Article Text
PDF
Abstract
Introduction Among youth with type 1 diabetes (T1D), longitudinal poor glycemic control is associated with adverse socioeconomic conditions at the neighborhood level. Child Opportunity Index (COI), which encompasses measures of education, health, environment, social, and economic factors, is associated with obesity in youth but has not been evaluated in youth with new-onset T1D or type 2 diabetes (T2D). We hypothesized that lower COI would be associated with adverse clinical outcomes at diabetes diagnosis, and due to differing risk factors and pathophysiology, that youth with new-onset T2D would have lower COI than youth with T1D.
Research design and methods Retrospective cohort of youth with new-onset diabetes admitted to a large academic pediatric hospital. COI was compared by diabetes type using t-tests and Χ2 tests. Multivariable linear and logistic regression analyses were used to evaluate associations between COI and clinical characteristics, stratified by diabetes type and adjusted for age and sex.
Results The cohort (n=484) differed in race and age by diabetes type (T1D: n=389; 10.0% black, 81.2% white; age 9.6±0.2 years; T2D: n=95; 44.2% black, 48.4% white; age 14.8±0.3 years). Youth with T2D had lower COI (p<0.001). Low COI was associated with diabetic ketoacidosis in T1D and T2D. Black youth with low COI had the highest hemoglobin A1c among youth with T2D and the highest obesity prevalence among youth with T1D.
Conclusions COI is associated with differing characteristics at diagnosis in youth-onset T1D and T2D but is worse among youth with T2D overall. These findings underscore the need to address socioeconomic adversity when designing interventions to reduce T2D risk and to improve outcomes at diabetes diagnosis in youth.
- Diabetic Ketoacidosis
- Diabetes Mellitus, Type 1
- Diabetes Mellitus, Type 2
- Pediatric Obesity
Data availability statement
Data are available upon reasonable request. De-identified data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Individual-level and neighborhood-level social determinants of health are increasingly recognized as key determinants of clinical outcomes in youth with type 1 diabetes and type 2 diabetes. Child Opportunity Index, a composite measure of education, health, environment, social, and economic factors important for child thriving, is associated with childhood obesity but has not been evaluated in the context of youth with newly diagnosed type 1 or type 2 diabetes.
WHAT THIS STUDY ADDS
In this cohort of 389 youth with type 1 diabetes and 95 youth with type 2 diabetes, Child Opportunity Index was significantly lower among youth with type 2 diabetes than type 1 diabetes. Among youth with either type 1 or type 2 diabetes, lower Child Opportunity Index was associated with higher odds of diabetic ketoacidosis. An interactive effect between race and Child Opportunity Index was demonstrated among youth with type 2 diabetes, with black youth living in neighborhoods with low Child Opportunity Index having the highest hemoglobin A1c at diagnosis.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Child Opportunity Index may represent a novel measure to guide future screening or intervention efforts aimed at reducing incidence of diabetic ketoacidosis among incident youth-onset diabetes, as well as earlier recognition of new-onset type 2 diabetes.
Introduction
Social determinants of health, the conditions in which individuals live, work, and play, are tightly linked with clinical outcomes in type 1 and type 2 diabetes (T1D, T2D) across the lifespan.1 2 Much of these findings have centered on individual-level poverty, food insecurity, and healthcare access among adults. However, neighborhood-level composite indices may illuminate how exposures to a variety of adverse conditions impact diabetes risk and outcomes. This is particularly important for youth-onset T2D given a rapidly rising incidence3–5 and early, severe outcomes that disproportionately impact minoritized individuals.6
A recently developed neighborhood-level index, Child Opportunity Index (COI), has emerged as a promising indicator of adverse health outcomes in youth, including obesity,7 recurrent diabetic ketoacidosis (DKA) in youth with T1D,8 and potentially avoidable hospitalization.9 This composite index, developed in 2014 and revised in 2020 to version 2.0, is available for nearly all census tracts within the USA and encompasses 29 measures of neighborhood-level resources influencing pediatric health and development, including but not limited to access to high-quality schools, healthy foods, green space, and toxin-free environments.10–12 COI is strongly associated with measures of structural racism, including historical redlining, blockbusting, and urban renewal,13 highlighting the intersectionality of race and racism with adverse social determinants of health.
We investigated whether COI was associated with clinical characteristics at diagnosis of T1D and T2D. We hypothesized that COI would be worse among youth with T2D and would be associated with more severe presentation, including more severe hyperglycemia and acidosis.
Research design and methods
Study cohort
This retrospective cohort study included youth aged 6 months–20 years who were diagnosed with T1D or T2D and admitted to UPMC Children’s Hospital of Pittsburgh, a large academic center, between January 1, 2021 and December 31, 2022, for management of new-onset diabetes. Electronic medical records were manually reviewed after approval from the University of Pittsburgh Institutional Review Board.
