---
title: "Appendix F: Case Study Library - Overview – The Public Health AI Handbook"
source: "https://publichealthaihandbook.com/appendices/case-study-overview.html"
author: "Bryan Tegomoh, MD, MPH"
last_updated: "2026-11-16"
doi: "10.5281/zenodo.18263441"
license: CC BY 4.0
generated-by: handbook-markdown-export
---

# Appendix F: Case Study Library - Overview

A collection of documented public health AI cases and explicitly labeled teaching scenarios, examining successes, failures, and implementation tradeoffs. Reported evidence is separated from illustrative code, outcome tables, and economic models.

How to Use This Appendix

- **Quick scan:** Review the summary table and key themes below
- **Specific domain:** Jump to relevant sections using the navigation links
- **Deep dive:** Read the [complete case studies](../appendices/case-studies.html)
- **Code implementation:** Cases include illustrative Python-style pseudocode and templates

Appendix Summary (TL;DR)

**The Big Picture:** The collection combines documented implementations, retrospective studies, policy controversies, and explicitly labeled teaching scenarios. It does not treat all entries as equivalent clinical deployments or calculate success rates from unlike evidence categories.

**Evidence categories:**

- **Documented implementations and platforms:** BlueDot, ProMED and HealthMap, IDx-DR, Google Flu Trends, Allegheny Family Screening Tool, Crisis Text Line, and MomConnect
- **Retrospective or research-stage systems:** the Veterans Affairs AKI prediction model, mammography models, sepsis reinforcement-learning research, COVID-19 prediction models, and AI-assisted drug discovery
- **Teaching scenarios:** the NHS-inspired disparity audit, hospital bed-allocation model, and AI extension to Project ECHO

**Top 10 Lessons Across All Cases:**

1.  **Technical performance does not establish clinical impact:** Retrospective model performance requires prospective evaluation
2.  **External validation is mandatory:** Performance must be tested across intended sites, populations, equipment, and time periods
3.  **Fairness requires active design:** Group performance and downstream allocation effects require direct measurement
4.  **Human oversight must be operationally defined:** Responsibility, escalation, and override pathways cannot remain slogans
5.  **Context matters profoundly:** The same algorithm performs differently across settings (equipment, workflows, populations)
6.  **Economic value requires local measurement:** Illustrative ROI models are not substitutes for observed costs and outcomes
7.  **Implementation is half the battle:** Algorithm quality is insufficient without workflow integration, training, and change management
8.  **Transparency builds trust:** Explainability and public documentation increase accountability
9.  **Continuous monitoring is required:** Performance can degrade over time (Google Flu Trends model drift)
10. **Some decisions should remain human:** Life-or-death decisions, such as ventilator allocation and crisis-line escalation, require human judgment

**Coverage:**

- **10 domains:** Surveillance, diagnostics, treatment, resource allocation, population health, health economics, mental health, drug discovery, rural health, maternal health
- Python-style pseudocode and implementation templates span surveillance, prediction, optimization, fairness, and monitoring
- The collection spans multiple generations of public health AI

**Key Insight for Practitioners:** External validation tests whether performance transports to different sites, populations, equipment, and time periods. The size and direction of the performance change are empirical questions, not a fixed percentage.

**Economic Reality:** Economic evaluation should report local implementation costs, operating costs, opportunity costs, measured outcomes, uncertainty, and the comparator. The bed-allocation and Project ECHO extensions are teaching models, not reported ROI studies.

**Fairness Reality:** Historical data can encode inequity. Fairness evaluation should examine group-specific errors, access, allocation, and outcomes, then test whether proposed mitigations improve the relevant harms.

**Takeaway:** Study failures before building. Recurrent problems include observational confounding, inadequate validation, workflow mismatch, unclear accountability, and inappropriate use cases.

