1 INTRODUCTION

Type 1 diabetes mellitus (T1D) is one of the most common chronic health conditions experienced by young people.¹ Worldwide, there are an estimated 2.58 million young people aged ≤19 years with T1D,² with a further 132,000 young people diagnosed with T1D annually.³ In Australia, there are an estimated 6,500 children aged ≤14 years living with T1D.⁴ T1D can influence the lives of young people in a variety of ways, including by affecting their general health,^{1, 5} family, peer and community interactions¹ and both their school performance^{2, 6} and attendance.⁷

T1D can be challenging to manage and without metabolic control, young people may experience acute complications, such as hypoglycaemia or ketoacidosis,² or more long-term chronic complications, such as retinopathy, neuropathy, or nephropathy.¹ There is also evidence that some young people with T1D experience mild deficits in their cognitive, motor, or visuospatial skills.^8-11 Less ideal glycaemic control leading to frequent hypoglycaemic or keto-acidotic episodes could affect neuropsychological development, and thus, cognitive performance, including attention and memory skills.^{12, 13} Cognitive deficits have the potential to influence academic performance among young people with T1D.^{2, 6}

There have been mixed findings on the impact of T1D on academic performance in young people.^{12, 14-16} Systematic reviews and meta-analyses have identified that young people with T1D can experience some cognitive impairment.^{10, 17, 18} Two studies conducted in Sweden and Arizona have identified that young people with T1D have poor academic performance compared to their peers or siblings without T1D,^{11, 19} while other studies conducted in Australia, Scotland, and Washington have all found no difference in academic performance compared to peers.^{14, 15, 20, 21} Three studies have identified differences in academic performance between sub-groups of young people with T1D, depending on their level of metabolic control, with young people with better metabolic control performing better academically.^{12, 14, 15} As previous research has indicated mixed findings in relation to whether T1D has an adverse impact on academic performance, the authors hypotheses that there is a difference between the academic performance of young people with T1D compared to their matched peers. Specifically, this study aims to compare scholastic performance and high school completion in young people hospitalized with T1D compared to matched peers not hospitalized with diabetes.

2 METHOD

This is a retrospective population-based, case-comparison cohort study of young people hospitalized with T1D aged ≤18 years in New South Wales (NSW), Australia using linked birth, health, education and mortality data collections from 1 January 2005 to 31 December 2018.²² Ethical approval and a waiver of consent was obtained from the NSW Population and Health Services Research Ethics Committee (2018HRE0904).

3 RESULTS

There were 833 young people hospitalized with T1D for whom a matched comparison was identified and who completed NAPLAN assessments in Grade 3; 673 in Grade 5; 516 in Grade 7; and 324 in Grade 9 during 2015–2018. There were 843 young people with T1D hospitalized with a matched peer identified who could have completed year 10; 775 in year 11; and 682 in year 12 of high school.

Across the school grades 3–9, there was a higher proportion of females (between 52.2% and 54.5%) compared to males and a higher proportion of young people in urban areas (between72.4% and 74.4%) compared to rural areas hospitalized with T1D. Most young people (≥79%) were from English-speaking households and almost all (≥97%) had no other chronic comorbidities identified. There were no significant differences between parental education or occupation for young people with T1D and their matched peers in each school grade. Young people hospitalized with T1D had higher mean number of ED visits, hospital admissions, and hospital LOS than their matched counterparts in each school grade (Table 1).

Young people predominantly attended government schools in major cities and there were no significant differences by school sector or location of the school for young people hospitalized with T1D, compared to their matched counterparts, in each grade. The proportion of students hospitalized with T1D who did not achieve the NMS for their numeracy or reading assessment in grades 3–9, did not differ significantly from their matched peers (Table 2).

After adjustment for covariates, young females hospitalized with T1D did not have a higher risk of not achieving the NMS compared to matched peers on school assessments for numeracy (ARR: 1.19; 95%CI 0.77–1.84) or reading (ARR: 0.98; 95%CI 0.63–1.50). Young males hospitalized with T1D similarly had no higher risk of not achieving the NMS for numeracy (ARR: 0.74; 95%CI 0.46–1.19) or reading (ARR: 0.85; 95%CI 0.58–1.24), compared to their matched counterparts (Table 3).

Young females hospitalized with T1D showed no difference in completion of year 10, but had a higher risk for not completing year 11 (ARR: 1.73; 95%CI 1.19–2.53) or year 12 (ARR: 1.65; 95%CI 1.17–2.33) compared to their matched peers. Young males hospitalized with T1D had no higher risk for not completing high school in years 10, 11, or 12, compared to their matched counterparts (Table 4).

