Parental Income Level and Risk of Developing Type 2 Diabetes in Youth

Key Points Question Is poverty associated with youth-onset type 2 diabetes? Findings In this nationwide, population-based cohort study of more than 5 million children and adolescents in Taiwan, those from a family with very low, low, or middle income had a significantly higher hazard of youth-onset type 2 diabetes than those from high-income families. Children and adolescents from very low, low, and middle income families also had a significantly higher hazard of all-cause mortality than those from high-income families. Meaning These findings suggest that low family income is associated with increased risk of type 2 diabetes and all-cause mortality among children and adolescents; further research to reveal the factors underlying this association may improve the accuracy of identifying individuals at greatest risk for developing type 2 diabetes in youth.


Introduction
Type 2 diabetes usually occurs after the age of 50 to 60 years. 1,2However, youth-onset type 2 diabetes, also known as juvenile or adolescent type 2 diabetes, occurs most often between the ages of 10 and 20 years. 3Type 1 diabetes is the main type of diabetes in people younger than 20 years.5][6] Reports 6,7 have found a higher prevalence of type 2 diabetes than type 1 diabetes in young people aged 10 to 20 years in some countries and certain racial and ethnic groups.Young people with a diagnosis of type 2 diabetes will live with the condition for a longer time than those who receive a diagnosis in adulthood. 4,5wever, youth with type 2 diabetes may be busy with schoolwork and often miss regular appointments and medications. 3,5They may be in a rebellious stage of adolescence and may feel pressured by families and schools. 3,8Insulin secretion in adolescents with type 2 diabetes tends to decline rapidly, making them susceptible to failure of oral hypoglycemic therapy. 5,94][5] Identifying patients at high risk for developing adolescent type 2 diabetes is crucial.
A US study 6 found that type 2 diabetes in youth was more prevalent among those from historically minoritized racial groups than among White individuals.In other countries, reports have found that youth-onset type 2 diabetes is more prevalent in racially oppressed groups, Indigenous, or financially deprived populations than in less-disadvantaged populations, 5,10 with the common characteristic being poverty. 11However, these are broad epidemiological findings and inferences.
Although poverty has been linked to the development of adult type 2 diabetes, 1,12 few studies have examined the association of poverty with the risk of youth-onset type 2 diabetes. 3,5Therefore, we conducted a population-based cohort study to compare the association of different family incomes with the hazard of youth-onset type 2 diabetes in a nation with universal health coverage.

Study Population and Data Source
This study was conducted in accordance with the tenets of the Declaration of Helsinki. 13The institutional review board of the National Health Research Institutes approved this study.Participant and practitioner information were deidentified and encrypted before release to protect individual privacy.The study received a waiver of informed consent from the institutional review board.This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for observational studies.
Participants in this study were identified from the National Health Insurance Research Database (NHIRD).In 1995, Taiwan established a national health insurance program to provide health care to its citizens.National Health Insurance is a compulsory program with the government as the sole purchaser.The public pays a small premium amount, while the government and the client pay the rest.Thus, in the year 2000, approximately 99% of the 23 million people in the country were enrolled in this health insurance scheme. 14Information on people's medical appointments, including their address, age, sex, diagnoses, procedures, prescriptions, and outpatient and inpatient care This cohort study included children and adolescents aged 0 to 19 years from the 2008 NHIRD full population.Exclusion criteria were age 20 years or older, diagnosis of type 1 or type 2 diabetes, gestational diabetes, or dialysis treatment (eTable 1 in Supplement 1). 15We divided the children and adolescents into 3 age groups (0-6 years, 7-13 years, and 14-19 years) to describe the basic

JAMA Network Open | Diabetes and Endocrinology
Parental Income Level and Risk of Developing Type 2 Diabetes in Youth characteristics of this study population.We also controlled for sex, overweight (ICD-9 and ICD-10-CM coding for overweight and body mass index [calculated as weight in kilograms divided by height in meters squared] 25-29), obesity (ICD-9 and ICD-10-CM coding for obesity, obesity complicating diseases, and body mass index 30-39), severe obesity (ICD-9 and ICD-10-CM coding for severe obesity, obesity undergoing bariatric surgery, and body mass index Ն40), smoking status, alcoholrelated disorders, hypertension, gout, psychiatric disorders (including mental disorders, schizophrenia, mood disorders, delusional disorders, psychosis, and pervasive developmental disorders), and Charlson Comorbidity Index scores (see eTable 1 in Supplement 1 for related definitions). 16Information on comorbidities and Charlson Comorbidity Index scores was obtained from the 2007 NHIRD records.To increase the diagnostic validity of comorbidities, we assumed at least 2 outpatient visits or 1 inpatient diagnosis.