------------------------------------------------------------------------

## Case Studies at a Glance

| \# | Case Study | Domain | Outcome | Key Metric |
|----|----|----|----|----|
| 1 | [BlueDot COVID-19 Detection](#cs1-bluedot) | Surveillance | Documented alert | Client alert on December 31, 2019 |
| 2 | [Google Flu Trends](#cs2-google-flu) | Surveillance | Failure and discontinuation | More than twofold overestimation (2012–2013) |
| 3 | [ProMED + HealthMap](#cs3-promed) | Surveillance | Complementary workflow | Automated aggregation plus expert interpretation |
| 4 | [IDx-DR Autonomous Diagnostic](#cs4-idxdr) | Diagnostics | Authorized device | FDA De Novo classification, 2018 |
| 5 | [DeepMind AKI Work](#cs5-deepmind-aki) | Diagnostics | Distinct systems | Streams was rule-based; VA model was retrospective |
| 6 | [Breast Cancer AI](#cs6-mammography) | Diagnostics | Mixed | Performance varies by product, population, equipment, threshold, and workflow |
| 7 | [Sepsis RL Treatment](#cs7-sepsis) | Treatment | Controversial | RCT needed |
| 8 | [COVID-19 Prediction Models](#cs8-covid-models) | Treatment | Mostly failed | All 232 at high or unclear risk of bias |
| 9 | [Ventilator Allocation](#cs9-ventilators) | Resources | Ethical and policy analysis | Human review, transparency, and appeals emphasized |
| 10 | [Allegheny Child Welfare](#cs10-allegheny) | Population Health | Controversial | Calibration and threshold-based group differences |
| 11 | [NHS-Inspired Disparity Audit](#cs11-nhs) | Population Health | Teaching scenario | Illustrative equity-audit workflow |
| 12 | [Hospital Bed Allocation](#cs12-bed-allocation) | Health Economics | Teaching scenario | Illustrative optimization and ROI model |
| 13 | [Crisis Text Line](#cs13-crisis-text) | Mental Health | Documented platform plus teaching scenario | Illustrative evaluation metrics |
| 14 | [AlphaFold Drug Discovery](#cs14-alphafold) | Drug Discovery | Research and development programs | Product-specific evidence required |
| 15 | [Project ECHO + AI](#cs15-echo) | Rural Health | Teaching scenario | Hypothetical AI extension |
| 16 | [MomConnect](../appendices/case-studies.html#case-study-16-momconnect) | Maternal Health | National platform | SMS messaging with NLP question matching and human escalation |

The full cases provide illustrative pseudocode and source-linked evidence.

------------------------------------------------------------------------

## Part I: Disease Surveillance and Outbreak Detection

### Case Study 1: BlueDot - Early COVID-19 Detection

**Context:** Automated global disease surveillance using news reports, flight data, and climate patterns.

**Key Achievement:**

- Alerted clients on **December 31, 2019**, the same date WHO’s China Country Office identified the Wuhan notice and before WHO’s January 4 public communication ([WHO, 2024](https://www.who.int/news/item/30-12-2024-milestone-covid-19-five-years-ago))
- Predicted initial spread destinations (Bangkok, Hong Kong, Tokyo, Seoul, Singapore)

**Limitations:**

- Could not predict pandemic severity or spread dynamics
- Early warning alone insufficient without policy action

**Technical Highlights:**

- Multi-source data integration (news in 65 languages, flight networks, climate data)
- NLP for disease mention extraction
- Geographic risk scoring
- 24/7 automated monitoring

**Lesson:** AI excels at early detection, not prediction of impact

**[→ Read full case study](../appendices/case-studies.html#case-study-1-bluedot)**

------------------------------------------------------------------------

### Case Study 2: Google Flu Trends - Rise and Fall

**Context:** Search query-based flu surveillance (2008–2015)

**Timeline:**

- **2008–2011:** Accurate predictions (correlation 0.90 with CDC data)
- **2012–2013:** Estimates exceeded CDC data more than twofold at the peak
- **2013–2015:** Continued overestimation; public estimates ended in 2015

**Why It Failed:**

- Correlation without causation and an unpublished methodology
- Search behavior changes (media coverage effects)
- No update mechanism (model drift)
- Overfitting to data artifacts

