HIDDENCARE • CDC NHANES Data • Algorithmic Epidemiology Lab

HiddenCare compares two ways of choosing who gets screened for diabetes when a clinic cannot test everyone.

Interactive research study on public health data. Not an individual diagnosis tool or hospital-deployed system.

When screening algorithms learn only from past diagnosis records, people who rarely visit doctors can be misclassified as low risk. This project compares how training label definitions reorder screening triage priorities across a population under finite testing capacity.

What I Built
Data Engineering: Curated and validated N = 12,459 adult records from CDC NHANES (2017 to 2023) with complex survey weights and missingness audits.
Model Comparison: Trained gradient-boosted models comparing administrative diagnosis labels (Model A) against biomarker-augmented labels (Model B: HbA1c ≥ 6.5%).
Screening Simulator: Built an interactive triage engine simulating 5% to 50% capacity constraints with dynamic queue thresholds.
Equity & Decision Audits: Quantified health insurance sensitivity divides, boundary queue swaps, and net monetary benefit via decision curves.

How to use: Adjust the screening budget slider below to compare how both models allocate test slots and detect hidden cases.

The 30-Second Translation • Plain Language Summary

When Health Algorithms Overlook People Who Don't Visit Doctors

1. Finite Clinic Budget

A community clinic only has enough funding to test a fraction of residents (e.g., 20% of the town). Who gets invited first?

2. The Healthcare Access Trap

Standard AI (Model A) trains on past doctor visits. Because uninsured people rarely see doctors, their records look healthy even when they are sick.

3. The Biological Rescue

At the 20% budget cutoff, Model B swaps 61 overlooked people into the queue, achieving a 3.0x higher hit rate (4.92% vs 1.64%, or 3 cases vs 1 case) in that marginal swapped cohort compared to Model A. Across the full 939-person queue, Model B detects 55 latent cases compared to 53 for Model A (+2 net cases overall).

Evaluation Cohort • Aug 2021 to Aug 2023
N = 4,698 Adults

Strictly previously undiagnosed adults with zero temporal overlap with training data. Biochemically confirmed 116 latent diabetes cases (HbA1c ≥ 6.5%).

• Unweighted Background Prevalence: 2.47%
• Predictors: Age, Sex, BMI, Waist Circumference
Training Benchmark • 2017 to Mar 2020
N = 7,761 Adults

Pre-pandemic baseline comparing administrative self-report labels against biomarker-augmented screening labels.

• Label A (Diagnosis Only): 1,257 cases (16.20%)
• Label B (Dx OR HbA1c ≥ 6.5%): 1,484 cases (19.12%)
01 / Simulation Engine

Screening Queue Triage Under Finite Capacity

Adjust screening capacity to simulate how proxy-trained versus biomarker-augmented triage prioritizes previously undiagnosed adults (N = 4,698 holdout).

5 50

Quick Budget Presets:

Random Baseline (20% Capacity)
23.2 / 116 cases
Sensitivity (% caught)20.0%
Precision (hit rate)2.47%
Tests per case (NNS)40.5 tests
Screening Quota939.0 / 4,698
Model A (Doctor History Label)
53.0 / 116 cases
Sensitivity (% caught)45.69%
Precision (hit rate)5.64%
Tests per case (NNS)17.7 tests
DiscriminationROC AUC 0.7300
Model B (CDC Blood Test Label) +2.0 Cases
55.0 / 116 cases
Sensitivity (% caught)47.41% (+1.72%)
Precision (hit rate)5.86%
Tests per case (NNS)17.1 tests
DiscriminationROC AUC 0.7410

Cumulative yield: How many latent diabetes cases (HbA1c ≥ 6.5%) are identified as community screening capacity expands from 5% to 50%.© 2026 Cao Thien An Nguyen • GitHub: aansensei

Screening efficiency: Number of diagnostic blood tests needed to find one latent case. Lower values indicate higher testing efficiency.© 2026 Cao Thien An Nguyen • GitHub: aansensei

02 / Boundary Analysis

The 61-Person Margin: Who Swaps at the 20% Cutoff?

