Overview
The development of type 2 diabetes in children and adolescents is a growing public health concern, especially among youth who already show signs of prediabetes. Prediabetes means blood sugar levels are higher than normal, but not yet high enough to meet the criteria for diabetes. Because many young people with prediabetes do not all progress at the same rate, clinicians need better tools to identify which patients are at highest risk and may benefit from closer follow-up and early intervention.
This study examined whether adding the Area Deprivation Index, or ADI, could improve prediction of type 2 diabetes development in youth with prediabetes. The ADI is a neighborhood-based measure of socioeconomic disadvantage that reflects factors such as income, education, employment, housing quality, and access to resources. It is based on a patient’s residential census block and is used to estimate the broader social conditions that may affect health.
Why socioeconomic status matters
Diabetes risk is shaped not only by biology, but also by social and environmental factors. A child living in a neighborhood with limited access to healthy food, safe places to exercise, transportation, or routine medical care may face more barriers to prevention and early treatment. These factors can influence weight gain, physical activity, medication adherence, and access to follow-up testing.
Traditional prediction models often focus on clinical data such as body mass index, family history, or laboratory values like hemoglobin A1c, also known as HbA1c. HbA1c reflects average blood sugar over the previous 2 to 3 months. While it is useful, it may not capture the full picture of risk. This study tested whether adding ADI to a model could make prediction more accurate.
Study design
The researchers used data from 665 patient encounters recorded in an electronic medical record system. These encounters involved youth diagnosed with prediabetes. The goal was to predict which patients would develop type 2 diabetes within 1 year of the prediabetes diagnosis.
To build the prediction models, the investigators used supervised machine learning methods. In simple terms, machine learning allows a computer to identify patterns in data and estimate outcomes based on previous cases. Two models were compared: one built with clinical data alone and another built with clinical data plus ADI. Model selection identified logistic regression as the best-performing approach in both data sets.
The best model using clinical data alone included HbA1c only. The best model using both clinical data and social data included HbA1c plus ADI. This result suggests that HbA1c was the strongest single predictor in this cohort, but neighborhood deprivation added meaningful information.
Main findings
Out of 665 patient encounters, 181 encounters, or 27.2%, progressed to type 2 diabetes within 1 year. That is a substantial proportion, showing that youth with prediabetes represent a clinically important high-risk group.
The performance of the two best models was measured using the area under the receiver operating characteristic curve, or AUC. The AUC is a common way to evaluate how well a model separates people who develop disease from those who do not. An AUC of 0.5 means the model performs no better than chance, while a value of 1.0 means perfect prediction.
The HbA1c-only model had an AUC of 0.68. When ADI was added, the AUC increased to 0.73. Although this is not a dramatic jump, it is a meaningful improvement in risk prediction. In practical terms, the combined model was better at identifying youth more likely to develop diabetes.
Clinical meaning
This study supports the idea that social determinants of health should be incorporated into clinical prediction tools. A child’s neighborhood can affect nutrition, physical activity opportunities, chronic stress, and access to medical care. These influences may help explain why some youth with similar laboratory results go on to develop diabetes while others do not.
For clinicians, this finding suggests that ADI may help refine decision-making in youth with prediabetes. It may support more targeted counseling, earlier referral to weight management or nutrition services, closer follow-up, and more proactive monitoring for rising HbA1c or other metabolic changes.
Importantly, ADI should not replace clinical judgment or standard medical evaluation. Rather, it may serve as an additional layer of information. A prediction model that combines biological and social factors is more likely to reflect the real-world complexity of diabetes risk.
How ADI can help in practice
The Area Deprivation Index is attractive because it uses readily available geographic information and can be linked to a patient’s address. In clinical settings, it may help flag youth who live in highly deprived communities and who may need extra support.
Potential uses include:
1. Prioritizing follow-up for patients at highest risk
2. Identifying families who may need help with transportation, food access, or community resources
3. Supporting earlier intervention in lifestyle programs
4. Helping health systems design more equitable diabetes prevention strategies
Because it reflects neighborhood conditions rather than individual income alone, ADI can capture structural barriers that are not always visible in routine clinic visits.
Limitations
As with any prediction study, there are important limitations. The data came from electronic medical records, which can include missing or incomplete information. The study was also based on patient encounters from a specific health care setting, so the findings may not apply equally to all populations or regions.
The model was evaluated over a 1-year period, which is useful for near-term prediction but does not answer longer-term risk. In addition, area-level socioeconomic measures like ADI are useful proxies, but they do not measure every aspect of a family’s lived experience. Individual factors such as diet, physical activity, stress, and genetics also matter.
Even so, the study provides strong support for including social determinants in pediatric diabetes risk models.
Implications for youth with prediabetes
Youth with prediabetes are not all the same. Some will return to normal glucose levels with lifestyle changes, while others will progress to type 2 diabetes relatively quickly. A more accurate prediction tool can help clinicians identify which children and adolescents need the most attention.
This is especially important because early diabetes in youth is often more aggressive than adult-onset diabetes and may lead to complications earlier in life. Preventing or delaying progression is therefore a major goal.
The study’s findings also highlight an equity issue: youth living in disadvantaged neighborhoods may face higher risk not because of biology alone, but because of the environments in which they live. Better prediction models can help health systems allocate resources more fairly.
Conclusion
In this study, adding the Area Deprivation Index to a prediction model improved the ability to forecast type 2 diabetes development in youth with prediabetes. The best model combined HbA1c and ADI and performed better than HbA1c alone.
These findings suggest that social context matters in diabetes prediction and that neighborhood-level socioeconomic measures can strengthen clinical algorithms. Incorporating ADI may help clinicians identify high-risk youth earlier and support more personalized, equitable prevention efforts.
