DiabetesRiskAI
Empirical Model Trained on 100,000 Patient Records|60-Second Test

Know your diabetes risk before it becomes irreversible.

1 in 3 adults have prediabetes, and over 80% don't know it. Our clinical screening engine combines the American Diabetes Association (ADA) criteria with machine learning to give you an actionable 1 to 10 risk rating.

100% Client-Side PrivacyNo Registration NeededPrintable Doctor Summary

96 Million

US adults have prediabetes (more than 1 in 3 people).

80%+

Are unaware they have prediabetes due to silent symptoms.

58% Drop

In diabetes risk through structured modest lifestyle interventions.

100,000

Patient records trained to calibrate our multivariate model.

Interactive Screener

Evidence-Based Diabetes Risk Assessment

Takes 60 seconds • 100% private on-device calculation • Calibrated on 100,000 patient records

Question 1 of 911% Completed
Step 1

Who are you taking this diabetes risk test for?

This helps us personalize question phrasing, doctor summaries, and action steps.

Competitor Comparison

Engineered to outperform the official CDC & ADA tools

The official tests from diabetes.org and cdc.gov were designed decades ago as static paper forms. We redesigned the screening experience with modern interactive engineering and real machine learning.

Core Screening CapabilityStandard ADA / CDC TestDiabetesRisk (Our Platform)
BMI Input ExperienceRigid height-weight grid tables with static columns✓Height in Ft/In (default) & Weight in KG (default) with 1-click toggles
Asian Population SensitivityBuried footnote about Asian lower weight risk✓Native BMI ≥ 23 adjusted risk threshold
Scoring ResolutionBinary "Low or High" cutoff (points ≥ 5)✓Granular 1 to 10 Risk Index + % Probability
Empirical Model BackingFixed manual point heuristics✓Logistic Regression on 100,000 Patient Records
Interactive "What-If" SimulatorNone (Static recommendations)✓Live simulator: see score drop with weight & exercise
Physician Report ExportGeneric brochure or email signup✓1-Click Printable Doctor Consultation Sheet
Clinical & Empirical Methodology

How our prediction engine works

A hybrid screening model harmonizing the American Diabetes Association (ADA) 60-second scoring matrix with multivariate logistic regression trained on 100,000 real patient records.

1. Training Dataset Architecture (100,000 Patient Records)

Our underlying empirical probabilities are calibrated against the Kaggle Diabetes Prediction Dataset (diabetes_prediction_dataset.csv), comprising 100,000 anonymized clinical patient observations.

Age Stratification

  • < 40 years old: 1.56% prevalence
  • 40 – 49 years: 6.84% prevalence (4.4x surge)
  • 50 – 59 years: 12.51% prevalence (8x surge)
  • 60+ years: 19.68% prevalence (12.6x surge)

Hypertension & Lifestyle Impact

  • Normotensive (No HTN): 6.93% prevalence
  • Diagnosed Hypertension: 27.90% prevalence (4x risk)
  • Male vs Female: 9.75% vs 7.62% prevalence
  • Current / Former Smokers: 10.21% – 17.00% prevalence

2. Empirical Logistic Regression Equation

We trained a regularized multivariate logistic regression model on the 100,000 patient dataset to calculate calibrated probability log-odds (z):

z = -6.7877
  + 4.0027 * (Age / 80)
  + 0.4203 * isMale
  + 0.7817 * hasHypertension
  + 0.3500 * onBPMedication
  + 0.4054 * isDailySmoker
  + 0.2620 * isFormerSmoker
  + 3.5044 * ((BMI - 15) / 35)
  + 0.8500 * hasFamilyHistory
  + 0.4500 * isPhysicallyInactive
  + 0.5000 * isPredisposedEthnicity

P(Diabetes) = 1 / (1 + exp(-z))

3. The Asian BMI ≥ 23 Threshold Explained

Both the World Health Organization (WHO) and the CDC have established that people of Asian descent typically possess a higher percentage of body fat and visceral adiposity at lower Body Mass Index (BMI) levels compared to people of European descent. Consequently, the threshold for increased type 2 diabetes risk is adjusted from BMI ≥ 25 down to BMI ≥ 23.

Why this matters: Millions of Asian individuals receive false reassurance from conventional BMI calculators that categorize a BMI of 24.5 as "normal", despite being at clinical risk for metabolic syndrome. Our calculator accounts for this physiological distinction automatically.