4 DISCUSSION

This study found no association between young people hospitalized with T1D and the risk of not achieving minimum standards for numeracy and reading in standardized school-based assessments, compared to matched peers for young females or males not hospitalized with diabetes. It also found that young females with T1D had a higher risk of not completing year 11 or 12, but not year 10, compared to matched peers. However, it did not find a higher risk of not completing high school for young males hospitalized with T1D, compared to their matched counterparts. This study has confirmed that academic performance is not adversely affected by T1D, although females with T1D were less likely to complete high school.

The impact of T1D on the lives and abilities of young people is complex and does not appear to be uniform. While T1D has been shown in previous research to negatively affect facets of academic skill development,¹¹ the current study found no discernible difference in academic performance compared to matched peers. Similar to the current study, comparisons of academic performance between young people with T1D and non-diabetic peers in Scotland¹⁴ and Denmark¹⁵ found no significant differences in performance. In addition, two studies conducted in Australia, one in Western Australia²⁰ and another in South Australia among grade 5 students,¹⁶ similarly found no differences in the academic performance of young people with T1D compared to non-diabetic peers.

Differences in the findings on the impact of T1D on academic performance may be explained by the fact that previous studies that found students with T1D had worse academic performance than peers, included student cohorts from the 1970s and 1980s¹⁹ or used subjective teacher ratings of performance, rather than standardized assessments.¹¹ The non-objective assessment of student performance may have resulted in biased reporting of academic performance. It is also likely that significant advances in treatment regimens and modalities, as well as the increased availability of diabetes insulin pump and continuous glucose monitoring (CGM) technologies, which have been associated with improved outcomes in severe hypoglycaemia³³ and diabetic ketoacidosis,¹³ as well as overall diabetes control,³⁴ have improved T1D outcomes²⁰ and subsequently reduced the potential for negative consequences of T1D on student academic performance.

One of the studies that identified decrements in academic performance for young people with T1D compared to healthy peers, did not find consistent performance across each subject.¹⁹ This Swedish study examined performance in mathematics, Swedish, English and sports in young people with T1D compared to peers and found an increased likelihood of not passing, or receiving a lower mark, for mathematics and Swedish, but no significant difference in performance in sport or English for young people with T1D¹⁹ compared to peers. However, the research did identify that young people with T1D had a reduced likelihood of achieving a high mark (e.g., distinction) in sport or English, compared to peers.¹⁹ That consistent performance decrements for young people with T1D are not evident across school subjects illustrates the complexity of the impact of T1D on academic performance. It is also possible that annual standardized school assessments, such as NAPLAN, may not adequately capture everyday difficulties that could be experienced by young people with T1D in the school environment,^{1, 16} and that more nuanced, frequent, and regular monitoring of academic performance is required.

The current study was unable to take into account age at onset of T1D and level of metabolic control (i.e., glycosylated hemoglobin AIc (HbA1c). One prior study found that young people who had early onset of T1D (i.e., <7 years of age) had poorer school performance compared to young people with later onset T1D.¹⁹ There is evidence of structural changes in the brains of young people diagnosed with T1D at a younger age,³⁵ thought to be due to episodes of significant hypoglycaemia, and it is possible that a longer disease duration of T1D, because of early onset age, may be associated with difficulties with cognition, memory, and attention.³⁶ Yet other studies did not identify any impact of age of onset for T1D on academic achievement.^{11, 12, 15, 21} It is probable that once an intensive treatment regimen has commenced for a young person, the age of onset could become less of an issue for academic performance,¹¹ particularly where insulin pump and integrated CGM technologies can ameliorate the incidence of profound hypo- and hyper-glycaemia.³³ However, most research-lacked information regarding adherence by young people with T1D to a treatment regimen.

In comparing academic performance on numeracy and literacy for sub-groups of young people with T1D, previous research found that young people who experienced metabolic control (i.e., HbA1c levels <8%) performed better academically than those with less ideal metabolic control (i.e., HbA1c levels >10%).¹² Similarly, research in Scotland¹⁴ identified that young people with T1D with better metabolic control performed better academically than students with poor metabolic control. Whereas, a study in Denmark¹⁵ found that young people with metabolic control performed worse on assessments of literacy and numeracy than students with poor metabolic control. The authors speculated that the negative association between metabolic control and school performance could be the result of the provision of increased learning support services to students with T1D.¹⁵