At exactly 20% capacity (K = 939 slots), 61 individuals differ between Model A and Model B queues. Inspect their joint rankings and individual clinical records.

Boundary swap precision: In the 61-person boundary swap at 20% budget, Model B achieves a 4.92% hit rate (3 / 61 cases) compared to Model A's 1.64% (1 / 61 cases).© 2026 Cao Thien An Nguyen • GitHub: aansensei

Joint priority ranking: Each dot represents a patient. Points above the diagonal line are prioritized higher by Model B than Model A.© 2026 Cao Thien An Nguyen • GitHub: aansensei

Inspect Marginal Patient Clinical Profiles (At K = 939 Boundary Cutoff)
Showing top 3 of 3 matching profiles
SEQN Demographics Anthropometry Access & Income Priority Ranking Observed HbA1c Clinical Status
#131367 58y / Female BMI 35.8 • Waist 120.5cm Insured • Missing PIR Rank A: #1038 • Rank B: #862 6.8% Rescued Case (HbA1c ≥ 6.5%)
#132056 65y / Female BMI 34.1 • Waist 108.8cm Uninsured • Low PIR Rank A: #1057 • Rank B: #915 8.2% Rescued Case (HbA1c ≥ 6.5%)
#142034 45y / Male BMI 44.2 • Waist 131.9cm Insured • Missing PIR Rank A: #1072 • Rank B: #829 6.8% Rescued Case (HbA1c ≥ 6.5%)
03 / Fairness & Disparity

Health Equity Audit Across Insurance and Income Strata

Evaluating whether training on clinical diagnosis history systematically penalizes individuals with lower healthcare access.

Structural Access Disparity & Statistical Honesty Callout

1. The 3x Insurance Sensitivity Divide: At the primary 20% capacity budget, both models successfully detect ~50.0% of latent cases among insured individuals (51/102 cases for Model A, 52/102 for Model B). Conversely, screening sensitivity drops to 15.4% (Model A) and 23.1% (Model B) among uninsured individuals, leaving over 75% of uninsured cases undetected.

2. Relative Biomarker Recovery: Incorporating the laboratory biomarker (Model B) allocates 6 additional screening slots to uninsured participants and catches +1 additional true case (+50% relative increase in yield), demonstrating that biomarker augmentation partially corrects access bias.

3. Small-Sample Integrity: In the holdout cohort of previously undiagnosed adults, the uninsured subgroup holds N = 422 participants with exactly 13 latent positive cases. In accordance with biostatistical best practices, we report exact participant counts and avoid over-extrapolating small-sample estimates into sweeping algorithmic fairness claims.

Health equity audit: Compares screening sensitivity between insured and uninsured populations, highlighting where biomarker training closes access-driven gaps.© 2026 Cao Thien An Nguyen • GitHub: aansensei

04 / Policy Economics

Decision Curve Analysis & Net Monetary Benefit

Quantifying financial and health utility trade-offs across screening test costs and complication avoidance values.

Decision curve analysis: Evaluates net clinical and economic benefit across decision threshold probabilities compared to universal or zero screening.© 2026 Cao Thien An Nguyen • GitHub: aansensei

Economic & Clinical Impact Ledger (Budget: 20%)
Allocation Policy Cases Found Total Testing Cost Cost per Case Detected Net Monetary Benefit (NMB)
Random Outreach 1011.9 $23,475 $1,012 $173,725
Model A (Diagnosis Only) 53.0 $23,475 $443 $427,025
Model B (Combined Biomarker) 55.0 $23,475 $427 $444,025
Zero Marginal Cost Policy Improvement
Because both models use the exact same testing quota (939.0 tests; $\Delta\text{Cost} = \$0$), shifting from Label A to Label B yields an Incremental Net Benefit of +$17,000 on this holdout cohort (or +$36,186 per 10,000 undiagnosed adults) by identifying 2.0 additional cases at zero additional laboratory expenditure.
Part 2 / Academic Dossier Poster & Citations

05 • How to Cite (APA 7th), Methodology, Research Poster & Sources

Inspect the vector research poster (36×24 in), review complex survey variance proofs, examine formal APA 7th literature citations, and explore author background.