**Proposed Remedies:**

- Combined with CDC data (ensemble approach)
- Increased transparency
- Regular recalibration
- Acknowledged limitations

**Lesson:** Correlation-based models need ongoing validation, transparency, and update mechanisms

**[→ Read full case study](../appendices/case-studies.html#case-study-2-google-flu)**

------------------------------------------------------------------------

### Case Study 3: ProMED-mail + HealthMap - Human-AI Collaboration

**Context:** Hybrid human-AI disease surveillance system

**Complementary Functions:**

- ProMED provides expert-curated outbreak reports and moderator context
- HealthMap aggregates and classifies online disease reports
- A combined workflow requires explicit review criteria and evaluation

**Key Innovation:** Automated collection can extend breadth while experts provide contextual interpretation

**Lesson:** Humans + AI \> Either alone

**[→ Read full case study](../appendices/case-studies.html#case-study-3-promed-healthmap)**

------------------------------------------------------------------------

## Part II: Diagnostic AI

### Case Study 4: IDx-DR - First Autonomous AI Diagnostic

**Regulatory Milestone:**

- **First FDA-authorized autonomous diagnostic AI** (April 2018)
- Provides a screening result without clinician image interpretation

**Performance:**

- Sensitivity: 87.2%
- Specificity: 90.7%
- Clinical trial: 900 patients, 10 sites

**Real-World Challenges:**

- Performance can change with image quality, population, camera, and workflow
- Reimbursement challenges

**Regulatory Pathway:**

- FDA De Novo classification
- Extensive clinical validation
- Post-market surveillance requirements

**Lesson:** Autonomous AI possible but requires extensive validation and monitoring

**[→ Read full case study](../appendices/case-studies.html#case-study-4-idx-dr)**

------------------------------------------------------------------------

### Case Study 5: DeepMind AKI Work - Two Distinct Systems

**Evidence boundary:**

- **Royal Free Streams:** A clinically deployed app using the NHS national rule-based AKI algorithm, not machine learning
- **VA prediction model:** A retrospective recurrent neural network trained on 703,782 US Veterans Affairs patients
- **Reported performance:** Predicted 55.8% of inpatient AKI and 90.2% of dialysis-requiring AKI up to 48 hours in advance at two false alerts per true alert ([Tomašev et al., 2019](https://doi.org/10.1038/s41586-019-1390-1))
- **Clinical impact:** Not established by a prospective deployment trial

**Why the distinction matters:**

- Royal Free’s data-sharing controversy concerns governance of Streams
- The VA model’s reported metrics concern retrospective prediction
- Alert burden, workflow integration, and patient outcomes require separate prospective evidence

**Critical Insight:** Retrospective prediction performance does not establish clinical utility

**Lessons for the Future:**

- Design with clinicians, not for them
- Provide actionable recommendations, not just predictions
- Integrate into existing workflows
- Clear value proposition required

**[→ Read full case study](../appendices/case-studies.html#case-study-5-deepmind-aki)**

------------------------------------------------------------------------

### Case Study 6: Breast Cancer Detection - Inconsistent Results

**Multiple Systems Evaluated:**

- Google Health/DeepMind
- Lunit INSIGHT MMG
- iCAD ProFound AI

**Performance Variability:**

- Performance changes are product-, population-, equipment-, threshold-, and workflow-specific
- Internal validation does not establish transportability
- Multi-site, multi-equipment testing is required before local deployment

**Success Story:**

- **Sweden (Lund):** 44% radiologist workload reduction, maintained detection rate
- **Key:** AI as concurrent reader, not replacement

**Lesson:** Internal validation insufficient; need multi-site, multi-equipment testing

**[→ Read full case study](../appendices/case-studies.html#case-study-6-breast-cancer)**

------------------------------------------------------------------------

## Part III: Treatment Optimization

### Case Study 7: Sepsis Treatment - AI-RL Controversy

**The AI Clinician (MIT):**

- Learned treatment policy from 100,000 ICU patients
- Recommended less fluid than standard guidelines
- **Controversial:** Observational data are subject to confounding