Economic Status
The economic status of the children and adolescents was divided into 4 groups according to their family's monthly income using data from the NHIRD (incomes are shown as US dollars throughout; as of October 31, 2023, US $1 = 32.47New Taiwan dollars): very low (ie, those recognized by the local government as living below the lowest living index, <$480), low (<$733; Taiwan's minimum monthly wage in 2009 was $733), middle ($733-$1499), and high (Ն$1500).The National Health Insurance administration waives premiums and copayments for very-low-income Medicare beneficiaries.

Main End Points
Youth-onset type 2 diabetes and all-cause mortality were the main end points of this study.To ensure diagnostic accuracy, we defined children and adolescents as having type 2 diabetes if they had 3 or more outpatient visits or 1 or more inpatient admission for type 2 diabetes within 1 year.The ICD-9 and ICD-10-CM coding algorithm for identifying type 2 diabetes was validated by a previous Taiwanese study, 17 with sensitivity of 90.9% and a positive predictive value of 92.0%.To reduce the misclassification of type 1 diabetes as type 2 diabetes, we excluded those taking insulin therapy within 3 months of diabetes diagnosis and continuing use or those receiving treatment in an emergency department or hospital for diabetic ketoacidosis (eAppendix in Supplement 1).All-cause mortality was determined from the death certificate and was verified with the National Death Registry.

Statistical Analysis
We performed the data analysis from June 9, 2022, to January 16, 2023.We used the χ 2 test to determine statistical differences between categorical variables and the Student t test for continuous variables.The incidence rate of type 2 diabetes or mortality during the follow-up period was estimated as the number of outcomes per 1000 person-years.Person-years were calculated as the time elapsed from the date of cohort entry (January 1, 2008) to the date of development of outcomes, withdrawal from the NHI program, or the end of follow-up (December 31, 2019), whichever came first.
Cox proportional hazards models were used to estimate adjusted hazard ratios (aHRs) for the development of youth-onset type 2 diabetes and all-cause mortality in the 3 lower income groups compared with the high-income group.The proportional hazards assumption was tested using the Schoenfeld residuals test and complementary log-log plots.The covariates used for adjustment in this study included clinical (age, sex, obesity, and smoking), medical (alcohol disorders, hypertension, gout, psychiatric disorders, and Charlson Comorbidity Index scores), and health care utilization (number of outpatient visits per year) related factors.We used this method to compare the hazard of inpatient, outpatient, emergency, and total diagnoses of youth-onset type 2 diabetes among adolescents in these 4 family income groups.Multiple Cox regression was used to estimate aHRs and 95% CIs.The 95% CI for the aHR was calculated assuming that the aHR followed a Poisson

Basic Characteristics
From

Main Outcomes
The incidence rates of total youth-onset type 2 diabetes were 0.52 cases per 1000 person-years for the very-low-income group, 0.40 cases per 1000 person-years for the low-income group, 0.35 cases per 1000 person-years for the middle-income group, and 0.28 cases per 1000 person-years for the high-income group ( The incidence of all-cause mortality was 0.52 cases per 1000 person-years in the very-lowincome group, 0.33 cases per 1000 person-years in the low-income group, 0.25 cases per 1000 for the middle-income group vs the high-income group.Adolescents from low-income families had a significantly higher cumulative incidence of total youth-onset type 2 diabetes (Figure 1A), inpatient diagnosed youth-onset type 2 diabetes (Figure 1B), emergency department diagnosed youth-onset type 2 diabetes (Figure 2A), and outpatient diagnosed youth-onset type 2 diabetes (Figure 2B) than those from high-income families.