4. 100% Client-Side Privacy Guarantee

Your health answers never leave your browser. All calculations, scoring algorithms, and simulated predictions execute purely via client-side JavaScript. We do not store your answers in databases, cookies, or external analytics servers.

Help & Clinical Insights

Frequently Asked Questions

Everything you need to know about prediabetes screening, scoring science, and laboratory testing.

What is prediabetes, and why is early detection so critical?↓

Prediabetes is a serious health condition where blood sugar levels are higher than normal, but not yet high enough to be diagnosed as type 2 diabetes. According to the CDC, approximately 96 million American adults—more than 1 in 3—have prediabetes. More than 80% don't know they have it. Prediabetes significantly raises the risk of type 2 diabetes, heart disease, and stroke. However, prediabetes is almost entirely reversible through modest lifestyle modifications.

How does this platform improve upon the official CDC and ADA risk tests?↓

The standard tests from diabetes.org and cdc.gov rely on static paper-based point checklists created decades ago. Our platform delivers significant improvements: (1) A real-time, unit-flexible BMI calculator (supporting lbs/kg and ft/cm without manual conversion), (2) Asian population sensitivity with the CDC-recommended BMI ≥ 23 threshold, (3) A granular 1 to 10 Risk Index score paired with empirical probabilities calibrated on 100,000 real patient records, (4) An interactive 'What-If' Lifestyle Simulator that shows in real time how much risk drops with modest weight loss and exercise, (5) Dual mode for testing yourself or a loved one, and (6) A one-click printable physician discussion guide.

What does my 1 to 10 Risk Index score mean?↓

Scores 1 to 3 represent Low Risk (statistical probability < 5%), indicating healthy metabolic status. Scores 4 to 6 represent Moderate / Elevated Risk (prediabetes danger zone), where lifestyle changes can proactively halt progression. Scores 7 to 8 represent High Risk, and scores 9 to 10 represent Very High Risk. If you score 5 or higher, clinical guidelines strongly recommend scheduling a checkup with your physician for standard diagnostic blood work.

What blood tests should I ask my doctor to order?↓

The primary diagnostic tests for prediabetes and type 2 diabetes are: (1) Hemoglobin A1c (HbA1c) test — measures your average blood sugar levels over the past 2 to 3 months (normal: < 5.7%, prediabetes: 5.7% to 6.4%, diabetes: ≥ 6.5%); (2) Fasting Plasma Glucose (FPG) test — measures blood glucose after an 8-hour overnight fast (normal: < 100 mg/dL, prediabetes: 100 to 125 mg/dL, diabetes: ≥ 126 mg/dL); (3) Oral Glucose Tolerance Test (OGTT) — measures glucose levels before and 2 hours after drinking a sweet liquid.

Why does high blood pressure (hypertension) increase diabetes risk?↓

In our analysis of 100,000 patient records from Kaggle, patients with hypertension exhibited a 27.9% diabetes prevalence compared to only 6.9% in normotensive patients—a 4-fold increase. Chronic hypertension causes vascular endothelial dysfunction, arterial stiffening, and reduced microcapillary blood flow to skeletal muscle, directly impairing glucose delivery and cellular insulin sensitivity.

Why are Asian Americans screened at a lower BMI threshold (BMI ≥ 23)?↓

Extensive clinical research reviewed by the WHO and CDC shows that individuals of Asian descent tend to accumulate higher percentages of visceral and ectopic body fat around vital organs at lower body weights. As a result, metabolic disturbances and insulin resistance develop at a BMI of 23 kg/m², rather than the conventional Western cutoff of 25 kg/m².

Can type 2 diabetes really be prevented or reversed?↓

Yes. The landmark National Diabetes Prevention Program (DPP) clinical trial demonstrated that individuals with prediabetes who lost just 5% to 7% of their body weight (about 10 to 14 pounds for a 200-pound person) and engaged in 150 minutes of weekly moderate exercise reduced their risk of developing type 2 diabetes by 58% (and by 71% for adults over age 60).

Is my personal health data saved or shared?↓

No. Your privacy is paramount. Our evaluation algorithms execute 100% client-side in your web browser. None of your answers, height, weight, or calculated scores are transmitted to our servers, stored in databases, or shared with third parties.

Ready to discover your diabetes risk score?

Takes less than 60 seconds. 100% private and on-device.