**The Confounding Problem:**

- Sicker patients receive more aggressive treatment → worse outcomes
- AI learns: More treatment → Worse outcomes (confounded)
- Reality: Treatment could not overcome initial severity

**Current Status:**

- The AI Clinician evidence remains retrospective and observational
- Prospective comparative evidence of clinical benefit is still required

**Lesson:** Reinforcement learning on observational data is hypothesis-generating, not practice-changing without prospective comparative validation

**[→ Read full case study](../appendices/case-studies.html#case-study-7-sepsis-rl)**

------------------------------------------------------------------------

### Case Study 8: COVID-19 Prediction Models - Limited Impact

**The Pandemic Rush:**

- 232 COVID models published by October 2020
- 98% had high risk of bias
- Only 1 externally validated with low bias
- Most never used clinically

**Common Problems:**

- Small sample sizes (\<500 patients)
- Lack of external validation
- Poor reporting standards
- Overfitting

**Models That Worked:**

- **4C Mortality Score (UK):** 35,000 patients, multiple sites, simple and interpretable
- **ISARIC-4C:** 75,000 patients, properly validated

**Lesson:** Urgency does not justify poor methods. Simple, validated models \> complex, unvalidated ones

**[→ Read full case study](../appendices/case-studies.html#case-study-8-covid-models)**

------------------------------------------------------------------------

## Part IV: Resource Allocation

### Case Study 9: Ventilator Allocation - Ethics Meets AI

**The Dilemma:**

- COVID-19 ventilator shortages required triage decisions
- AI systems proposed for allocation
- Published triage frameworks, such as Pittsburgh’s, used clinical assessment instead

**The Trilemma (Cannot maximize all three):**

1.  **Utility** (save most lives)
2.  **Fairness** (equal treatment)
3.  **Autonomy** (individual rights)

**Reasons to Limit Algorithmic Allocation:**

- Insufficient accuracy (COVID-19 survival models had C-statistics of 0.70–0.80)
- Bias concerns (perpetuate historical inequities)
- Legal risks (disability discrimination)
- Trust and legitimacy issues

**Features of Published Triage Frameworks:**

- Human clinical assessment with ethical oversight
- Triage officers (experienced clinicians)
- Appeals process
- Re-evaluation every 48–120 hours

**Lesson:** Some decisions should remain human. AI can inform but not decide life-or-death allocation.

**[→ Read full case study](../appendices/case-studies.html#case-study-9-ventilator-allocation)**

------------------------------------------------------------------------

## Part V: Population Health and Health Equity

### Case Study 10: Allegheny Family Screening Tool - Algorithmic Child Welfare

**Context:** Risk assessment for child welfare referrals (used since 2016)

**Performance:**

- Predicts child removal risk (AUC 0.76)
- Used by caseworkers to prioritize investigations

**Fairness Findings:**

- The audit found broadly similar calibration patterns across groups
- Threshold-based placement and error rates varied, with the most pronounced false-positive-rate imbalance reported for the Mixed-race subgroup
- Group-specific errors require direct measurement rather than inference from overall accuracy

**The Feedback Loop Problem:**

- Historical over-surveillance of Black and low-income families
- More system contact → Higher risk scores
- Higher scores → More investigation
- Cycle perpetuates

**Responses:**

- Public documentation and transparency
- Community engagement
- Regular fairness audits
- Human override capability maintained

**Ongoing Debate:**

- **Supporters:** More consistent than unaided human judgment
- **Critics:** Automates and scales existing discrimination
- Both perspectives have validity

**Lesson:** Historical bias in data perpetuates inequality. Transparency and community input are essential.

**[→ Read full case study](../appendices/case-studies.html#case-study-10-allegheny)**

------------------------------------------------------------------------

### Case Study 11: Hypothetical NHS-Inspired Disparity Audit

Documented UK COVID-19 disparities motivate this teaching scenario. The AI implementation, care-pathway findings, interventions, and outcome figures below are illustrative, not reported NHS results.

