# Clinical Cases: Health Technology in Medicine

## Case 1: AI-Assisted Radiology Diagnosis Error

### Patient Presentation
**Demographics:** 54-year-old female office manager

**Chief Complaint:** "My doctor said the AI found something on my chest X-ray and I need more tests."

**History of Present Illness:**
A 54-year-old woman underwent a routine pre-operative chest X-ray prior to elective cholecystectomy. The radiology department utilizes a commercially available AI-assisted detection system for preliminary reads. The AI algorithm flagged a 1.2 cm opacity in the right upper lobe as "suspicious for malignancy" with a confidence score of 87%. The patient was informed of the finding and referred for CT imaging.

Upon CT chest with contrast, no discrete pulmonary nodule was identified. The opacity seen on chest X-ray corresponded to an overlap of the first rib and clavicle creating a pseudo-nodule artifact. The AI system had generated a false-positive result, causing significant patient anxiety and a two-week delay in her elective surgery.

A retrospective review of the AI system's performance over the prior quarter revealed a false-positive rate of 11.3% for pulmonary nodule detection, with bone overlap artifacts being the most common cause (38% of false positives). The attending radiologist had not reviewed the AI-flagged image before the result was communicated to the referring physician.

**Past Medical History:**
- Cholelithiasis (symptomatic)
- Hypertension, well-controlled
- Anxiety disorder

**Medications:**
- Lisinopril 10 mg daily
- Sertraline 50 mg daily

**Social History:**
- Never smoker
- Social alcohol use
- Works as office manager, sedentary lifestyle

**Family History:**
- Mother: breast cancer at age 67
- No family history of lung cancer

### Physical Examination
- **Vital Signs:** BP 138/86 mmHg, HR 88 bpm, RR 16/min, Temp 36.8°C, SpO2 99% on room air
- **General:** Anxious-appearing female, tearful during examination
- **Lungs:** Clear to auscultation bilaterally, no wheezes or crackles
- **Cardiovascular:** Regular rate and rhythm, no murmurs
- **Abdomen:** Mild RUQ tenderness, positive Murphy's sign

### Workup and Results

**Laboratory Studies:**
| Test | Result | Reference Range |
|------|--------|-----------------|
| WBC | 7,200/µL | 4,500-11,000/µL |
| Hemoglobin | 13.8 g/dL | 12.0-16.0 g/dL |
| CEA | 1.2 ng/mL | <3.0 ng/mL |
| ALT | 42 U/L | 7-56 U/L |
| AST | 38 U/L | 10-40 U/L |

**Imaging/Additional Studies:**
- **Chest X-ray (AI-flagged):** 1.2 cm opacity in right upper lobe, AI confidence score 87% for malignancy
- **CT Chest with contrast:** No pulmonary nodule identified; opacity on X-ray corresponds to first rib-clavicle overlap artifact; lungs clear bilaterally
- **AI System Audit:** False-positive rate 11.3% over prior quarter; bone overlap artifacts accounted for 38% of false positives

### Clinical Image

![AI Radiology False Positive Diagram](case_01_image.jpg)

*Diagram illustrating how overlapping bone structures on chest X-ray can create pseudo-nodule artifacts that trigger false-positive AI detection alerts. Source: Educational illustration.*

### Diagnosis
**AI-Generated False-Positive Pulmonary Nodule Detection Due to Bone Overlap Artifact**

**Key Diagnostic Criteria:**
- AI system flagged opacity with high confidence score
- CT chest confirmed no true pulmonary nodule
- Opacity corresponded to anatomical bone overlap (first rib and clavicle)
- Workflow gap: AI result communicated before radiologist review

### Treatment Plan
1. Reassurance and patient education regarding the false-positive finding
2. No pulmonary follow-up required
3. Proceed with elective cholecystectomy as originally planned
4. Implement mandatory radiologist verification before AI results are communicated to referring physicians
5. Retrain AI model with augmented dataset including bone overlap artifacts
6. Establish departmental protocol for AI-assisted reads requiring attending sign-off

### Key Learning Points
- AI diagnostic tools are adjuncts to, not replacements for, physician clinical judgment
- False-positive results from AI systems can cause significant patient harm through anxiety, unnecessary testing, and treatment delays
- Bone overlap artifacts on chest X-ray are a well-known source of false-positive AI detections
- Institutional protocols must mandate physician review of AI-flagged findings before results reach patients or referring providers
- Regular performance audits of AI diagnostic systems are essential for quality assurance