The current study found that young females hospitalized with T1D had a higher risk of not completing high school than matched peers. Similarly, Fletcher and Richards³⁷ identified that young people with T1D were 5% to 7% more likely to drop out of high school than their peers. Adolescent females with T1D experience more issues with metabolic control,³⁸ more distress,³⁹ and have a higher risk of developing an eating disorder than young males.^{40, 41} For females with T1D, transitioning from childhood to adolescence may have wider psychosocial implications that contribute to young females having higher risk of early school leaving. The additional psychological distress and a higher prevalence of eating disorders among young people with T1D⁴² attests to the need for regular psychological screening of young people with T1D.⁴³

Other studies did not find a similar association with T1D and high-school dropout. In Scotland, there was no increase in the proportion of young people with T1D leaving school before age 16 compared to peers.¹⁴ Likewise, in Finland there was no difference in young people with T1D leaving school at age 16 compared to peers, including no difference by sex,⁴⁴ nor in Sweden.¹⁹ However, in Sweden, once the age of onset of T1D was taken into account, children diagnosed with T1D before 2 years of age had a higher likelihood of dropping out of high school.¹⁹ The differences found between studies could relate to case identification criteria, with the current study only including young people hospitalized with T1D, therefore was more likely to include more poorly controlled cases. Other studies identified cases from clinical registers or from prescribed medications to identify any young person with T1D,⁴⁴ which could also include young people who had not been hospitalized with T1D.

Advances in treatment modalities,^{15, 20} compared to some of the earlier studies of the impact of T1D on academic performance, better glycaemic control through improved diabetes management,^{11, 20} such as using insulin pumps rather than insulin injection modalities,¹⁶ use of CGM technology, especially where CGM is linked to pump technology, and better communication and training regarding T1D management strategies, including with school staff^{1, 16} are all likely to have contributed to obtaining the similar educational outcomes for young people with T1D compared to non-diabetic peers. Further support for students with T1D is being provided in Australia with the launch in 2020 of the national diabetes in schools program in Australia.⁴⁵ The program aims to provide information, training, and support for educators, families, and health professionals to better support young people with T1D.⁴⁵

Further research examining T1D and academic performance in Australia could link health and education performance data to primary care attendances, diabetes registry and pharmaceutical dispensing data, as this would allow confirmation of age at T1D diagnosis, diabetes management, and outcomes. Additional research is also required to further tease out the impact of recent versus latent T1D diagnosis on a young person's academic performance and also identify if there are any later health and social outcomes, including on employment opportunities.

This study's strengths were that it was a large population-level cohort study, linking health and standardized academic assessment records over a 13-year period and was able to adjust for key factors that may influence academic performance, such as socioeconomic status and parental education. However, the study had some limitations. The study only included young people who had been hospitalized with T1D, so did not include young people presenting solely to tertiary center diabetes outpatient departments, primary care or other health professionals for treatment. Although, hospitalized young people with T1D were more likely to have less perfect diabetes control. There was no information in the hospitalization data regarding metabolic control (i.e. HbA1c levels, time in range on CGM or the proportion of time high, very high, or low), number of severe hypo/hyperglycaemic episodes, frequency of diabetic ketoacidosis, prescribed insulin doses/kg, age of T1D onset for the young person, or access to specialist care from endocrinologists. Only chronic health conditions relevant to a hospital admission are recorded in diagnosis classifications and it is likely that some comorbidities were not identified. This is particularly the case for the comparison cohort, where 68.6% had not been hospitalized, leaving no opportunity to identify comorbid conditions, despite the 3-year lookback period.

A small number of residential postcodes were not able to be provided to the authors and socioeconomic status for these records was not able to be identified. The recency of postcode of residence may vary between data collections, which may affect estimates of socioeconomic status. Data validity was not able to be assessed and it is possible that there could be some data misclassification in hospital or education records. Visits to private hospital EDs were not able to be accessed for this study. A higher proportion of young people with T1D were absent for NAPLAN assessments compared to their matched counterparts. The current study was not able to take into account school clustering and no information was available concerning any tutoring or supplementary education services that a young person may have received.

5 CONCLUSION

There was no difference in the academic performance of either young females or males with T1D compared to matched peers. Better glucose control and awareness of strategies to manage T1D, including diabetes education programs for young people in schools, may explain the absence of school performance decrements for students with T1D. However, young females hospitalized with T1D did have a higher risk of leaving high school than their matched peers. Additional research is needed to explore psychological and social factors, as well as T1D control factors, that may contribute to this increased risk, and suggests the need for routine psychological screening as recommended in best practice guidelines.

CONFLICT OF INTEREST

The authors declare they have no conflict of interest.

AUTHOR CONTRIBUTIONS

RM, CC, and AM were all involved in study concept and design. RM acquired and organized the data, conducted the analysis and wrote the first draft of the manuscript. TB-P provided statistical assistance for the GLMM analysis. All authors were involved in interpretation of data and critical revision of the manuscript.