**Illustrative premise:**

- AI identified disparities in **human** care delivery, not AI decisions
- Used as diagnostic tool for systemic racism
- Findings led to concrete policy changes

**Illustrative disparities found:**

- Black patients: **2.5x mortality rate** (1.8x after adjusting for comorbidities)
- **8-hour longer** admission wait times for Black patients
- Lower ICU admission rates despite similar severity
- Lower guideline-concordant care rates

**Illustrative interventions:**

- Enhanced translation services (24/7 availability)
- Cultural competency training (mandatory)
- Community health workers
- Care pathway standardization
- Real-time disparity monitoring dashboards

**Illustrative results after two years:**

- Admission disparities reduced **40%**
- ICU access disparities reduced **25%**
- Mortality disparities reduced **15%**
- Disparities persist despite measurable progress

**Lesson:** AI can expose systemic problems for intervention. Used correctly, it is a tool for justice, not just a source of bias.

**[→ Read full case study](../appendices/case-studies.html#case-study-11-uk-nhs-disparities)**

------------------------------------------------------------------------

## Part VI: Health Economics

### Case Study 12: Hypothetical AI-Driven Hospital Bed Allocation

This Johns Hopkins-inspired teaching scenario preserves an optimization and economic-analysis template. The results, equity effects, ROI, and named replication sites are illustrative, not reported Johns Hopkins outcomes.

**Illustrative implementation:**

**Challenge:** Balance competing objectives:

- Efficiency (maximize utilization)
- Access (minimize wait times)
- Quality (appropriate care level)
- Equity (fair access across populations)

**Illustrative results:**

- Bed utilization: **82% → 88%** (+6 percentage points)
- ED wait times: **4.2 → 3.0 hours** (28% reduction)
- Ambulance diversions: **45% reduction**
- Elective surgery delays: **35% reduction**

**Illustrative economic model:**

- **3-Year ROI:** 2,054%
- **Total costs:** \$650,000
- **Total benefits:** \$14,004,000
- **Net benefit:** \$13,354,000
- **Payback period:** 2.3 months

**Equity Impact:**

- Illustrative reduction in racial disparities: over 80%
- Fairness constraints embedded in optimization
- Wait time disparities: Black patients +1.2 hours → +0.2 hours

**Potential evaluation settings:**

- Hospitals with comparable capacity constraints

**Lesson:** Optimization with explicit fairness constraints should be evaluated for both efficiency and equity effects

**[→ Read full case study](../appendices/case-studies.html#case-study-12-hospital-bed-allocation)**

------------------------------------------------------------------------

## Part VII: Mental Health AI

### Case Study 13: Crisis Text Line and an Illustrative AI-Triage Evaluation

Crisis Text Line has used NLP to support prioritization. The performance, capacity, outcome, and replication figures below are illustrative unless linked to a point source.

**Context:**

- High volume of text-based crisis conversations
- Suicide is a leading cause of death in the US
- Minutes matter in prevention

**Illustrative impact metrics:**

- Wait times for highest-risk: **45 min → 3 min** (93% reduction)
- Sensitivity: **92%** (detecting high-risk)
- Estimated **250 lives saved** over 7 years (conservative)
- False negative rate: **8%** (a residual risk that requires human safeguards)

**Safety Features:**

- Multiple screening layers (keywords → ML → human counselor)
- Conservative thresholds (high sensitivity, accept some false positives)
- Human counselor maintains final authority
- Continuous conversation monitoring
- Supervisor alerts for escalation

**Illustrative counselor impact:**

- 40% efficiency increase
- Better workload management
- Reduced burnout
- Context provided before conversation

**Challenges:**

- False negatives (8% miss high-risk individuals)
- Privacy concerns (AI analyzing sensitive content)
- Bias risks (addressed through continuous auditing)
- Preventing over-reliance (training emphasizes human judgment)

**Potential applicability:**

- 988 Suicide & Crisis Lifeline (US)
- Samaritans (UK)
- Lifeline Australia
- Crisis Services Canada

