,17

Finally, lower intelligence in childhood is related to lower educational attainment. Educational attainment may influence one’s ability to process information and assess risks, one’s occupation and physical environment, and the type of society and culture in which one lives. For example, one’s level of risk is affected by factors such as the environment in which one lives and works and pressure from peers to engage in risk-taking behaviors. The attenuating effect of education on the intelligence–injury association suggests that this may represent an important pathway, but interventional studies would be required to determine whether general educational interventions reduce injury risk.

In addition, prospective cohort studies with detailed measurements of education, cognition, socioeconomic position, and behaviors from across the life course, including throughout childhood, adolescence, and early and later adulthood, would be useful for determining the most important pathways between childhood intelligence and adult injury. If information on preinjury occupation and other socioeconomic characteristics had been available for the present cohort, we would have been able to explore these pathways further.

Two recent systematic reviews of early learning and school readiness interventions,18,19 one of which focused on randomized trials only,18 concluded that these programs had important effects on reading, arithmetic ability, and general intelligence that extended to adolescents of secondary-school age. Thus, such programs may provide a means through which childhood intelligence can be enhanced. If associations between childhood intelligence and later adult morbidity and mortality—including the associations with adult injuries described here—are causal, then these interventions may have an impact on these later adult mortality and morbidity outcomes. Long-term follow-up of intervention studies would be required to make such an assessment.

Our findings show that childhood intelligence is associated with hospital admissions for injuries in adulthood. This association may in part explain the relationship between childhood intelligence and adult mortality uncovered in earlier studies, and childhood intelligence may represent a modifiable factor that can contribute to reducing adult injury risks.

Table
TABLE 1— Baseline Characteristics of Participants in the Aberdeen Children of the 1950s Cohort: Aberdeen, Scotland
TABLE 1— Baseline Characteristics of Participants in the Aberdeen Children of the 1950s Cohort: Aberdeen, Scotland
CharacteristicSample (n = 12 150), No. (%) or Mean (SD)Missing Data, No. (%)
Girl5868 (48.3)0 (0)
IQ score
    Age 7 y107.1 (16.4)471 (3.9)
    Age 9 y111.3 (17.0)764 (6.3)
    Age 11 y104.1 (13.4)882 (7.3)
Father’s occupational social class at time of birth 681 (5.6)
    I/II (highest)1163 (9.6) 
    III nonmanual1335 (11.0) 
    III manual5319 (43.8) 
    IV1689 (13.9) 
    V (lowest)1963 (16.2) 
Mother’s number of pregnancies 1 (<1)
    13991 (32.8) 
    23505 (28.9) 
    32202 (18.1) 
    41208 (9.9) 
    ≥ 51243 (10.2) 
Mother’s height, in 0 (0)
    ≤60a3101 (25.5) 
    611911 (15.7) 
    622169 (17.9) 
    631777 (14.6) 
    641489 (12.3) 
    ≥ 651703 (14.0) 
Mother’s age at delivery, y 0 (0)
    15–19567 (4.6) 
    20–243798 (31.3) 
    25–293777 (31.1) 
    30–342546 (21.0) 
    35–391108 (9.1) 
    ≥ 40 354 (2.9)
Born outside marriage555 (4.6)0 (0)
Gestational age, wk 1262 (10.0)
    < 37760 (7.0) 
    37–407805 (71.7) 
    > 402323 (21.3) 
Birthweight, lb7.27b (1.13)22 (0.2)
Height at primary school entry, in42.8 (4.1)431 (3.5)
Weight at primary school entry, lb42.0 (6.4)511 (4.2)
Educational attainment 5067 (41.7)
    No formal qualifications1624 (22.9) 
    School leaving certificate173 (2.4) 
    Certificate of secondary education165 (2.3) 
    Ordinary-level qualifications1684 (23.8) 
    Advanced-level qualifications926 (13.1) 
    Higher national certificate1138 (16.1) 
    University degree1374 (19.4) 

a 152 cm or less.

b 3.27 kg.

