RF01 Educational outcomes among children with type 1 diabetes: whole-of-population study. (3rd September 2019)

• Record Type: Journal Article
• Authors: Begum, M; Chittleborough, CR; Pilkington, RM; Mittinty, M; Lynch, JW; Penno, M; Smithers, LG
• Is Part Of: Journal of epidemiology and community health. Volume 73(2019)Supplement 1
• Page Start: A55
• Page End: A56
• Publication Date: 2019-09-03
• Subjects: Education -- type-1 diabetes -- Augmented-Inverse-probability-treatment-weighting; Public health -- Periodicals; Epidemiology -- Periodicals; 614.4
• DOI: 10.1136/jech-2019-SSMabstracts.116
• Languages: English
• ISSNs: 0143-005X

Abstract

Abstract : Background: Evidence about the impact of type 1 diabetes (T1D) on educational outcomes is mixed. Despite advances in clinical care and intensive insulin treatment regimens, achieving optimum metabolic control is a challenge in pediatric populations with T1D. Poor metabolic control leading to hyperglycemia or hypoglycemia can potentially have implications for children's educational outcomes. In the last decade, there has been substantial improvement in T1D management, therefore the objective of this study was to estimate to what extent T1D is linked to children's educational outcomes. Methods: This whole-of-population study (n=61, 445) used de-identified, administrative linked data from the South Australian Early Childhood Data Project (births 1999–2013). This study examined the impact of T1D on reading, writing, spelling, grammar and numeracy scores of children in year 5 (age 10 years), assessed by the National Assessment Program-Literacy and Numeracy (NAPLAN) in 2008–2015. Children with T1D were identified from hospitalization data (2001–2014) using ICD-10-AM diagnosis codes (E10, E101-E109). The effect of T1D on the five NAPLAN domains (continuous variables) was estimated by augmented inverse probability treatment weighting (AIPW). AIPW includes; 1) creation of weights and, 2) using those weights in the outcome regression in a way such that the final estimates of the treatment effect is unbiased, even if the weights regression or the outcome regression is incorrect. We explored two associations between T1D and educational outcomes; 1) T1D versus non-T1D, 2) time since diagnosis (≤2 years, 3–10 years) versus non-T1D. Additionally, to address the problem of missing data we used multiple imputation. Results: Among 61, 445 children born in South Australia and who had undertaken NAPLAN assessments, 162 had been diagnosed with T1D. There was no difference in the mean reading, writing, spelling, and grammar and numeracy scores of children with and without T1D. For example, the crude mean reading score was 482.8 with a standard deviation of 78.9, and the average treatment effect was 6.84 (95% CI -6.25, 19.92), which reflects a negligible difference in the mean reading scores of children with and without T1D. There was also no difference in educational outcome between children who were recently diagnosed (exposed to T1D for ≤2years), or those who were exposed to T1D for 3–10 years at the time of NAPLAN assessment, compared with non-T1D. Conclusion: This whole-of-population study demonstrated that children with T1D are not performing poorly on literacy or numeracy at year 5. This could be due to improved T1D management in South Australia