**Lesson:** High-stakes applications require extreme caution, multiple safety layers, and human authority

**[→ Read full case study](../appendices/case-studies.html#case-study-13-crisis-text-line)**

------------------------------------------------------------------------

## Part VIII: Drug Discovery

### Case Study 14: AlphaFold and AI-Accelerated Drug Discovery

**The AlphaFold Advance:**

- Addressed the **50-year protein structure prediction problem**
- CASP14 competition: **median GDT score of 92.4**
- Hours of computation vs months of lab work
- Democratized structural biology

**AI drug-discovery evidence boundary:**

- Multiple AI-assisted candidates have entered clinical development
- Timeline, cost, and success-rate claims require product-specific evidence
- Trial status should be checked against current registries and sponsor reports

**Where AI Helped:**

- Virtual screening (faster candidate triage)
- Lead optimization (predict properties)
- Target identification (multi-omics analysis)
- Protein structure prediction (major advance)

**Where AI Fell Short of Hype:**

- “AI eliminates need for chemists” → Still need expert chemists
- “AI drugs have higher success rates” → Too early to tell
- “AI eliminates animal testing” → Still required by regulators
- “10x faster overall” → Clinical trials remain the rate-limiting step

**Examples of AI-assisted candidates:**

- **Exscientia DSP-1181:** 5-HT1A agonist for OCD; entered phase I and was later discontinued
- **Insilico INS018_055, now rentosertib:** Investigational treatment for idiopathic pulmonary fibrosis
- **BenevolentAI BEN-2293:** Investigational treatment studied in atopic dermatitis
- **Relay Therapeutics RLY-4008:** Investigational treatment for FGFR2-altered cancers

**Economic evidence boundary:**

- Investment, valuation, time-savings, and cost-savings claims require dated product-level or portfolio-level sources
- Trial entry is not equivalent to approval, comparative effectiveness, or positive return on investment

**Lesson:** Progress is real but more modest than promotional claims. Experimental and clinical validation remain essential.

**[→ Read full case study](../appendices/case-studies.html#case-study-14-alphafold)**

------------------------------------------------------------------------

## Part IX: Rural Health

### Case Study 15: Hypothetical AI Extension to Project ECHO

Project ECHO’s hub-and-spoke model is evidence-based. The AI implementation, outcome tables, quotations, ROI, and replication claims below are illustrative, not reported Project ECHO results.

**Context:**

- Limited local specialist access in rural areas
- Longer specialist wait times than in urban areas
- Long travel distances for specialist care
- Higher mortality rates than in urban areas

**The ECHO Model:**

- Hub-and-spoke (specialists mentor PCPs)
- Case-based learning
- “Moving knowledge, not patients”
- Community of practice

**AI Enhancements:**

- Clinical decision support for PCPs
- Automated case classification
- Remote monitoring with AI triage
- Predictive analytics for high-risk patients

**Illustrative New Mexico pilot results:**

**Access Improvements:**

- PCP confidence: **4.2 → 7.8 out of 10** (+86%)
- Cases managed locally: **45% → 72%** (+27 points)
- Specialist referrals: **-38%** reduction
- Wait times: **6.5 → 2.1 weeks** (for cases still needing specialist)

**Clinical Outcomes:**

- Diabetes control: **32% → 51%** at goal (+19 points)
- Hypertension control: **48% → 64%** at goal (+16 points)
- Hepatitis C cure rate: **67% → 89%** (+22 points)
- Hospitalization rate: **-23%** reduction

**Economic Impact:**

- **3-Year ROI:** 840%
- **Cost per patient/year:** \$8,500 (traditional) → \$6,100 (ECHO+AI)
- **Savings:** \$2,400 per patient per year
- **Total savings:** \$324 million (45,000 patients over 3 years)

**Provider Impact:**

- Satisfaction: 6.2 → 8.7 out of 10
- Burnout: 58% → 34% reporting burnout

**Patient Impact:**

- Fewer 3-hour drives to specialists
- Local care with specialist backing
- Satisfaction: 7.1 → 8.9 out of 10

**Illustrative scale assumptions:**

- A replication plan could span **120 clinics**
- It could reach **~200,000 patients**
- Payer and Medicaid coverage would follow prospective evaluation

**Lesson:** Technology + human networks \> Either alone. Sustainability, ROI, and equity benefits require measured evaluation.