Table
TABLE 2— Gender-Adjusted Associations (With 95% Confidence Intervals [CIs]) of Childhood Intelligence and Other Early Life Characteristics With at Least 1 Unintentional Injury Hospital Admission in Adulthood: Aberdeen Children of the 1950s Cohort, Aberdeen, Scotland
TABLE 2— Gender-Adjusted Associations (With 95% Confidence Intervals [CIs]) of Childhood Intelligence and Other Early Life Characteristics With at Least 1 Unintentional Injury Hospital Admission in Adulthood: Aberdeen Children of the 1950s Cohort, Aberdeen, Scotland
 No. of AccidentsGender-Adjusted Hazard Ratio (95% CI)P
IQ score
    Age 7 y (per SD)10430.75 (0.70, 0.80)<.001
    Age 9 y (per SD)10430.76 (0.72, 0.81)<.001
    Age 11 y (per SD)10430.72 (0.68, 0.77)<.001
Father’s occupational social class at time of birth  <.001
    I/II (highest)651.00 
    III nonmanual871.08 (0.78, 1.50) 
    III manual4581.45 (1.11, 1.89) 
    IV2122.06 (1.54, 2.75) 
    V (lowest)2211.83 (1.36, 2.46) 
Mother’s number of pregnancies  <.001
    13011.00 
    23001.15 (0.98, 1.35) 
    31911.17 (0.98, 1.40) 
    41091.21 (0.97, 1.51) 
    ≥ 51421.50 (1.22, 1.84) 
Mother’s age at delivery, y  .40
    15–19551.17 (0.89, 1.53) 
    20–348571.00 
    ≥ 351311.09 (0.90, 1.31) 
Born outside marriage  .30
    No9911.00 
    Yes521.16 (0.88, 1.54) 
Mother’s height, in  .02
    ≤ 60a2901.00 
    611740.96 (0.80, 1.17) 
    621770.87 (0.72, 1.05) 
    631590.96 (0.79, 1.17) 
    641230.89 (0.72, 1.11) 
    ≥ 651200.76 (0.62, 0.94) 
Birthweight (per z score)10430.95 (0.88, 1.02).10
Height at school entry (per z score)10430.86 (0.81, 0.92)<.001
Weight at school entry (per z score)10430.92 (0.84, 1.01).08
Educational attainment  <.001
    None3991.00 
    School leaving certificate260.67 (0.41, 1.12) 
    Certificate of secondary education240.90 (0.55, 1.47) 
    Ordinary-level qualifications2510.76 (0.64, 0.91) 
    Advanced-level qualifications820.46 (0.35, 0.58) 
    Higher national certificate1620.71 (0.57, 0.87) 
    University degree990.43 (0.33, 0.57) 

Note. Analyses were conducted with 10 multiple imputation data sets, which allows all 11 103 participants to contribute to the analyses. For the intelligence measures, hazard ratios refer to each 1-standard-deviation increase in score. The hazard ratio for birthweight was that for a 1-standard-deviation increase standardized for differences in birthweight owing to differences in gestational age and gender.

a152 cm or less.

Table
TABLE 3— Multivariate Associations of Childhood Intelligence at the Age of 7 Years With at Least 1 Unintentional Injury Hospital Admission in Cohort Overall and Stratified by Intelligence Score: Aberdeen Children of the 1950s Cohort, Aberdeen, Scotland
TABLE 3— Multivariate Associations of Childhood Intelligence at the Age of 7 Years With at Least 1 Unintentional Injury Hospital Admission in Cohort Overall and Stratified by Intelligence Score: Aberdeen Children of the 1950s Cohort, Aberdeen, Scotland
 Hazard Ratioa (95% Confidence Interval)
 Model 1bModel 2cModel 3dModel 4eModel 5fModel 6g
All participants0.75 (0.70, 0.80)0.79 (0.73, 0.84)0.79 (0.73, 0.84)0.79 (0.74, 0.85)0.83 (0.76, 0.89)0.85 (0.78, 0.91)
Participants with IQ score < 1000.72 (0.62, 0.83)0.73 (0.63, 0.85)0.73 (0.63, 0.85)0.75 (0.64, 0.86)0.76 (0.65, 0.88)0.76 (0.65, 0.88)
Participants with IQ score ≥ 1000.81 (0.71, 0.92)0.83 (0.73, 0.94)0.82 (0.73, 0.93)0.82 (0.73, 0.93)0.97 (0.81, 1.13)0.97 (0.81, 1.14)

Note. Analyses were conducted with 10 multiple imputation data sets, which allowed all 11 103 participants to contribute to the analyses.

a For at least 1 hospital admission categorized as an unintentional injury per each increase of 1-standard-deviation in intelligence score at the age of 7 years.

b Adjusted for gender only.

c Same as model 1 plus indicators of socioeconomic position: father’s occupational social class at time of birth, mother’s previous number of pregnancies, born outside marriage, and mother’s age and height.

d Same as model 2 plus birthweight.

e Same as model 3 plus childhood height and weight.

f Adjusted for educational attainment only.

g Same as model 4 plus educational attainment.

The Aberdeen Children of the 1950s Study was funded as a component project of the Medical Research Council (grant G0828205). A project on cognition and adult health in the cohort was supported by the Chief Scientist Office, Scottish Executive Health Department (grant CZG/2/203), which also funded Heather Clark. Debbie A. Lawlor was funded by a United Kingdom Department of Health Career Scientist Award (award PHCSA).

We are very grateful to Raymond Illsley for providing us with the data from the Aberdeen Child Development Survey and for his advice about the study. Graeme Ford played a crucial role in identifying individual cohort members and in helping us initiate the process of revitalizing the cohort. Sally Macintyre, Doris Campbell, George Davey Smith, Marion Hall, Bianca De Stavola, Susan Morton, David Batty, David Godden, Diana Kuh, Glyn Lewis, and Viveca Östberg collaborated with us in revitalizing the cohort. Margaret Beveridge assisted with study management.

We also thank staff at the Information and Statistics Division (Edinburgh), the General Register Office (Scotland), and the National Health Service Central Register (Southport) for their substantial contributions, and John Lemon, who undertook the linkage to the Aberdeen Maternity and Neonatal Databank. Finally, we thank the study participants who responded to a mailed questionnaire 40 years after the original survey was completed.