Subgroup Analyses
Among the different subgroups of children and adolescents, in addition to family income, adolescents who were older, female, obese, and had dyslipidemia, gout, and psychiatric disorders details, are recorded in the NHIRD.The International Classification of Diseases, Ninth Revision (ICD-9) and International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) are used for diagnosis.The NHIRD is linked to the National Death Registry to verify death information.
distribution.Statistical significance was defined as a 2-tailed P < .05.SAS statistical software version JAMA Network Open | Diabetes and Endocrinology Parental Income Level and Risk of Developing Type 2 Diabetes in Youth JAMA Network Open.2023;6(11):e2345812. doi:10.1001/jamanetworkopen.2023.45812(Reprinted) November 30, 2023 3/11 Downloaded from jamanetwork.comby guest on 12/08/2023 9.4 (SAS Institute) and Stata statistical software version 16.1 (StataCorp) were used for statistical analyses.Kaplan-Meier methods and log-rank tests were used to examine the cumulative incidence of all and inpatient, outpatient, or emergency diagnoses of youth-onset type 2 diabetes among adolescents in these 4 income groups.Subgroup analyses were performed for the risk of youth-onset type 2 diabetes and all-cause mortality among 4 family income groups associated with age, sex, overweight, obesity, severe obesity, smoking, alcohol, hypertension, dyslipidemia, gout, and psychiatric disorders.For secondary and subgroup analyses, the potential for type I error due to multiple comparisons requires that the significance threshold be adjusted to P < .005.We performed an additional analysis by restricting the study participants to those aged 7 to 19 years and assessed the outcomes of youth-onset type 2 diabetes and hospitalization for diabetes complications among the 4 groups of children and adolescents.

Figure 1 .
Figure 1.Kaplan-Meier Curves for the Incidence of Total and Inpatient Diagnosis of Type 2 Diabetes in Youth From Different Income Groups

Table 1 .
the 2008 Taiwan NHIRD, we identified 5 182 893 children and adolescents aged 0 to 19 years Characteristics of Children and Adolescents Aged 0 to 19 Years in 2008 in Taiwan by Family Income -income group were older, more obese, and had more psychiatric disorders than those in the other 3 groups.The median (SD) follow-up period was 9.0 (0.3) years.

Table 2 )
. Compared with the high-income group, the aHRs for risk of youth-95% CI, 1.44-1.91)for the low-income group, and 1.42 (95% CI, 1.21-1.66)for the middle-income group vs the high-income group.The aHRs for the risk of receiving a diagnosis of youth-onset type 2 diabetes in the emergency department were 2.00 (95% CI, 1.61-2.48)for the very-low-income group, 1.57 (95% CI, 1.38-1.77)for the low-income group, and 1.38 (95% CI, 1.20-1.58)for the middle-income group vs the high-income group.The aHRs for the risk of outpatient diagnosed youth-onset type 2 diabetes were 1.47 (95% CI, 1.32-1.64)for the very-low-income group, 1.30 (95% CI, 1.23-1.37)for the low-income group, and 1.24 (95% CI, 1.17-1.32)for the middle-income group vs the high-income group.

Table 2 .
Risk of All-Cause Mortality and Diabetes Incidence Among Children and Adolescents Aged 0 to 19 Years in 2008 in Taiwan by Family Income Group Model was adjusted age, sex, overweight, obesity, severe obesity, smoking, alcohol, hypertension, dyslipidemia, gout, psychiatric disorders, Charlson Comorbidity Index scores, and frequency of outpatient visits per year.Downloaded from jamanetwork.combyguest on 12/08/2023 person-years in the middle-income group, and 0.21 cases per 1000 person-years in the high-income group (Table2).The aHRs for the risk of all-cause mortality were 2.18 (95% CI, 1.97-2.41)forthe very-low-income group, 1.51 (95% CI, 1.42-1.60)forthe low-income group, and 1.22 (95% CI, 1.14-1.31) a Graphs show incidence rates for total (A) and inpatient diagnosis (B) of type 2 diabetes.The P value in the log-rank tests compares the very low, low, and moderate income groups with the high income group.Figure 2. Kaplan-Meier Curves for the Incidence of Emergency Department and Outpatient Diagnosis of Type 2 Diabetes in Youth From Different Income Groups Graphs show incidence rates for emergency department (A) and outpatient (B) diagnosis of type 2 diabetes.The P value in the log-rank tests compares the very low, low, and moderate income groups with the high income group.