**[→ Read full case study](../appendices/case-studies.html#case-study-15-project-echo)**

------------------------------------------------------------------------

## Key Themes Across the Case Collection

### 1. Technical Success ≠ Clinical Impact

**Evidence:** The retrospective VA AKI model and COVID-19 prediction-model reviews show why model metrics must not be treated as clinical outcomes.

**Implication:** Must measure patient-centered endpoints, not just algorithm performance

------------------------------------------------------------------------

### 2. External Validation is Mandatory

**Evidence:** Mammography AI (internal AUC 0.95 → external 0.82), COVID models (98% high bias)

**Implication:** Internal test performance may not transport. Validate on different populations, sites, equipment, and time periods.

------------------------------------------------------------------------

### 3. Fairness Requires Active Design

**Evidence:** The Allegheny audit documents group-specific performance concerns. The bed-allocation example illustrates how fairness constraints could be evaluated.

**Implication:** Algorithms can perpetuate bias unless explicitly designed for fairness. Regular auditing is essential.

------------------------------------------------------------------------

### 4. Human-AI Collaboration Optimal

**Evidence:** ProMED and Crisis Text Line illustrate distinct human-review roles. The Project ECHO extension is a hypothetical model for designing such collaboration.

**Implication:** AI provides scale and consistency, humans provide judgment and accountability. Hybrid \> either alone.

------------------------------------------------------------------------

### 5. Context Matters Profoundly

**Evidence:** Mammography AI performance varies by equipment and setting; ventilator-allocation frameworks differed across institutions and ethical approaches.

**Implication:** Same algorithm performs differently in different settings. Must adapt to local context.

------------------------------------------------------------------------

### 6. Economic Value Requires Local Measurement

**Evidence:** The bed-allocation, Project ECHO, and Crisis Text Line examples show the inputs an economic evaluation would need. Their illustrative calculations are not observed returns.

**Implication:** Measure implementation costs, operating costs, uncertainty, outcomes, and the relevant comparator before making an ROI claim.

------------------------------------------------------------------------

### 7. Implementation Is Half the Battle

**Evidence:** Streams highlights governance and implementation requirements. The ECHO scenario illustrates training and change-management questions.

**Implication:** Algorithm quality alone is insufficient. Implementation must address change management, training, workflow integration.

------------------------------------------------------------------------

### 8. Transparency Builds Trust

**Evidence:** Allegheny and ProMED provide documented transparency mechanisms. The NHS-inspired scenario illustrates how disparity reporting could support accountability.

**Implication:** Explainability supports clinician review. Public documentation increases accountability.

------------------------------------------------------------------------

### 9. Continuous Monitoring Required

**Evidence:** Google Flu Trends illustrates model drift; authorized devices require postmarket surveillance; the bed-allocation scenario illustrates operational monitoring.

**Implication:** Performance can degrade over time and requires ongoing evaluation and model updates.

------------------------------------------------------------------------

### 10. Some Decisions Should Remain Human

**Evidence:** Ventilator-allocation frameworks emphasize accountable review; Crisis Text Line illustrates a human escalation pathway.