Human Participation Protection The revitalization of the Aberdeen Children of the 1950s Study cohort was approved by the Scottish multicenter research ethics committee and local research ethics committees, along with the Scottish Privacy Advisory Committee. Participants responding to the questionnaire provided informed consent to be involved in the study.

References

1. Peden M, McGee K, Sharma G. The Injury Chart-book: A Graphical Overview of the Global Burden of Injuries. Geneva, Switzerland: World Health Organization; 2002. Google Scholar
2. Berger LR, Mohan D. Injury Control: A Global View. Oxford, England: Oxford University Press; 1996. Google Scholar
3. Batty GD, Deary IJ. Early life intelligence and adult health: emerging associations, plausible mechanisms, and public health importance [editorial]. BMJ.2004;329:585–586. Crossref, MedlineGoogle Scholar
4. Hart CL, Taylor MD, Davey Smith G, et al. Childhood IQ, social class, deprivation, and their relationships with mortality and morbidity risk in later life: prospective observational study linking the Scottish Mental Survey 1932 and the Midspan Studies. Psychosom Med.2003;65:877–883. Crossref, MedlineGoogle Scholar
5. Osler M, Andersen AM, Due P, Lund R, Damsgaard MT, Holstein BE. Socioeconomic position in early life, birth weight, childhood cognitive function, and adult mortality: a longitudinal study of Danish men born in 1953. J Epidemiol Community Health.2003;57:681–686. Crossref, MedlineGoogle Scholar
6. Whalley LJ, Deary IJ. Longitudinal cohort study of childhood IQ and survival up to age 76. BMJ.2001;322:819–822. Crossref, MedlineGoogle Scholar
7. Kuh D, Richards M, Hardy R, Butterworth S, Wadsworth MEJ. Childhood cognitive ability and deaths up until middle age: a post-war birth cohort study. Int J Epidemiol.2004;33:408–413. Crossref, MedlineGoogle Scholar
8. O’Toole BI. Intelligence and behavior and motor vehicle accident mortality. Accid Anal Prev.1990;22: 211–221. Crossref, MedlineGoogle Scholar
9. Batty GD, Morton SMB, Campbell D, et al. The Aberdeen Children of the 1950s cohort study: background, methods, and follow-up information on a new resource for the study of life-course and intergenerational effects on health. Paediatr Perinat Epidemiol.2004;18:221–239. Crossref, MedlineGoogle Scholar
10. Illsley R, Wilson F. Longitudinal studies in Aberdeen, Scotland: the Aberdeen Child Development Survey. In: Mednick S, Baert A, Bachmann B, eds. Prospective Longitudinal Research: An Empirical Basis for the Primary Prevention of Psychosocial Disorders. Oxford, England: Oxford University Press; 1981:66–68. Google Scholar
11. International Classification of Diseases, Ninth Revision. Geneva, Switzerland: World Health Organization; 1980. Google Scholar
12. International Classification of Diseases, Tenth Revision. Geneva, Switzerland: World Health Organization; 1992. Google Scholar
13. Lawlor DA, Batty GD, Morton SM, et al. Early life predictors of childhood intelligence: evidence from the Aberdeen children of the 1950s study. J Epidemiol Community Health.2005;59:656–663. Crossref, MedlineGoogle Scholar
14. Nishiwaki Y, Clark H, Morton SM, Leon DA. Early life factors, childhood cognition and postal questionnaire response rate in middle age: the Aberdeen Children of the 1950s study. BMC Med Res Methodol.2005;5:16. Crossref, MedlineGoogle Scholar
15. Royston P. Multiple imputation of missing values. Stata J.2004;4:227–241. Google Scholar
16. Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons Inc; 1987. Google Scholar
17. Deary IJ, Der G. Reaction time explains IQ’s association with death. Psychol Sci.2005;16:64–69. Crossref, MedlineGoogle Scholar
18. Ramey CT, Ramey SL. Early learning and school readiness: can early intervention make a difference? Merrill-Palmer Q.2004;50:471–491. CrossrefGoogle Scholar
19. Barnett WS, Camilli G. Compensatory preschool education, cognitive development, and “race.” In: Fish JM, ed. Race and Intelligence: Separating Science From Myth. London, England: Lawrence Erlbaum Associates; 2002:369–406. Google Scholar

Related

No related items

TOOLS

SHARE

ARTICLE CITATION

Debbie A. Lawlor, PhD, Heather Clark, MSc, and David A. Leon, PhDDebbie A. Lawlor is with the Department of Social Medicine, University of Bristol, Bristol, England. Heather Clark is with the Dugald Baird Centre, University of Aberdeen, Aberdeen, Scotland. David A. Leon is with the Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, England. “Associations Between Childhood Intelligence and Hospital Admissions for Unintentional Injuries in Adulthood: The Aberdeen Children of the 1950s Cohort Study”, American Journal of Public Health 97, no. 2 (February 1, 2007): pp. 291-297.

https://doi.org/10.2105/AJPH.2005.080168

PMID: 17194859