Outcome measures and covariates
Clinical outcomes included DKA (based on pH <7.3, metabolic acidosis with bicarbonate <15 mmol/L or clinical notes if laboratory data were incomplete), pH, serum bicarbonate, glucose, and hemoglobin A1c (HbA1c). Body mass index (BMI) was calculated using admission weight and height. BMI Z-score was determined using Centers for Disease Control and Prevention growth charts, and obesity defined as Z-score ≥1.64.14 Age, sex, race, ethnicity, and zip code were manually extracted from the medical record. Diabetes type was based on clinical notes and diabetes autoantibody status.
As described above, COI is a measure of neighborhood-level resources that combines 29 indicators of three opportunity domains (educational, health and environmental, and social and economic). COI ranks neighborhoods from very low to very high opportunity.8 9 Zip code-based nationally normed COI was obtained using the COI V.2.0 database that included 2020 US Postal Service zip codes, the closest available to the study period.15
Analysis
Patient characteristics are reported using summary statistics. Continuous variables were compared by diabetes type using t-test and Wilcoxon rank-sum test, depending on distribution. Categorical variables were compared using the Χ2 test. Multivariable linear and logistic regression analyses were used to evaluate associations between COI and diabetes outcomes in separate models, adjusting for age and sex, stratified by diabetes type. Due to the infrequency of high/very high COIs among youth with T2D, these categories were combined with moderate COI, and low/very low COIs were combined, leading to a two-level categorization of COI (not low=moderate, high, very high; low=low, very low) for multivariable analyses. This binary classification was also used in unadjusted logistic regression to determine the OR of low/very low COI by diabetes type. Due to the anticipated collinearity of race and COI, intersectional race-COI categories were created (eg, black-not low, black-low). Missing data were not imputed. All analyses were performed using Stata V.17 (StataCorp; College Station, Texas, USA) and used a two-sided alpha=0.05 for statistical significance.
Results
Cohort characteristics
A total of 548 patients with new-onset diabetes mellitus were admitted during the 2-year study period. Of these, n=484 (T1D: n=389; T2D: n=95) were ≤20 years old and had available COI and evaluable DKA status upon admission (online supplemental figure 1). Nearly all (T1D: 91.3%, T2D: 92.8%) were either white or black, so race was collapsed into white, black and other, which included Asian or Indian, multiracial, American Indian or not reported (table 1). Youth with T2D were older on average and had higher prevalence of obesity than youth with T1D. Youth with T1D were predominantly white, while white and black races were nearly equally represented in T2D. HbA1c did not differ by diabetes type, but patients with T1D had lower pH and bicarbonate concentration.
Supplemental material
Cohort characteristics, T1D versus T2D
As shown in table 1, half of youth with T2D resided in neighborhoods with low or very low COI, in comparison with less than one-quarter of youth with T1D (low/very low COI in T2D vs T1D: OR 3.5 (95% CI 2.2, 5.7)). However, this difference by diabetes type was driven by white youth (low/very low COI in white youth with T1D vs T2D: 16.8% vs 47.9%; OR 4.6 (95% CI 2.4, 8.7)). COI distribution was similar between T1D and T2D for black youth and youth of other races (figure 1). Notably, although in T1D, low/very low COI was significantly more common in black youth (OR 5.8 (95% CI 2.9, 11.6)) and youth of other races (OR 3.1 (95% CI 1.4, 6.5)) than white youth, odds of low/very low COI did not differ significantly by racial group in T2D.
COI by diabetes type in each racial group, demonstrating higher exposure to low COI for white youth with T2D versus T1D. In contrast, low and very low COI was common among black youth and youth of other races regardless of diabetes type. COI, Child Opportunity Index; T1D, type 1 diabetes; T2D, type 2 diabetes.
Association of race-COI and clinical presentation at diabetes onset
As shown in table 2, in multivariable regression among youth with T1D, DKA was more common among black youth and youth of other races with low COI (66.7% (95% CI 46.6% to 86.7%) and 76.9% (95% CI 54.1% to 99.8%), respectively), versus white youth with not-low COI (42.2% (95% CI 36.3% to 48.1%)) (figure 2A). In contrast, among youth with T2D, DKA was most common among white youth with low COI (31.8% (95% CI 12.5% to 51.1%) vs 4.2% (95% CI −3.8% to 12%) among white youth with not-low COI) (figure 2B). Among youth with T1D, obesity was most common among black youth with low COI (36.8% (95% CI 15.3% to 58.4%)) but did not otherwise differ by race and COI (figure 3A). In contrast, obesity was equally common among all race-COI groups in T2D, ranging from 70.8% to 95.5% among each subgroup (figure 3B). HbA1c did not differ significantly across race-COI groups within T1D. However, among youth with T2D, HbA1c was 1.6% higher in black youth with low COI compared with white youth with not-low COI (12.3% (95% CI 11.3%, 13.3%) vs 10.7% (95% CI 9.7%, 11.8%)). Severity of acidosis as assessed by pH <7 and bicarbonate <5 mmol/L did not differ by race-COI group.