**Implication:** Life-or-death decisions require human judgment. AI should inform, not decide.

------------------------------------------------------------------------

## Evidence Classification

**Documented deployments and platforms:**

1.  BlueDot
2.  ProMED and HealthMap
3.  IDx-DR
4.  Google Flu Trends
5.  Allegheny Family Screening Tool
6.  Crisis Text Line
7.  MomConnect

**Research-stage or retrospective evidence:**

1.  Veterans Affairs AKI prediction model
2.  Mammography AI studies
3.  Sepsis reinforcement-learning research
4.  COVID-19 prediction models
5.  AI-assisted drug discovery

**Explicitly hypothetical teaching scenarios:**

1.  NHS-inspired disparity audit
2.  Hospital bed allocation
3.  AI extension to Project ECHO

------------------------------------------------------------------------

## Using This Appendix

### For Students

- **Start here:** Read cases relevant to your interests
- **Study implementations:** Cases include illustrative Python-style pseudocode and templates
- **Analyze outcomes:** What worked vs what did not, and why
- **Extract lessons:** Apply to your own projects

### For Practitioners

- **Before implementation:** Review cases in your domain
- **Learn from mistakes:** Study the failures to avoid repeating them
- **Adapt code:** Use examples as starting templates
- **Evaluate properly:** Follow validation frameworks demonstrated

### For Researchers

- **Identify gaps:** What has not been studied yet?
- **Deep dives:** Follow references for full literature review
- **Benchmark your work:** Compare against the documented results
- **Contribute evidence:** Help build the evidence base

### For Policymakers

- **Understand impact:** See real-world effects, not just promises
- **Evidence-based policy:** Design regulations based on actual outcomes
- **Prioritize investments:** Distinguish measured economic outcomes from illustrative models
- **Equity focus:** Learn from documented cases and clearly labeled teaching scenarios

------------------------------------------------------------------------

## Collection Scope

### Coverage

- **Geographic:** Includes US, UK, South African, and global examples
- **Domains:** Surveillance, diagnostics, treatment, resource allocation, population health, health economics, mental health, drug discovery, rural health, and maternal health
- **Evidence types:** Deployments, regulatory records, retrospective studies, reviews, controversies, and teaching scenarios

### Technical Depth

- **Code examples:** Illustrative pseudocode and implementation templates
- **Algorithms covered:** CNN, RNN, RL, NLP, optimization, ensemble methods
- **Frameworks:** TensorFlow, PyTorch, scikit-learn, XGBoost, SHAP, Fairlearn

### Evidence Base

- **References:** Primary, regulatory, official, and peer-reviewed sources linked at the point of use
- **Evaluation focus:** Validation, transportability, workflow, economic assumptions, safety, and fairness

------------------------------------------------------------------------

## Updates and Contributions

This appendix is continuously updated. For corrections or to suggest additional case studies, contact the author at [bryantegomoh.com](https://bryantegomoh.com/).

------------------------------------------------------------------------

## Citation

If you use these case studies in your work, please cite:

``` sourceCode
@incollection{tegomoh2025casestudies,
 title = {Case Study Library: Real-World AI in Public Health},
 booktitle = {The Public Health AI Handbook: Evaluating AI Tools for Public Health Practice},
 author = {Tegomoh, Bryan},
 year = {2025},
 doi = {10.5281/zenodo.18263441},
 url = {https://publichealthaihandbook.com/appendices/case-study-overview.html}
}
```

See [How to Cite This Handbook](../references.html#how-to-cite-this-handbook) for additional citation formats.

------------------------------------------------------------------------

## Next Steps

Continue with the full cases and supporting resources:

**[→ Read Complete Case Studies](../appendices/case-studies.html)**: Full technical details, code, and analysis

**[→ Code Repository Guide](../appendices/code-repository.html)**: Access companion code and examples

**[→ Further Reading](../appendices/resources.html)**: Curated resources for continued learning

------------------------------------------------------------------------

*For complete technical details, methodology, code implementations, and full analysis, see the [full case studies](../appendices/case-studies.html).*

![](../images/bryan-tegomoh-author.jpg)

[Bryan Tegomoh, MD, MPH](https://bryantegomoh.com/)

Physician-scientist and epidemiologist in health AI and biosecurity.

[Follow @BryanTegomoh](https://x.com/intent/follow?screen_name=BryanTegomoh) [Follow on LinkedIn](https://www.linkedin.com/comm/mynetwork/discovery-see-all?usecase=PEOPLE_FOLLOWS&followMember=bryan-tegomoh-md-mph) [Support](../support.html)