---

## Case 2: Telemedicine Management of Chronic Disease

### Patient Presentation
**Demographics:** 68-year-old male retired farmer, rural community

**Chief Complaint:** "My blood sugars have been all over the place and I can't get to the clinic easily."

**History of Present Illness:**
A 68-year-old male with a 15-year history of type 2 diabetes mellitus presents via telemedicine video visit from his home, which is 87 miles from the nearest endocrinology clinic. He was transitioned to telemedicine follow-up 6 months ago after his endocrinologist implemented a remote patient monitoring (RPM) program. He uses a continuous glucose monitor (CGM) that transmits data in real-time to his care team.

Review of his CGM data over the past 14 days reveals a time-in-range (70-180 mg/dL) of only 42% (target >70%), with significant postprandial hyperglycemia reaching 280-320 mg/dL after dinner and overnight hypoglycemia with glucose values dropping to 54-62 mg/dL between 2:00 AM and 4:00 AM. His most recent HbA1c was 8.9%, up from 7.8% three months ago.

During the video visit, the patient demonstrates his injection technique, which reveals he has been injecting his evening insulin into a lipohypertrophic area on his abdomen. He also reports skipping his morning metformin dose because it causes nausea, and he did not know he could take it with food to reduce GI side effects.

**Past Medical History:**
- Type 2 diabetes mellitus (15 years)
- Hypertension
- Diabetic peripheral neuropathy
- Chronic kidney disease stage 3a (eGFR 52 mL/min)
- Osteoarthritis of both knees

**Medications:**
- Insulin glargine 32 units at bedtime
- Insulin lispro 8 units before meals
- Metformin 1000 mg twice daily (admits to skipping AM dose)
- Lisinopril 20 mg daily
- Gabapentin 300 mg TID
- Aspirin 81 mg daily

**Social History:**
- Retired farmer, lives with wife in rural area
- Nearest pharmacy 35 miles away (uses mail-order)
- Limited broadband internet; uses cellular hotspot for telemedicine
- Former smoker (quit 10 years ago, 20 pack-year history)

**Family History:**
- Father: type 2 diabetes, died of MI at age 72
- Mother: hypertension, stroke at age 78

### Physical Examination
- **Vital Signs (patient-reported home readings):** BP 148/88 mmHg, HR 76 bpm, Weight 98 kg (BMI 32.1)
- **General (video assessment):** Overweight male, appears well, no acute distress
- **Skin (shown on camera):** Lipohypertrophic nodule approximately 3 cm diameter on left lower abdomen at injection site; feet examined via camera showing intact skin, no ulcers
- **Extremities:** Mild bilateral lower extremity edema noted on video

### Workup and Results

**Laboratory Studies (drawn at local community lab 3 days prior):**
| Test | Result | Reference Range |
|------|--------|-----------------|
| HbA1c | 8.9% | <7.0% (target) |
| Fasting glucose | 186 mg/dL | 70-100 mg/dL |
| Creatinine | 1.4 mg/dL | 0.7-1.3 mg/dL |
| eGFR | 52 mL/min/1.73m² | >60 mL/min/1.73m² |
| Urine albumin/creatinine ratio | 88 mg/g | <30 mg/g |
| Total cholesterol | 218 mg/dL | <200 mg/dL |
| LDL | 128 mg/dL | <100 mg/dL |
| Potassium | 4.6 mEq/L | 3.5-5.0 mEq/L |

**CGM Data (14-day summary):**
- Time in range (70-180 mg/dL): 42%
- Time above range (>180 mg/dL): 48%
- Time below range (<70 mg/dL): 10%
- Glucose management indicator (GMI): 8.7%
- Coefficient of variation: 41% (target <36%)

### Clinical Image

![Telemedicine RPM Dashboard](case_02_image.jpg)

*Illustration of a remote patient monitoring dashboard showing CGM data trends, ambulatory glucose profile, and time-in-range statistics used for telemedicine diabetes management. Source: Educational illustration.*

### Diagnosis
**Poorly Controlled Type 2 Diabetes Mellitus with Insulin Lipohypertrophy, Nocturnal Hypoglycemia, and Medication Non-adherence, Managed via Telemedicine with Remote Patient Monitoring**