Predicted marginal per cent of youth with DKA on admission for new-onset (A) T1D and (B) T2D, based on multivariable logistic models adjusted for age and sex. DKA was more common among youth with T1D than T2D but differed by race and COI. * indicates significant (p<0.05) difference from the reference category (marked with arrow) of white youth with not-low COI. B-L, black-low COI; B-NL, black-not-low COI; COI, Child Opportunity Index; DKA, diabetic ketoacidosis; O-L, other race-low COI; O-NL, other race-not-low COI; T1D, type 1 diabetes; T2D, type 2 diabetes; W-L, white-low COI; W-NL, white-not-low COI.
Predicted marginal per cent of youth with obesity (BMI Z-score ≥1.64) on admission for new-onset (A) T1D and (B) T2D, based on multivariable logistic models adjusted for age and sex. Obesity was more common among youth with T2D than T1D but was highest among black youth with low COI in the T1D cohort. * indicates significant (p<0.05) difference from the reference category (marked with arrow) of white youth with not-low COI. B-L, black-low COI; BMI, body mass index; B-NL, black-not-low COI; COI, Child Opportunity Index; DKA, diabetic ketoacidosis; O-L, other race-low COI; O-NL, other race-not-low COI; T1D, type 1 diabetes; T2D, type 2 diabetes; W-L, white-low COI; W-NL, white-not-low COI.
Adjusted marginal estimates of diabetes-related outcomes from multivariable regression models, stratified by diabetes type
Conclusions
In a large retrospective cohort of youth with new-onset T1D and T2D, low and very low COIs were more than twice as common among youth with T2D and were present in half of the cohort, underscoring that adverse socioeconomic status is a common experience for youth with T2D.16 COI was associated with DKA in both T1D and T2D and may therefore be a useful risk-stratification marker when designing population-level interventions to prevent DKA at diagnosis. COI’s differential associations with obesity (in T1D only) and HbA1c (in T2D only) reflect the differences in pathophysiology of these diseases but also that adverse clinical characteristics, in general, are more common among youth with low COI.
In youth with new-onset T1D, COI was associated with obesity. This is consistent with findings in the general pediatric population, in which a dose–response association between COI and BMI was previously demonstrated.7 The COI–obesity association may reflect differing availability of safe places for physical activity between neighborhoods with higher versus lower COI.17 18 Notably, we did not detect a difference in obesity prevalence among youth with T2D. This is likely due to the existence of very few youth without obesity in this sample, limiting the power of this analysis. In contrast, obesity was less prevalent among youth with new-onset T1D. It is also possible that youth with new-onset T2D from neighborhoods with low COI, in whom HbA1c tended to be higher, experienced weight loss due to unrecognized and untreated hyperglycemia, thus explaining our null findings. Our finding that COI is associated with DKA at admission is consistent with those from Bergmann et al, who found that lower COI was associated with higher probability of DKA readmission within 1 year among youth with T1D.8 Bergmann et al also found that black youth remained at highest risk of recurrent DKA, even after accounting for COI.
Although race is a social construct, the impact of structural racism can become embodied19 in intergenerational risk, with greater risk of youth-onset T2D conferred by maternal diabetes during pregnancy and other family history of T2D.20 The potential combined effect of adverse socioeconomic exposures and population ancestry-based risk was demonstrated in a study by Cromer et al, who found that an area-level measure of socioeconomic status was associated with obesity and T2D in adults, with an additive effect among those with higher genetic risk scores.21 These findings, along with our demonstration of low COI in half of youth with T2D, underscore that in order to better understand T2D risk, social determinants of health must be considered, rather than population ancestry alone.22
Limitations of our study include the single center and limited time range, beginning during the COVID-19 pandemic. However, youth-onset T2D incidence rose significantly during this period,23 and our study further characterizes this high-risk population. Because our study only included youth with T2D who were admitted for diabetes management, youth in this study may represent a group at higher risk of adverse COI than those whose new-onset T2D was addressed in the outpatient setting. Finally, our cohort reflects the very low population of Hispanic/Latinx youth in our region,24 an important limitation given the higher prevalence of youth-onset T2D in this population.
In conclusion, COI is significantly lower in youth with T2D and is strongly associated with clinical presentation in youth with new-onset diabetes, including presence of DKA at diagnosis. COI may be a particularly important marker of risk for DKA and for youth-onset T2D and should be the focus of future studies to guide screening and prevention efforts.
Data availability statement
Data are available upon reasonable request. De-identified data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
Approval was granted by the University of Pittsburgh Institutional Review Board (STUDY21020049).
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors MEV, SA, and IL designed the study. IL, NM, and KH performed data collection. MEV performed data analysis. KH and MEV wrote the manuscript. SA, IL, NM and YHH provided critical review of the manuscript. MEV is the guarantor of this work, and as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Funding MEV was supported by NIH (K23DK125719). KH was supported by the Richard L Day Endowed Chair, held by SA.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.