Table 1

Summary of findings table

The association between T1DM and educational attainment in childhood
Patient or population: individuals who have undertaken high stakes testing at the end of compulsory schooling, under 18 years of age.
Setting: school.
Intervention: known diagnosis of T1DM.
Comparison: no known diagnosis of T1DM.
OutcomesImpactParticipants
(studies) (n)
Certainty of the evidence
(GRADE)
Mean final grade of all school subjects from compulsory schooling at age 16 (Mean final grade)Both papers found significantly lower attainment of between 0.07 and 0.08 marks achieved by children with T1DM compared with their non-diabetic counterparts.(Two observational studies)⨁◯◯◯
VERY LOW*†‡§
Mean final grade of specific subjects (Maths, Swedish, English and Sports) from compulsory school at age 16Dahlquist and Källén tested for differences in the odds for each level of attainment by subject in the raw categories and in an alternative four categories where the raw numeric scores (1988–1997) were manipulated to have the same number of categories as the alphabetical marks (1998–2003). In Sports, children with T1DM had a greater odds of a low score and a lower odds of a high score. In Maths and Swedish, children with T1DM had a higher odds of a low score (no clear differences for high scores), and there were no clear effects for English.
Persson et al showed a similar pattern, though all the comparisons of levels within subjects were statistically significant (though comparison for the mean attainment scores for Maths was not significant), with the largest differences in attainment for Sports and the smallest differences for Maths.
(Two observational studies)⨁◯◯◯
VERY LOW*†‡§
GRADE Working Group grades of evidence
High certainty:
We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty:
We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty:
Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect.
Very low certainty:
We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.
  • *Both included studies control for gender, academic year and maternal education. Dahlquist and Källén additionally condition on maternal age and parity, while Persson et al additionally control for paternal education, long-term parental income and parental country of origin.

  • †Heterogeneity could not be reliably assessed as only two papers met final inclusion criteria (one study with updated methodology). As a result, there is a substantial amount of duplication of cases from the databases, with both papers including a cohort of participants born 1973–1978.

  • ‡Publication bias could not be reliably assessed as only two papers met final inclusion criteria (one study with updated methodology).

  • §Total number of participants unable to be combined because of the unknown quantity of data duplication from the same cohort of participants born 1973–1978. (With T1DM—Dahlquist n=5159 and Persson n=2392. Without T1DM—Dahlquist n=1 330 968, Persson n=9563.)

  • GRADE, Grading of Recommendations Assessment, Development and Evaluation; T1DM, type 1 diabetes mellitus.