**Key Diagnostic Criteria:**
- HbA1c 8.9% (worsening from 7.8%)
- CGM showing time-in-range of only 42%
- Nocturnal hypoglycemia pattern (2:00-4:00 AM)
- Postprandial hyperglycemia after dinner
- Lipohypertrophy at injection site identified via video examination
- Medication non-adherence (skipping morning metformin)

### Treatment Plan
1. **Injection site rotation:** Educate patient to rotate injection sites, avoiding the lipohypertrophic area; when switching sites, reduce insulin glargine dose by 20% (to 26 units) as absorption will improve in non-lipohypertrophic tissue
2. **Insulin adjustment:** Reduce bedtime glargine to 26 units to address nocturnal hypoglycemia; increase dinner lispro to 10 units for postprandial coverage
3. **Metformin adherence:** Instruct to take morning metformin with breakfast to reduce GI side effects; consider switching to extended-release formulation
4. **Add statin:** Start atorvastatin 40 mg daily for LDL >100 with diabetic nephropathy
5. **RPM intensification:** Increase CGM data review to weekly for the next month; set up automated alerts for glucose <70 mg/dL
6. **Follow-up:** Telemedicine visit in 2 weeks to review CGM data after changes; repeat HbA1c in 3 months
7. **Local coordination:** Send updated care plan to patient's local primary care provider for in-person foot exam and blood pressure reassessment

### Key Learning Points
- Telemedicine with remote patient monitoring can effectively manage chronic diseases in rural and underserved populations, overcoming geographic barriers
- CGM data review during telemedicine visits provides actionable glycemic pattern data that HbA1c alone cannot reveal
- Video-based physical examination, while limited, can identify important findings such as lipohypertrophy and foot abnormalities
- Insulin absorption is significantly impaired at lipohypertrophic sites, requiring dose adjustment when rotating to new sites
- Successful telemedicine chronic disease management requires integration with local laboratory and pharmacy services

---

## Case 3: Wearable Device Detection of Atrial Fibrillation

### Patient Presentation
**Demographics:** 47-year-old male software engineer

**Chief Complaint:** "My smartwatch keeps telling me I have an irregular heartbeat."

**History of Present Illness:**
A 47-year-old male presents to his primary care physician after his consumer smartwatch (photoplethysmography-based heart rhythm monitoring feature) generated three "irregular rhythm" notifications over the past week. The notifications occurred during periods of rest -- twice while sitting at his desk and once while watching television in the evening. He denies palpitations, chest pain, dyspnea, lightheadedness, or syncope during these episodes.

He purchased the smartwatch 2 months ago primarily for fitness tracking. The device's irregular rhythm notification feature uses a photoplethysmography (PPG) sensor to detect pulse irregularity and, when triggered, prompts the user to record a single-lead electrocardiogram (ECG) via the watch. The patient recorded ECGs during two of the three notifications. Both recordings, which he shows on his smartphone app, demonstrate an irregularly irregular rhythm with absent P waves consistent with atrial fibrillation, with ventricular rates of 92 and 108 bpm and recording durations of 30 seconds each.

The patient has no prior history of atrial fibrillation or cardiac disease. He does report a 3-year history of untreated obstructive sleep apnea (diagnosed by home sleep study but declined CPAP) and drinks 3-4 craft beers on weekend evenings.

**Past Medical History:**
- Obstructive sleep apnea (moderate, AHI 22; untreated)
- Obesity (BMI 31.4)
- Pre-hypertension

**Medications:**
- None

**Social History:**
- Software engineer, sedentary work
- Alcohol: 3-4 beers on Friday and Saturday evenings
- No tobacco or illicit drug use
- Married, two children
- Exercises sporadically (walks 2-3 times per week)

**Family History:**
- Father: atrial fibrillation diagnosed at age 62, on anticoagulation
- Mother: hypothyroidism

### Physical Examination
- **Vital Signs:** BP 134/82 mmHg, HR 78 bpm (regular at time of visit), RR 14/min, Temp 36.9°C, SpO2 97% on room air, BMI 31.4
- **General:** Well-appearing, obese male in no acute distress
- **HEENT:** Mallampati class III, large neck circumference (43 cm)
- **Cardiovascular:** Regular rate and rhythm at time of exam, no murmurs, rubs, or gallops; no JVD
- **Lungs:** Clear to auscultation bilaterally
- **Extremities:** No peripheral edema, pulses 2+ throughout

