542BRITISH MEDICAL JOURNAL 26 FEBRUARY 1977 domised,5 and our results suggest that claims of benefit based on such trials should not be readily accepted.We thank Mr J Ransom for technical help, and Dr S Variend for reviewing the tumour histology in the patients studied.

Congenital colour vision defects (CVD) are common, inherited (most commonly X linked), non-progressive, and untreatable disorders.1 2 Screening children for these disorders is established practice in the United Kingdom, primarily so that those affected can be advised about occupational preclusions.2 Population based work on the broader impact of colour vision defects is, however, limited. Participants, methods, and resultsWe investigated the association between CVD and education and unintentional injury in the 1958 British birth cohort.3 4 Despite attrition, people remaining were representative of the original sample, including with respect to colour vision status. The latter was assessed in 12 534 children aged 11 years using the Ishihara test, 1 with CVD being the inability to identify all 24 plates. Corrected distance acuity was measured with Snellen charts. We analysed educational, perceptual, and motor skills tests done at 7, 11, and 16 years 3 4 together with highest educational qualification by 33 (none, below O level or equivalent, O level or equivalent, A level or equivalent, or higher). We converted education test scores to z scores 3 and assessed the effect of CVD with multivariate linear regression. We analysed unintentional injuries requiring hospital care by CVD status and sex.Overall, 431 of 6422 boys (6.7%; 95% confidence interval 6.1% to 7.3%) and 68 of 6112 girls (1.1%; 0.8% to 1.4%) had CVD. The distribution of corrected visual acuity did not vary by colour vision status ( 2 trend, P = 0.12). Birthweight, social class at birth, family size, and parental education, all associated with education, were accounted for in the present analysis, although not associated with CVD. At 7 years, CVD and mathematics and reading scores were not significantly associated, after adjustment for age at testing and factors described above (table). At 16, after additional adjustment for prior test scores, children with CVD scored higher than those without; but the small differences, although statistically significant, were functionally unimportant (0.08 standard deviations; 0.002 to 0.16; P = 0.05 for mathematics and 0.07; 0.002 to 0.14; P = 0.04 for reading). There were no significant differences, by colour vision status, for boys or girls, in scores for "copy a design" or "draw a man" at 7 years. Highest educational qualification and colour vision status were not associated for either men or women ( 2 trend, P = 0.07 and P = 0.61). Risk of unintentional injury did not differ significantly (table). Overall, 8.9% (8.2% to 9.6%) of females and 19.2% (18.2% to 20.2%) of males had road injuries as a driver by 33 years; people with CVD reported fewer unintentional injuries (P = 0.08 and P = 0.05). At 33 years, 30% (28.9% to 31.2%) of men reported unintentional injuries in the workplace, without any increased risk in those with colour vision defects (P = 0.293).

SummaryIn a national sample of 16-year-old girls who were aged 12 when the rubella vaccine programme was implemented in 1970, 71% were reported to have received rubella vaccine. There was a high regional disparity in the uptake of rubella vaccine: 81% of girls living in Scotland had been vaccinated but only 61% of girls living in Wales. Similarly there was a difference in reported vaccine uptake according to the family social background, the lowest proportion vaccinated came from professional and unskilled manual families. Girls attending independent schools also had a lower vaccine uptake than girls in schools maintained by the local educational authorities. If rubella immunisation is to be effective uptake of vaccine must increase to almost 100%.Child Development Study5 a medical examination fprm was completed by school doctors for 11 686 children, 5541 of whom were girls. The results of a physical examination and details of the child's medical history, including a question whether the child had been vaccinated against rubella, were available. This information was obtained from interviews with the girls or their parents and from school medical records, which were available for the majority. Of the 5541 girls with completed medical forms, information on rubella vaccine uptake was available for 5097. For the remaining 444 (8(") it was not known if rubella vaccine had been given or the question had not been answered.The reported uptake of rubella vaccine was examined by region, social class (the latter being defined according to the father's occupation6) and the type of school attended. Of the 5097 girls, 3621 (71 %) were reported to have received rubella vaccine but there were considerable regional differences in vaccine uptake (see figure). The highest proportion of girls vaccinated was in Scotland and the South of England, and the lowest was in the North of England and Wales.When the uptake of rubella vaccine was examined according to the child's socioeconomic background, the highest proportion of Introduction Rubella vaccines became available in Britain in 1970, and in the same year a selective vaccination programme was introduced in which girls between their 11th and 14th birthdays were offered vaccine at school without preliminary serological tests for susceptibility to rubella. In the first phase priority was given to the older girls-that is, those aged 13.1 Further recommendations were made by the Joint Committee on Vaccination and

Deaths from Non-accidental Injuries in Childhood SIR,-Publicity given to speculative estimates running as high as 750 a year of the number of small children battered to death by their parents has led us to examine the Registrar General's statistics for faftal accidents in children under 5 years. The Registrar General's statistics for

No abstract