### Workup and Results

**Laboratory Studies:**
| Test | Result | Reference Range |
|------|--------|-----------------|
| TSH | 2.4 mIU/L | 0.4-4.0 mIU/L |
| Free T4 | 1.1 ng/dL | 0.8-1.8 ng/dL |
| BMP | Normal | - |
| Magnesium | 1.9 mg/dL | 1.7-2.2 mg/dL |
| BNP | 68 pg/mL | <100 pg/mL |
| CBC | Normal | - |
| Hemoglobin | 15.2 g/dL | 13.5-17.5 g/dL |
| Creatinine | 0.9 mg/dL | 0.7-1.3 mg/dL |

**Imaging/Additional Studies:**
- **Smartwatch ECG recordings (x2):** Irregularly irregular rhythm, absent P waves, ventricular rates 92 and 108 bpm, consistent with atrial fibrillation
- **In-office 12-lead ECG:** Normal sinus rhythm at 78 bpm, no ST changes, normal intervals
- **Echocardiogram:** Normal LV size and function (EF 60%), left atrial volume index 32 mL/m² (mildly enlarged), no valvular abnormalities
- **14-day continuous ambulatory ECG patch:** Three episodes of atrial fibrillation captured (durations 22 minutes, 3.8 hours, and 45 minutes), all asymptomatic; maximum ventricular rate 124 bpm; total AF burden 2.1%
- **CHA₂DS₂-VASc Score:** 0 (age <65, no HTN, DM, stroke, vascular disease, or female sex)

### Clinical Image

![Smartwatch AF Detection Pathway](case_03_image.jpg)

*Diagram showing the clinical pathway from consumer wearable device irregular rhythm detection through confirmatory medical-grade monitoring and treatment decision-making. Source: Educational illustration.*

### Diagnosis
**Paroxysmal Atrial Fibrillation Detected by Consumer Wearable Device, with Contributing Obstructive Sleep Apnea**

**Key Diagnostic Criteria:**
- Consumer smartwatch PPG sensor detected irregular rhythm on three occasions
- Smartwatch single-lead ECG recordings showed irregularly irregular rhythm without P waves
- Confirmed by 14-day ambulatory ECG monitor: three AF episodes (longest 3.8 hours)
- All episodes asymptomatic
- Left atrial volume mildly enlarged on echocardiogram
- Untreated moderate obstructive sleep apnea as contributing factor

### Treatment Plan
1. **Anticoagulation decision:** CHA₂DS₂-VASc score 0; anticoagulation not currently indicated per guidelines; reassess annually or with development of new risk factors
2. **Rate vs. rhythm control:** Given low AF burden (2.1%) and asymptomatic episodes, adopt watchful waiting with rate control strategy; no antiarrhythmic drugs at this time
3. **Address modifiable risk factors:**
   - Initiate CPAP therapy for obstructive sleep apnea (treatment of OSA reduces AF recurrence by 40-50%)
   - Weight loss counseling (target BMI <30)
   - Alcohol moderation (reduce to ≤2 standard drinks/occasion; consider elimination trial)
   - Structured exercise program
4. **Monitoring:** Continue smartwatch monitoring as adjunct; repeat 14-day monitor in 6 months to reassess AF burden
5. **Follow-up:** Cardiology follow-up in 3 months; reassess symptoms and AF burden; echocardiogram annually to monitor left atrial size

### Key Learning Points
- Consumer wearable devices with PPG-based irregular rhythm detection have demonstrated positive predictive values of 71-84% for atrial fibrillation in validation studies
- Wearable-detected arrhythmias require confirmation with medical-grade monitoring (ambulatory ECG patch or Holter) before treatment decisions
- Obstructive sleep apnea is an independent risk factor for atrial fibrillation, and its treatment significantly reduces AF recurrence
- CHA₂DS₂-VASc scoring guides anticoagulation decisions; a score of 0 in males generally does not warrant anticoagulation despite confirmed AF
- The detection of asymptomatic AF by consumer devices raises important clinical questions about screening, overdiagnosis, and the threshold for intervention